AT32UC3L064-ZAUR

Features • High-performance, Low-power 32-bit Atmel® AVR® Microcontroller • • • • • • • • • • • • • • • • • – Compact Single-cycle RISC Instruction Set Including DSP Instructions – Read-modify-write Instructions and Atomic Bit Manipulation – Performance • Up to 64DMIPS Running at 50MHz from Flash (1 Flash Wait State) • Up to 36DMIPS Running at 25MHz from Flash (0 Flash Wait State) – Memory Protection Unit (MPU) • Secure Access Unit (SAU) providing User-defined Peripheral Protection picoPower® Technology for Ultra-low Power Consumption Multi-hierarchy Bus System – High-performance Data Transfers on Separate Buses for Increased Performance – 12 Peripheral DMA Channels Improve Speed for Peripheral Communication Internal High-speed Flash – 64Kbytes, 32Kbytes, and 16Kbytes Versions – Single-cycle Access up to 25MHz – FlashVault Technology Allows Pre-programmed Secure Library Support for End User Applications – Prefetch Buffer Optimizing Instruction Execution at Maximum Speed – 100,000 Write Cycles, 15-year Data Retention Capability – Flash Security Locks and User-defined Configuration Area Internal High-speed SRAM, Single-cycle Access at Full Speed – 16Kbytes (64Kbytes and 32Kbytes Flash), or 8Kbytes (16Kbytes Flash) Interrupt Controller (INTC) – Autovectored Low-latency Interrupt Service with Programmable Priority External Interrupt Controller (EIC) Peripheral Event System for Direct Peripheral to Peripheral Communication System Functions – Power and Clock Manager – SleepWalking Power Saving Control – Internal System RC Oscillator (RCSYS) – 32KHz Oscillator – Multipurpose Oscillator and Digital Frequency Locked Loop (DFLL) Windowed Watchdog Timer (WDT) Asynchronous Timer (AST) with Real-time Clock Capability – Counter or Calendar Mode Supported Frequency Meter (FREQM) for Accurate Measuring of Clock Frequency Six 16-bit Timer/Counter (TC) Channels – External Clock Inputs, PWM, Capture, and Various Counting Capabilities 36 PWM Channels (PWMA) – 8-bit PWM with a Source Clock up to 150MHz Four Universal Synchronous/Asynchronous Receiver/Transmitters (USART) – Independent Baudrate Generator, Support for SPI – Support for Hardware Handshaking One Master/Slave Serial Peripheral Interfaces (SPI) with Chip Select Signals – Up to 15 SPI Slaves can be Addressed Two Master and Two Slave Two-wire Interface (TWI), 400kbit/s I2C-compatible One 8-channel Analog-to-digital Converter (ADC) with up to 12 Bits Resolution – Internal Temperature Sensor 32-bit Atmel AVR Microcontroller AT32UC3L064 AT32UC3L032 AT32UC3L016 32099I–01/2012 AT32UC3L016/32/64 • Eight Analog Comparators (AC) with Optional Window Detection • Capacitive Touch (CAT) Module • • • • • – Hardware-assisted Atmel® AVR® QTouch® and Atmel® AVR® QMatrix Touch Acquisition – Supports QTouch and QMatrix Capture from Capacitive Touch Sensors QTouch Library Support – Capacitive Touch Buttons, Sliders, and Wheels – QTouch and QMatrix Acquisition On-chip Non-intrusive Debug System – Nexus Class 2+, Runtime Control, Non-intrusive Data and Program Trace – aWire Single-pin Programming Trace and Debug Interface Muxed with Reset Pin – NanoTrace Provides Trace Capabilities through JTAG or aWire Interface 48-pin TQFP/QFN/TLLGA (36 GPIO Pins) Five High-drive I/O Pins Single 1.62-3.6 V Power Supply 2 32099I–01/2012 AT32UC3L016/32/64 1. Description The Atmel® AVR® AT32UC3L016/32/64 is a complete system-on-chip microcontroller based on the AVR32 UC RISC processor running at frequencies up to 50MHz. AVR32 UC is a high-performance 32-bit RISC microprocessor core, designed for cost-sensitive embedded applications, with particular emphasis on low power consumption, high code density, and high performance. The processor implements a Memory Protection Unit (MPU) and a fast and flexible interrupt controller for supporting modern and real-time operating systems. The Secure Access Unit (SAU) is used together with the MPU to provide the required security and integrity. Higher computation capability is achieved using a rich set of DSP instructions. The AT32UC3L016/32/64 embeds state-of-the-art picoPower technology for ultra-low power consumption. Combined power control techniques are used to bring active current consumption down to 165 µA/MHz, and leakage down to 9nA while still retaining a bank of backup registers. The device allows a wide range of trade-offs between functionality and power consumption, giving the user the ability to reach the lowest possible power consumption with the feature set required for the application. The Peripheral Direct Memory Access (DMA) controller enables data transfers between peripherals and memories without processor involvement. The Peripheral DMA controller drastically reduces processing overhead when transferring continuous and large data streams. The AT32UC3L016/32/64 incorporates on-chip Flash and SRAM memories for secure and fast access. The FlashVault technology allows secure libraries to be programmed into the device. The secure libraries can be executed while the CPU is in Secure State, but not read by nonsecure software in the device. The device can thus be shipped to end customers, who will be able to program their own code into the device to access the secure libraries, but without risk of compromising the proprietary secure code. The External Interrupt Controller (EIC) allows pins to be configured as external interrupts. Each external interrupt has its own interrupt request and can be individually masked. The Peripheral Event System allows peripherals to receive, react to, and send peripheral events without CPU intervention. Asynchronous interrupts allow advanced peripheral operation in low power sleep modes. The Power Manager (PM) improves design flexibility and security. The Power Manager supports SleepWalking functionality, by which a module can be selectively activated based on peripheral events, even in sleep modes where the module clock is stopped. Power monitoring is supported by on-chip Power-on Reset (POR), Brown-out Detector (BOD), and Supply Monitor (SM). The device features several oscillators, such as Digital Frequency Locked Loop (DFLL), Oscillator 0 (OSC0), and system RC oscillator (RCSYS). Either of these oscillators can be used as source for the system clock. The DFLL is a programmable internal oscillator from 40 to 150MHz. It can be tuned to a high accuracy if an accurate refernce clock is running, e.g. the 32 KHz crystal oscillator. The Watchdog Timer (WDT) will reset the device unless it is periodically serviced by the software. This allows the device to recover from a condition that has caused the system to be unstable. The Asynchronous Timer (AST) combined with the 32KHz crystal oscillator supports powerful real-time clock capabilities, with a maximum timeout of up to 136 years. The AST can operate in counter mode or calendar mode. 3 32099I–01/2012 AT32UC3L016/32/64 The Frequency Meter (FREQM) allows accurate measuring of a clock frequency by comparing it to a known reference clock. The device includes six identical 16-bit Timer/Counter (TC) channels. Each channel can be independently programmed to perform frequency measurement, event counting, interval measurement, pulse generation, delay timing, and pulse width modulation. The Pulse Width Modulation controller (PWMA) provides 8-bit PWM channels which can be synchronized and controlled from a common timer. One PWM channel is available for each I/O pin on the device, enabling applications that require multiple PWM outputs, such as LCD backlight control. The PWM channels can operate independently, with duty cycles set independently from each other, or in interlinked mode, with multiple channels changed at the same time. The AT32UC3L016/32/64 also features many communication interfaces, like USART, SPI, and TWI, for communication intensive applications. The USART supports different communication modes, like SPI Mode and LIN Mode. A general purpose 8-channel ADC is provided, as well as eight analog comparators (AC). The ADC can operate in 10-bit mode at full speed or in enhanced mode at reduced speed, offering up to 12-bit resolution. The ADC also provides an internal temperature sensor input channel. The analog comparators can be paired to detect when the sensing voltage is within or outside the defined reference window. The Capacitive Touch (CAT) module senses touch on external capacitive touch sensors, using the QTouch technology. Capacitive touch sensors use no external mechanical components, unlike normal push buttons, and therefore demand less maintenance in the user application. The CAT module allows up to 17 touch sensors, or up to 16 by 8 matrix sensors to be interfaced. One touch sensor can be configured to operate autonomously without software interaction, allowing wakeup from sleep modes when activated. Atmel offers the QTouch library for embedding capacitive touch buttons, sliders, and wheels functionality into AVR microcontrollers. The patented charge-transfer signal acquisition offers robust sensing and includes fully debounced reporting of touch keys as well as Adjacent Key Suppression® (AKS®) technology for unambiguous detection of key events. The easy-to-use QTouch Suite toolchain allows you to explore, develop, and debug your own touch applications. The AT32UC3L016/32/64 integrates a class 2+ Nexus 2.0 On-chip Debug (OCD) System, with non-intrusive real-time trace and full-speed read/write memory access, in addition to basic runtime control. The NanoTrace interface enables trace feature for aWire- or JTAG-based debuggers. The single-pin aWire interface allows all features available through the JTAG interface to be accessed through the RESET pin, allowing the JTAG pins to be used for GPIO or peripherals. 4 32099I–01/2012 AT32UC3L016/32/64 2. Overview Block Diagram DATAOUT aWire NEXUS CLASS 2+ OCD MEMORY PROTECTION UNIT INSTR INTERFACE DATA INTERFACE M M M SAU S HSB-PB BRIDGE B GENERALPURPOSE I/Os S POWER MANAGER CLOCK CONTROLLER SLEEP CONTROLLER RESET CONTROLLER 64/32/16 KB FLASH M S CONFIGURATION PA PB S HIGH SPEED BUS MATRIX S/M LOCAL BUS 16/8 KB SRAM REGISTERS BUS PERIPHERAL DMA CONTROLLER HSB-PB BRIDGE A DMA RESET_N JTAG INTERFACE LOCAL BUS INTERFACE CAPACITIVE TOUCH MODULE DMA EVTO_N TCK TDO TDI TMS AVR32UC CPU FLASH CONTROLLER MCKO MDO[5..0] MSEO[1..0] EVTI_N MEMORY INTERFACE Block Diagram USART0 USART1 USART2 USART3 DMA Figure 2-1. SPI DMA 2.1 TWI MASTER 0 TWI MASTER 1 DIS VDIVEN CSA[16:0] CSB[16:0] SMP SYNC RXD TXD CLK RTS, CTS RCSYS RC32K XIN32 XOUT32 OSC32K XIN0 XOUT0 OSC0 SYSTEM CONTROL INTERFACE DFLL INTERRUPT CONTROLLER EXTINT[5..1] NMI PWMA[35..0] NPCS[3..0] TWCK TWD TWALM TWCK DMA RC120M TWI SLAVE 0 TWI SLAVE 1 8-CHANNEL ADC INTERFACE FREQUENCY METER TRIGGER AD[8..0] A[2..0] TIMER/COUNTER 0 TIMER/COUNTER 1 B[2..0] CLK[2..0] ASYNCHRONOUS TIMER WATCHDOG TIMER TWALM ADVREFP EXTERNAL INTERRUPT CONTROLLER PWM CONTROLLER TWD PA PB ADP[1..0] DMA RC32OUT MISO, MOSI GENERAL PURPOSE I/Os SCK GCLK[4..0] AC INTERFACE GLUE LOGIC CONTROLLER ACBP[3..0] ACBN[3..0] ACAP[3..0] ACAN[3..0] ACREFN OUT[1:0] IN[7..0] 5 32099I–01/2012 AT32UC3L016/32/64 2.2 Configuration Summary Table 2-1. Configuration Summary Feature AT32UC3L064 AT32UC3L032 AT32UC3L016 Flash 64KB 32KB 16KB SRAM 16KB 16KB 8KB GPIO 36 High-drive pins 5 External Interrupts 6 TWI 2 USART 4 Peripheral DMA Channels 12 Peripheral Event System 1 SPI 1 Asynchronous Timers 1 Timer/Counter Channels 6 PWM channels 36 Frequency Meter 1 Watchdog Timer 1 Power Manager 1 Secure Access Unit 1 Glue Logic Controller 1 Oscillators ADC Digital Frequency Locked Loop 40-150MHz (DFLL) Crystal Oscillator 3-16MHz (OSC0) Crystal Oscillator 32KHz (OSC32K) RC Oscillator 120MHz (RC120M) RC Oscillator 115kHz (RCSYS) RC Oscillator 32kHz (RC32K) 8-channel 12-bit Temperature Sensor 1 Analog Comparators 8 Capacitive Touch Module 1 JTAG 1 aWire 1 Max Frequency Packages 50MHz TQFP48/QFN48/TLLGA48 6 32099I–01/2012 AT32UC3L016/32/64 3. Package and Pinout 3.1 Package The device pins are multiplexed with peripheral functions as described in Section 3.2. TQFP48/QFN48 Pinout 36 35 34 33 32 31 30 29 28 27 26 25 PA14 VDDANA ADVREFP GNDANA PB08 PB07 PB06 PB09 PA04 PA11 PA13 PA20 Figure 3-1. PA15 PA16 PA17 PA19 PA18 VDDIO GND PB11 GND PA10 PA12 VDDIO 37 38 39 40 41 42 43 44 45 46 47 48 24 23 22 21 20 19 18 17 16 15 14 13 PA21 PB10 RESET_N PB04 PB05 GND VDDCORE VDDIN PB01 PA07 PA01 PA02 12 11 10 9 8 7 6 5 4 3 2 1 PA05 PA00 PA06 PA22 PB03 PB02 PB00 PB12 PA03 PA08 PA09 GND 7 32099I–01/2012 AT32UC3L016/32/64 TLLGA48 Pinout 37 36 35 34 33 32 31 30 29 28 27 26 25 PA15 PA14 VDDANA ADVREFP GNDANA PB08 PB07 PB06 PB09 PA04 PA11 PA13 PA20 Figure 3-2. PA16 PA17 PA19 PA18 VDDIO GND PB11 GND PA10 PA12 VDDIO 24 23 22 21 20 19 18 17 16 15 14 38 39 40 41 42 43 44 45 46 47 48 PA21 PB10 RESET_N PB04 PB05 GND VDDCORE VDDIN PB01 PA07 PA01 13 12 11 10 9 8 7 6 5 4 3 2 1 PA02 PA05 PA00 PA06 PA22 PB03 PB02 PB00 PB12 PA03 PA08 PA09 GND 8 32099I–01/2012 AT32UC3L016/32/64 3.2 Peripheral Multiplexing on I/O lines 3.2.1 Multiplexed signals Each GPIO line can be assigned to one of the peripheral functions.The following table describes the peripheral signals multiplexed to the GPIO lines. Table 3-1. GPIO Controller Function Multiplexing 48pin PIN G P I O 11 PA00 0 14 PA01 13 GPIO Function Pin Type A B C VDDIO Normal I/O USART0 TXD USART1 RTS SPI NPCS[2] 1 VDDIO Normal I/O USART0 RXD USART1 CTS SPI NPCS[3] USART1 CLK PWMA PWMA[1] PA02 2 VDDIO Highdrive I/O USART0 RTS ADCIFB TRIGGER USART2 TXD TC0 A0 4 PA03 3 VDDIO Normal I/O USART0 CTS SPI NPCS[1] USART2 TXD 28 PA04 4 VDDIO Normal I/O SPI MISO TWIMS0 TWCK 12 PA05 5 VDDIO Normal I/O (TWI) SPI MOSI 10 PA06 6 VDDIO Highdrive I/O, 5V tolerant 15 PA07 7 VDDIO 3 PA08 8 2 PA09 46 Supply D E F G H SCIF GCLK[0] CAT CSA[2] ACIFB ACAP[0] TWIMS0 TWALM CAT CSA[1] PWMA PWMA[2] ACIFB ACBP[0] USART0 CLK CAT CSA[3] TC0 B0 PWMA PWMA[3] ACIFB ACBN[3] USART0 CLK CAT CSB[3] USART1 RXD TC0 B1 PWMA PWMA[4] ACIFB ACBP[1] TWIMS1 TWCK USART1 TXD TC0 A1 PWMA PWMA[5] ACIFB ACBN[0] SPI SCK USART2 TXD USART1 CLK TC0 B0 PWMA PWMA[6] Normal I/O (TWI) SPI NPCS[0] USART2 RXD TWIMS1 TWALM TWIMS0 TWCK PWMA PWMA[7] VDDIO Highdrive I/O USART1 TXD SPI NPCS[2] TC0 A2 ADCIFB ADP[0] PWMA PWMA[8] 9 VDDIO Highdrive I/O USART1 RXD SPI NPCS[3] TC0 B2 ADCIFB ADP[1] PWMA PWMA[9] SCIF GCLK[2] EIC EXTINT[1] CAT CSB[4] PA10 10 VDDIO Normal I/O TWIMS0 TWD PWMA PWMA[10] ACIFB ACAP[1] SCIF GCLK[2] CAT CSA[5] 27 PA11 11 VDDIN Normal I/O 47 PA12 12 VDDIO Normal I/O 26 PA13 13 VDDIN Normal I/O 36 PA14 14 VDDIO 37 PA15 15 38 PA16 39 41 PWMA PWMA[0] TC0 A0 ACIFB ACAN[0] CAT CSA[7] TWIMS0 TWD CAT CSB[7] SCIF GCLK[1] CAT CSB[1] EIC EXTINT[0] CAT CSB[2] CAT CSA[4] PWMA PWMA[11] USART2 CLK TC0 CLK1 CAT SMP PWMA PWMA[12] ACIFB ACAN[1] SCIF GCLK[3] CAT CSB[5] GLOC OUT[0] GLOC IN[7] TC0 A0 SCIF GCLK[2] PWMA PWMA[13] CAT SMP EIC EXTINT[2] CAT CSA[0] Normal I/O ADCIFB AD[0] TC0 CLK2 USART2 RTS CAT SMP PWMA PWMA[14] SCIF GCLK[4] CAT CSA[6] VDDIO Normal I/O ADCIFB AD[1] TC0 CLK1 GLOC IN[6] PWMA PWMA[15] CAT SYNC EIC EXTINT[3] CAT CSB[6] 16 VDDIO Normal I/O ADCIFB AD[2] TC0 CLK0 GLOC IN[5] PWMA PWMA[16] ACIFB ACREFN EIC EXTINT[4] CAT CSA[8] PA17 17 VDDIO Normal I/O (TWI) TWIMS1 TWD PWMA PWMA[17] CAT SMP CAT DIS CAT CSB[8] PA18 18 VDDIO Normal I/O GLOC IN[4] PWMA PWMA[18] CAT SYNC EIC EXTINT[5] CAT CSB[0] TC0 A1 ADCIFB AD[4] TC0 B1 USART2 CTS 9 32099I–01/2012 AT32UC3L016/32/64 Table 3-1. GPIO Controller Function Multiplexing 40 PA19 19 VDDIO Normal I/O ADCIFB AD[5] TC0 A2 TWIMS1 TWALM PWMA PWMA[19] CAT SYNC 25 PA20 20 VDDIN Normal I/O USART2 TXD TC0 A1 GLOC IN[3] PWMA PWMA[20] SCIF RC32OUT USART2 RXD TWIMS0 TWD TC0 B1 ADCIFB TRIGGER PWMA PWMA[21] PWMA PWMAOD[21] 24 PA21 21 VDDIN Normal I/O (TWI, 5V tolerant SMBus) 9 PA22 22 VDDIO Normal I/O USART0 CTS USART2 CLK TC0 B2 CAT SMP PWMA PWMA[22] ACIFB ACBN[2] 6 PB00 32 VDDIO Normal I/O USART3 TXD ADCIFB ADP[0] SPI NPCS[0] TC0 A1 PWMA PWMA[23] ACIFB ACAP[2] 16 PB01 33 VDDIO Highdrive I/O USART3 RXD ADCIFB ADP[1] SPI SCK TC0 B1 PWMA PWMA[24] 7 PB02 34 VDDIO Normal I/O USART3 RTS USART3 CLK SPI MISO TC0 A2 PWMA PWMA[25] 8 PB03 35 VDDIO Normal I/O USART3 CTS USART3 CLK SPI MOSI TC0 B2 VDDIN Normal I/O (TWI, 5V tolerant SMBus) TC1 A0 USART1 RTS USART1 CLK TC1 B0 USART1 CTS 21 PB04 36 CAT CSA[10] CAT CSA[12] SCIF GCLK[0] CAT SMP CAT CSB[10] TC1 A0 CAT CSA[9] TC1 A1 CAT CSB[9] ACIFB ACAN[2] SCIF GCLK[1] CAT CSB[11] PWMA PWMA[26] ACIFB ACBP[2] TC1 A2 CAT CSA[11] TWIMS0 TWALM PWMA PWMA[27] PWMA PWMAOD[27] TWIMS1 TWCK CAT CSA[14] USART1 CLK TWIMS0 TWCK PWMA PWMA[28] PWMA PWMAOD[28] SCIF GCLK[3] CAT CSB[14] 20 PB05 37 VDDIN Normal I/O (TWI, 5V tolerant SMBus) 30 PB06 38 VDDIO Normal I/O TC1 A1 USART3 TXD ADCIFB AD[6] GLOC IN[2] PWMA PWMA[29] ACIFB ACAN[3] EIC EXTINT[0] CAT CSB[13] 31 PB07 39 VDDIO Normal I/O TC1 B1 USART3 RXD ADCIFB AD[7] GLOC IN[1] PWMA PWMA[30] ACIFB ACAP[3] EIC EXTINT[1] CAT CSA[13] 32 PB08 40 VDDIO Normal I/O TC1 A2 USART3 RTS ADCIFB AD[8] GLOC IN[0] PWMA PWMA[31] CAT SYNC EIC EXTINT[2] CAT CSB[12] 29 PB09 41 VDDIO Normal I/O TC1 B2 USART3 CTS USART3 CLK PWMA PWMA[32] ACIFB ACBN[1] EIC EXTINT[3] CAT CSB[15] 23 PB10 42 VDDIN Normal I/O TC1 CLK0 USART1 TXD USART3 CLK EIC EXTINT[4] CAT CSB[16] 44 PB11 43 VDDIO Normal I/O TC1 CLK1 USART1 RXD 5 PB12 44 VDDIO Normal I/O TC1 CLK2 TWIMS1 TWALM GLOC OUT[1] PWMA PWMA[33] ADCIFB TRIGGER PWMA PWMA[34] CAT VDIVEN EIC EXTINT[5] CAT CSA[16] CAT SYNC PWMA PWMA[35] ACIFB ACBP[3] SCIF GCLK[4] CAT CSA[15] See Section 3.3 for a description of the various peripheral signals. Refer to ”Electrical Characteristics” on page 770 for a description of the electrical properties of the pin types used. 3.2.1.1 TWI, 5V Tolerant, and SMBUS Pins Some Normal I/O pins offer TWI, 5V Tolerant, and SMBUS features. These features are only available when either of the TWI functions or the PWMAOD function in the PWMA are selected for these pins. 10 32099I–01/2012 AT32UC3L016/32/64 Refer to the ”TWI Pin Characteristics(1)” on page 778 for a description of the electrical properties of the TWI, 5V Tolerant, and SMBUS pins. 3.2.2 Peripheral Functions Each GPIO line can be assigned to one of several peripheral functions. The following table describes how the various peripheral functions are selected. The last listed function has priority in case multiple functions are enabled on the same pin. Table 3-2. 3.2.3 Function Description GPIO Controller Function multiplexing GPIO and GPIO peripheral selection A to H Nexus OCD AUX port connections OCD trace system aWire DATAOUT aWire output in two-pin mode JTAG port connections JTAG debug port Oscillators OSC0, OSC32 JTAG Port Connections If the JTAG is enabled, the JTAG will take control over a number of pins, irrespectively of the I/O Controller configuration. Table 3-3. 3.2.4 Peripheral Functions JTAG Pinout 48-pin Pin Name JTAG Pin 11 PA00 TCK 14 PA01 TMS 13 PA02 TDO 4 PA03 TDI Nexus OCD AUX Port Connections If the OCD trace system is enabled, the trace system will take control over a number of pins, irrespectively of the I/O Controller configuration. Two different OCD trace pin mappings are possible, depending on the configuration of the OCD AXS register. For details, see the AVR32 UC Technical Reference Manual. Table 3-4. Nexus OCD AUX Port Connections Pin AXS=1 AXS=0 EVTI_N PA05 PB08 MDO[5] PA10 PB00 MDO[4] PA18 PB04 MDO[3] PA17 PB05 MDO[2] PA16 PB03 11 32099I–01/2012 AT32UC3L016/32/64 Table 3-4. 3.2.5 Pin AXS=1 AXS=0 MDO[1] PA15 PB02 MDO[0] PA14 PB09 EVTO_N PA04 PA04 MCKO PA06 PB01 MSEO[1] PA07 PB11 MSEO[0] PA11 PB12 Oscillator Pinout The oscillators are not mapped to the normal GPIO functions and their muxings are controlled by registers in the System Control Interface (SCIF). Please refer to the SCIF chapter for more information about this. Table 3-5. 3.2.6 Nexus OCD AUX Port Connections Oscillator Pinout 48-pin Pin Oscillator Function 3 PA08 XIN0 46 PA10 XIN32 26 PA13 XIN32_2 2 PA09 XOUT0 47 PA12 XOUT32 25 PA20 XOUT32_2 Other Functions The functions listed in Table 3-6 are not mapped to the normal GPIO functions. The aWire DATA pin will only be active after the aWire is enabled. The aWire DATAOUT pin will only be active after the aWire is enabled and the 2_PIN_MODE command has been sent. The WAKE_N pin is always enabled. Please refer to Section 6.1.4 on page 40 for constraints on the WAKE_N pin. Table 3-6. Other Functions 48-pin Pin Function 27 PA11 WAKE_N 22 RESET_N aWire DATA 11 PA00 aWire DATAOUT 12 32099I–01/2012 AT32UC3L016/32/64 3.3 Signal Descriptions The following table gives details on signal name classified by peripheral. Table 3-7. Signal Descriptions List Signal Name Function Type Active Level Comments Analog Comparator Interface - ACIFB ACAN3 - ACAN0 Negative inputs for comparators "A" Analog ACAP3 - ACAP0 Positive inputs for comparators "A" Analog ACBN3 - ACBN0 Negative inputs for comparators "B" Analog ACBP3 - ACBP0 Positive inputs for comparators "B" Analog ACREFN Common negative reference Analog ADC Interface - ADCIFB AD8 - AD0 Analog Signal Analog ADP1 - ADP0 Drive Pin for resistive touch screen Output TRIGGER External trigger Input aWire - AW DATA aWire data I/O DATAOUT aWire data output for 2-pin mode I/O Capacitive Touch Module - CAT CSA16 - CSA0 Capacitive Sense A I/O CSB16 - CSB0 Capacitive Sense B I/O DIS Discharge current control Analog SMP SMP signal Output SYNC Synchronize signal VDIVEN Voltage divider enable Input Output External Interrupt Controller - EIC NMI Non-Maskable Interrupt Input EXTINT5 - EXTINT1 External interrupt Input Glue Logic Controller - GLOC IN7 - IN0 Inputs to lookup tables OUT1 - OUT0 Outputs from lookup tables Input Output JTAG module - JTAG TCK Test Clock Input TDI Test Data In Input TDO Test Data Out TMS Test Mode Select Output Input 13 32099I–01/2012 AT32UC3L016/32/64 Table 3-7. Signal Descriptions List Power Manager - PM RESET_N Reset Input Low Pulse Width Modulation Controller - PWMA PWMA35 - PWMA0 PWMA channel waveforms Output PWMAOD35 PWMAOD0 PWMA channel waveforms, open drain mode Output Not all channels support open drain mode System Control Interface - SCIF GCLK4 - GCLK0 Generic Clock Output Output RC32OUT RC32K output at startup Output XIN0 Crystal 0 Input Analog/ Digital XIN32 Crystal 32 Input (primary location) Analog/ Digital XIN32_2 Crystal 32 Input (secondary location) Analog/ Digital XOUT0 Crystal 0 Output Analog XOUT32 Crystal 32 Output (primary location) Analog XOUT32_2 Crystal 32 Output (secondary location) Analog Serial Peripheral Interface - SPI MISO Master In Slave Out I/O MOSI Master Out Slave In I/O NPCS3 - NPCS0 SPI Peripheral Chip Select I/O SCK Clock I/O Low Timer/Counter - TC0, TC1 A0 Channel 0 Line A I/O A1 Channel 1 Line A I/O A2 Channel 2 Line A I/O B0 Channel 0 Line B I/O B1 Channel 1 Line B I/O B2 Channel 2 Line B I/O CLK0 Channel 0 External Clock Input Input CLK1 Channel 1 External Clock Input Input CLK2 Channel 2 External Clock Input Input Two-wire Interface - TWIMS0, TWIMS1 TWALM SMBus SMBALERT I/O TWCK Two-wire Serial Clock I/O TWD Two-wire Serial Data I/O Low Universal Synchronous/Asynchronous Receiver/Transmitter - USART0, USART1, USART2, USART3 14 32099I–01/2012 AT32UC3L016/32/64 Table 3-7. Signal Descriptions List CLK Clock CTS Clear To Send RTS Request To Send RXD Receive Data Input TXD Transmit Data Output Note: I/O Input Low Output Low 1. ADCIFB: AD3 does not exist. Table 3-8. Signal Description List, Continued Signal Name Function Type Active Level Comments Power VDDCORE Core Power Supply / Voltage Regulator Output Power Input/Output 1.62V to 1.98V VDDIO I/O Power Supply Power Input 1.62V to 3.6V. VDDIO should always be equal to or lower than VDDIN. VDDANA Analog Power Supply Power Input 1.62V to 1.98V ADVREFP Analog Reference Voltage Power Input 1.62V to 1.98V VDDIN Voltage Regulator Input Power Input 1.62V to 3.6V (1) GNDANA Analog Ground Ground GND Ground Ground Auxiliary Port - AUX MCKO Trace Data Output Clock Output MDO5 - MDO0 Trace Data Output Output MSEO1 - MSEO0 Trace Frame Control Output EVTI_N Event In EVTO_N Event Out Input Low Output Low General Purpose I/O pin PA22 - PA00 Parallel I/O Controller I/O Port 0 I/O PB12 - PB00 Parallel I/O Controller I/O Port 1 I/O 1. See Section 6.1 on page 36 15 32099I–01/2012 AT32UC3L016/32/64 3.4 3.4.1 I/O Line Considerations JTAG Pins The JTAG is enabled if TCK is low while the RESET_N pin is released. The TCK, TMS, and TDI pins have pull-up resistors when JTAG is enabled. The TCK pin always has pull-up enabled during reset. The TDO pin is an output, driven at VDDIO, and has no pull-up resistor. The JTAG pins can be used as GPIO pins and multiplexed with peripherals when the JTAG is disabled. Please refer to Section 3.2.3 on page 11 for the JTAG port connections. 3.4.2 PA00 Note that PA00 is multiplexed with TCK. PA00 GPIO function must only be used as output in the application. 3.4.3 RESET_N Pin The RESET_N pin is a schmitt input and integrates a permanent pull-up resistor to VDDIN. As the product integrates a power-on reset detector, the RESET_N pin can be left unconnected in case no reset from the system needs to be applied to the product. The RESET_N pin is also used for the aWire debug protocol. When the pin is used for debugging, it must not be driven by external circuitry. 3.4.4 TWI Pins PA21/PB04/PB05 When these pins are used for TWI, the pins are open-drain outputs with slew-rate limitation and inputs with spike filtering. When used as GPIO pins or used for other peripherals, the pins have the same characteristics as other GPIO pins. Selected pins are also SMBus compliant (refer to Section 3.2 on page 9). As required by the SMBus specification, these pins provide no leakage path to ground when the AT32UC3L016/32/64 is powered down. This allows other devices on the SMBus to continue communicating even though the AT32UC3L016/32/64 is not powered. After reset a TWI function is selected on these pins instead of the GPIO. Please refer to the GPIO Module Configuration chapter for details. 3.4.5 TWI Pins PA05/PA07/PA17 When these pins are used for TWI, the pins are open-drain outputs with slew-rate limitation and inputs with spike filtering. When used as GPIO pins or used for other peripherals, the pins have the same characteristics as other GPIO pins. After reset a TWI function is selected on these pins instead of the GPIO. Please refer to the GPIO Module Configuration chapter for details. 3.4.6 GPIO Pins All the I/O lines integrate a pull-up resistor. Programming of this pull-up resistor is performed independently for each I/O line through the GPIO Controller. After reset, I/O lines default as inputs with pull-up resistors disabled, except PA00. PA20 selects SCIF-RC32OUT (GPIO Function F) as default enabled after reset. 3.4.7 High-Drive Pins The five pins PA02, PA06, PA08, PA09, and PB01 have high-drive output capabilities. Refer to Section 32. on page 770 for electrical characteristics. 16 32099I–01/2012 AT32UC3L016/32/64 3.4.8 RC32OUT Pin 3.4.8.1 Clock output at startup After power-up, the clock generated by the 32kHz RC oscillator (RC32K) will be output on PA20, even when the device is still reset by the Power-On Reset Circuitry. This clock can be used by the system to start other devices or to clock a switching regulator to rise the power supply voltage up to an acceptable value. The clock will be available on PA20, but will be disabled if one of the following conditions are true: • PA20 is configured to use a GPIO function other than F (SCIF-RC32OUT) • PA20 is configured as a General Purpose Input/Output (GPIO) • The bit FRC32 in the Power Manager PPCR register is written to zero (refer to the Power Manager chapter) The maximum amplitude of the clock signal will be defined by VDDIN. Once the RC32K output on PA20 is disabled it can never be enabled again. 3.4.8.2 3.4.9 XOUT32_2 function PA20 selects RC32OUT as default enabled after reset. This function is not automatically disabled when the user enables the XOUT32_2 function on PA20. This disturbs the oscillator and may result in the wrong frequency. To avoid this, RC32OUT must be disabled when XOUT32_2 is enabled. ADC Input Pins These pins are regular I/O pins powered from the VDDIO. However, when these pins are used for ADC inputs, the voltage applied to the pin must not exceed 1.98V. Internal circuitry ensures that the pin cannot be used as an analog input pin when the I/O drives to VDD. When the pins are not used for ADC inputs, the pins may be driven to the full I/O voltage range. 17 32099I–01/2012 AT32UC3L016/32/64 4. Processor and Architecture Rev: 2.1.0.0 This chapter gives an overview of the AVR32UC CPU. AVR32UC is an implementation of the AVR32 architecture. A summary of the programming model, instruction set, and MPU is presented. For further details, see the AVR32 Architecture Manual and the AVR32UC Technical Reference Manual. 4.1 Features • 32-bit load/store AVR32A RISC architecture – – – – – 15 general-purpose 32-bit registers 32-bit Stack Pointer, Program Counter and Link Register reside in register file Fully orthogonal instruction set Privileged and unprivileged modes enabling efficient and secure operating systems Innovative instruction set together with variable instruction length ensuring industry leading code density – DSP extension with saturating arithmetic, and a wide variety of multiply instructions • 3-stage pipeline allowing one instruction per clock cycle for most instructions – Byte, halfword, word, and double word memory access – Multiple interrupt priority levels • MPU allows for operating systems with memory protection • Secure State for supporting FlashVault technology 4.2 AVR32 Architecture AVR32 is a new, high-performance 32-bit RISC microprocessor architecture, designed for costsensitive embedded applications, with particular emphasis on low power consumption and high code density. In addition, the instruction set architecture has been tuned to allow a variety of microarchitectures, enabling the AVR32 to be implemented as low-, mid-, or high-performance processors. AVR32 extends the AVR family into the world of 32- and 64-bit applications. Through a quantitative approach, a large set of industry recognized benchmarks has been compiled and analyzed to achieve the best code density in its class. In addition to lowering the memory requirements, a compact code size also contributes to the core’s low power characteristics. The processor supports byte and halfword data types without penalty in code size and performance. Memory load and store operations are provided for byte, halfword, word, and double word data with automatic sign- or zero extension of halfword and byte data. The C-compiler is closely linked to the architecture and is able to exploit code optimization features, both for size and speed. In order to reduce code size to a minimum, some instructions have multiple addressing modes. As an example, instructions with immediates often have a compact format with a smaller immediate, and an extended format with a larger immediate. In this way, the compiler is able to use the format giving the smallest code size. Another feature of the instruction set is that frequently used instructions, like add, have a compact format with two operands as well as an extended format with three operands. The larger format increases performance, allowing an addition and a data move in the same instruction in a 18 32099I–01/2012 AT32UC3L016/32/64 single cycle. Load and store instructions have several different formats in order to reduce code size and speed up execution. The register file is organized as sixteen 32-bit registers and includes the Program Counter, the Link Register, and the Stack Pointer. In addition, register R12 is designed to hold return values from function calls and is used implicitly by some instructions. 4.3 The AVR32UC CPU The AVR32UC CPU targets low- and medium-performance applications, and provides an advanced On-Chip Debug (OCD) system, no caches, and a Memory Protection Unit (MPU). Java acceleration hardware is not implemented. AVR32UC provides three memory interfaces, one High Speed Bus master for instruction fetch, one High Speed Bus master for data access, and one High Speed Bus slave interface allowing other bus masters to access data RAMs internal to the CPU. Keeping data RAMs internal to the CPU allows fast access to the RAMs, reduces latency, and guarantees deterministic timing. Also, power consumption is reduced by not needing a full High Speed Bus access for memory accesses. A dedicated data RAM interface is provided for communicating with the internal data RAMs. A local bus interface is provided for connecting the CPU to device-specific high-speed systems, such as floating-point units and I/O controller ports. This local bus has to be enabled by writing a one to the LOCEN bit in the CPUCR system register. The local bus is able to transfer data between the CPU and the local bus slave in a single clock cycle. The local bus has a dedicated memory range allocated to it, and data transfers are performed using regular load and store instructions. Details on which devices that are mapped into the local bus space is given in the CPU Local Bus section in the Memories chapter. Figure 4-1 on page 20 displays the contents of AVR32UC. 19 32099I–01/2012 AT32UC3L016/32/64 OCD interface Reset interface Overview of the AVR32UC CPU Interrupt controller interface Figure 4-1. OCD system Power/ Reset control AVR32UC CPU pipeline MPU 4.3.1 High Speed Bus slave CPU Local Bus master CPU Local Bus High Speed Bus master High Speed Bus High Speed Bus High Speed Bus master High Speed Bus Data memory controller Instruction memory controller CPU RAM Pipeline Overview AVR32UC has three pipeline stages, Instruction Fetch (IF), Instruction Decode (ID), and Instruction Execute (EX). The EX stage is split into three parallel subsections, one arithmetic/logic (ALU) section, one multiply (MUL) section, and one load/store (LS) section. Instructions are issued and complete in order. Certain operations require several clock cycles to complete, and in this case, the instruction resides in the ID and EX stages for the required number of clock cycles. Since there is only three pipeline stages, no internal data forwarding is required, and no data dependencies can arise in the pipeline. Figure 4-2 on page 21 shows an overview of the AVR32UC pipeline stages. 20 32099I–01/2012 AT32UC3L016/32/64 Figure 4-2. The AVR32UC Pipeline MUL IF ID Prefetch unit Decode unit Regfile Read ALU LS 4.3.2 Multiply unit Regfile write ALU unit Load-store unit AVR32A Microarchitecture Compliance AVR32UC implements an AVR32A microarchitecture. The AVR32A microarchitecture is targeted at cost-sensitive, lower-end applications like smaller microcontrollers. This microarchitecture does not provide dedicated hardware registers for shadowing of register file registers in interrupt contexts. Additionally, it does not provide hardware registers for the return address registers and return status registers. Instead, all this information is stored on the system stack. This saves chip area at the expense of slower interrupt handling. 4.3.2.1 Interrupt Handling Upon interrupt initiation, registers R8-R12 are automatically pushed to the system stack. These registers are pushed regardless of the priority level of the pending interrupt. The return address and status register are also automatically pushed to stack. The interrupt handler can therefore use R8-R12 freely. Upon interrupt completion, the old R8-R12 registers and status register are restored, and execution continues at the return address stored popped from stack. The stack is also used to store the status register and return address for exceptions and scall. Executing the rete or rets instruction at the completion of an exception or system call will pop this status register and continue execution at the popped return address. 4.3.2.2 Java Support AVR32UC does not provide Java hardware acceleration. 4.3.2.3 Memory Protection The MPU allows the user to check all memory accesses for privilege violations. If an access is attempted to an illegal memory address, the access is aborted and an exception is taken. The MPU in AVR32UC is specified in the AVR32UC Technical Reference manual. 4.3.2.4 Unaligned Reference Handling AVR32UC does not support unaligned accesses, except for doubleword accesses. AVR32UC is able to perform word-aligned st.d and ld.d. Any other unaligned memory access will cause an 21 32099I–01/2012 AT32UC3L016/32/64 address exception. Doubleword-sized accesses with word-aligned pointers will automatically be performed as two word-sized accesses. The following table shows the instructions with support for unaligned addresses. All other instructions require aligned addresses. Table 4-1. 4.3.2.5 Instructions with Unaligned Reference Support Instruction Supported Alignment ld.d Word st.d Word Unimplemented Instructions The following instructions are unimplemented in AVR32UC, and will cause an Unimplemented Instruction Exception if executed: • All SIMD instructions • All coprocessor instructions if no coprocessors are present • retj, incjosp, popjc, pushjc • tlbr, tlbs, tlbw • cache 4.3.2.6 CPU and Architecture Revision Three major revisions of the AVR32UC CPU currently exist. The device described in this datasheet uses CPU revision 3. The Architecture Revision field in the CONFIG0 system register identifies which architecture revision is implemented in a specific device. AVR32UC CPU revision 3 is fully backward-compatible with revisions 1 and 2, ie. code compiled for revision 1 or 2 is binary-compatible with revision 3 CPUs. 22 32099I–01/2012 AT32UC3L016/32/64 4.4 4.4.1 Programming Model Register File Configuration The AVR32UC register file is shown below. Figure 4-3. The AVR32UC Register File Application Supervisor INT0 Bit 31 Bit 31 Bit 31 Bit 0 Bit 0 INT1 Bit 0 INT2 Bit 31 Bit 0 INT3 Bit 31 Bit 0 Bit 31 Bit 0 Exception NMI Bit 31 Bit 31 Bit 0 Secure Bit 0 Bit 31 Bit 0 PC LR SP_APP R12 R11 R10 R9 R8 INT0PC R7 INT1PC R6 FINTPC R5 SMPC R4 R3 R2 R1 R0 PC LR SP_SYS R12 R11 R10 R9 R8 INT0PC R7 INT1PC R6 FINTPC R5 SMPC R4 R3 R2 R1 R0 PC LR SP_SYS R12 R11 R10 R9 R8 INT0PC R7 INT1PC R6 FINTPC R5 SMPC R4 R3 R2 R1 R0 PC LR SP_SYS R12 R11 R10 R9 R8 INT0PC R7 INT1PC R6 FINTPC R5 SMPC R4 R3 R2 R1 R0 PC LR SP_SYS R12 R11 R10 R9 R8 INT0PC R7 INT1PC R6 FINTPC R5 SMPC R4 R3 R2 R1 R0 PC LR SP_SYS R12 R11 R10 R9 R8 INT0PC R7 INT1PC R6 FINTPC R5 SMPC R4 R3 R2 R1 R0 PC LR SP_SYS R12 R11 R10 R9 R8 INT0PC R7 INT1PC R6 FINTPC R5 SMPC R4 R3 R2 R1 R0 PC LR SP_SYS R12 R11 R10 R9 R8 INT0PC R7 INT1PC R6 FINTPC R5 SMPC R4 R3 R2 R1 R0 PC LR SP_SEC R12 R11 R10 R9 R8 INT0PC R7 INT1PC R6 FINTPC R5 SMPC R4 R3 R2 R1 R0 SR SR SR SR SR SR SR SR SR SS_STATUS SS_ADRF SS_ADRR SS_ADR0 SS_ADR1 SS_SP_SYS SS_SP_APP SS_RAR SS_RSR 4.4.2 Status Register Configuration The Status Register (SR) is split into two halfwords, one upper and one lower, see Figure 4-4 and Figure 4-5. The lower word contains the C, Z, N, V, and Q condition code flags and the R, T, and L bits, while the upper halfword contains information about the mode and state the processor executes in. Refer to the AVR32 Architecture Manual for details. Figure 4-4. The Status Register High Halfword Bit 31 Bit 16 SS LC 1 - - DM D - M2 M1 M0 EM I3M I2M FE I1M I0M GM 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 1 Bit nam e Initial value G lobal Interrupt M ask Interrupt Level 0 M ask Interrupt Level 1 M ask Interrupt Level 2 M ask Interrupt Level 3 M ask Exception M ask M ode Bit 0 M ode Bit 1 M ode Bit 2 Reserved Debug State Debug State M ask Reserved Secure State 23 32099I–01/2012 AT32UC3L016/32/64 Figure 4-5. The Status Register Low Halfword Bit 15 Bit 0 - T - - - - - - - - L Q V N Z C Bit name 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Initial value Carry Zero Sign Overflow Saturation Lock Reserved Scratch Reserved 4.4.3 Processor States 4.4.3.1 Normal RISC State The AVR32 processor supports several different execution contexts as shown in Table 4-2. Table 4-2. Overview of Execution Modes, their Priorities and Privilege Levels. Priority Mode Security Description 1 Non Maskable Interrupt Privileged Non Maskable high priority interrupt mode 2 Exception Privileged Execute exceptions 3 Interrupt 3 Privileged General purpose interrupt mode 4 Interrupt 2 Privileged General purpose interrupt mode 5 Interrupt 1 Privileged General purpose interrupt mode 6 Interrupt 0 Privileged General purpose interrupt mode N/A Supervisor Privileged Runs supervisor calls N/A Application Unprivileged Normal program execution mode Mode changes can be made under software control, or can be caused by external interrupts or exception processing. A mode can be interrupted by a higher priority mode, but never by one with lower priority. Nested exceptions can be supported with a minimal software overhead. When running an operating system on the AVR32, user processes will typically execute in the application mode. The programs executed in this mode are restricted from executing certain instructions. Furthermore, most system registers together with the upper halfword of the status register cannot be accessed. Protected memory areas are also not available. All other operating modes are privileged and are collectively called System Modes. They have full access to all privileged and unprivileged resources. After a reset, the processor will be in supervisor mode. 4.4.3.2 Debug State The AVR32 can be set in a debug state, which allows implementation of software monitor routines that can read out and alter system information for use during application development. This implies that all system and application registers, including the status registers and program counters, are accessible in debug state. The privileged instructions are also available. All interrupt levels are by default disabled when debug state is entered, but they can individually be switched on by the monitor routine by clearing the respective mask bit in the status register. 24 32099I–01/2012 AT32UC3L016/32/64 Debug state can be entered as described in the AVR32UC Technical Reference Manual. Debug state is exited by the retd instruction. 4.4.3.3 4.4.4 Secure State The AVR32 can be set in a secure state, that allows a part of the code to execute in a state with higher security levels. The rest of the code can not access resources reserved for this secure code. Secure State is used to implement FlashVault technology. Refer to the AVR32UC Technical Reference Manual for details. System Registers The system registers are placed outside of the virtual memory space, and are only accessible using the privileged mfsr and mtsr instructions. The table below lists the system registers specified in the AVR32 architecture, some of which are unused in AVR32UC. The programmer is responsible for maintaining correct sequencing of any instructions following a mtsr instruction. For detail on the system registers, refer to the AVR32UC Technical Reference Manual. Table 4-3. System Registers Reg # Address Name Function 0 0 SR Status Register 1 4 EVBA Exception Vector Base Address 2 8 ACBA Application Call Base Address 3 12 CPUCR CPU Control Register 4 16 ECR Exception Cause Register 5 20 RSR_SUP Unused in AVR32UC 6 24 RSR_INT0 Unused in AVR32UC 7 28 RSR_INT1 Unused in AVR32UC 8 32 RSR_INT2 Unused in AVR32UC 9 36 RSR_INT3 Unused in AVR32UC 10 40 RSR_EX Unused in AVR32UC 11 44 RSR_NMI Unused in AVR32UC 12 48 RSR_DBG Return Status Register for Debug mode 13 52 RAR_SUP Unused in AVR32UC 14 56 RAR_INT0 Unused in AVR32UC 15 60 RAR_INT1 Unused in AVR32UC 16 64 RAR_INT2 Unused in AVR32UC 17 68 RAR_INT3 Unused in AVR32UC 18 72 RAR_EX Unused in AVR32UC 19 76 RAR_NMI Unused in AVR32UC 20 80 RAR_DBG Return Address Register for Debug mode 21 84 JECR Unused in AVR32UC 22 88 JOSP Unused in AVR32UC 23 92 JAVA_LV0 Unused in AVR32UC 25 32099I–01/2012 AT32UC3L016/32/64 Table 4-3. System Registers (Continued) Reg # Address Name Function 24 96 JAVA_LV1 Unused in AVR32UC 25 100 JAVA_LV2 Unused in AVR32UC 26 104 JAVA_LV3 Unused in AVR32UC 27 108 JAVA_LV4 Unused in AVR32UC 28 112 JAVA_LV5 Unused in AVR32UC 29 116 JAVA_LV6 Unused in AVR32UC 30 120 JAVA_LV7 Unused in AVR32UC 31 124 JTBA Unused in AVR32UC 32 128 JBCR Unused in AVR32UC 33-63 132-252 Reserved Reserved for future use 64 256 CONFIG0 Configuration register 0 65 260 CONFIG1 Configuration register 1 66 264 COUNT Cycle Counter register 67 268 COMPARE Compare register 68 272 TLBEHI Unused in AVR32UC 69 276 TLBELO Unused in AVR32UC 70 280 PTBR Unused in AVR32UC 71 284 TLBEAR Unused in AVR32UC 72 288 MMUCR Unused in AVR32UC 73 292 TLBARLO Unused in AVR32UC 74 296 TLBARHI Unused in AVR32UC 75 300 PCCNT Unused in AVR32UC 76 304 PCNT0 Unused in AVR32UC 77 308 PCNT1 Unused in AVR32UC 78 312 PCCR Unused in AVR32UC 79 316 BEAR Bus Error Address Register 80 320 MPUAR0 MPU Address Register region 0 81 324 MPUAR1 MPU Address Register region 1 82 328 MPUAR2 MPU Address Register region 2 83 332 MPUAR3 MPU Address Register region 3 84 336 MPUAR4 MPU Address Register region 4 85 340 MPUAR5 MPU Address Register region 5 86 344 MPUAR6 MPU Address Register region 6 87 348 MPUAR7 MPU Address Register region 7 88 352 MPUPSR0 MPU Privilege Select Register region 0 89 356 MPUPSR1 MPU Privilege Select Register region 1 26 32099I–01/2012 AT32UC3L016/32/64 Table 4-3. 4.5 System Registers (Continued) Reg # Address Name Function 90 360 MPUPSR2 MPU Privilege Select Register region 2 91 364 MPUPSR3 MPU Privilege Select Register region 3 92 368 MPUPSR4 MPU Privilege Select Register region 4 93 372 MPUPSR5 MPU Privilege Select Register region 5 94 376 MPUPSR6 MPU Privilege Select Register region 6 95 380 MPUPSR7 MPU Privilege Select Register region 7 96 384 MPUCRA Unused in this version of AVR32UC 97 388 MPUCRB Unused in this version of AVR32UC 98 392 MPUBRA Unused in this version of AVR32UC 99 396 MPUBRB Unused in this version of AVR32UC 100 400 MPUAPRA MPU Access Permission Register A 101 404 MPUAPRB MPU Access Permission Register B 102 408 MPUCR MPU Control Register 103 412 SS_STATUS Secure State Status Register 104 416 SS_ADRF Secure State Address Flash Register 105 420 SS_ADRR Secure State Address RAM Register 106 424 SS_ADR0 Secure State Address 0 Register 107 428 SS_ADR1 Secure State Address 1 Register 108 432 SS_SP_SYS Secure State Stack Pointer System Register 109 436 SS_SP_APP Secure State Stack Pointer Application Register 110 440 SS_RAR Secure State Return Address Register 111 444 SS_RSR Secure State Return Status Register 112-191 448-764 Reserved Reserved for future use 192-255 768-1020 IMPL IMPLEMENTATION DEFINED Exceptions and Interrupts In the AVR32 architecture, events are used as a common term for exceptions and interrupts. AVR32UC incorporates a powerful event handling scheme. The different event sources, like Illegal Op-code and interrupt requests, have different priority levels, ensuring a well-defined behavior when multiple events are received simultaneously. Additionally, pending events of a higher priority class may preempt handling of ongoing events of a lower priority class. When an event occurs, the execution of the instruction stream is halted, and execution is passed to an event handler at an address specified in Table 4-4 on page 31. Most of the handlers are placed sequentially in the code space starting at the address specified by EVBA, with four bytes between each handler. This gives ample space for a jump instruction to be placed there, jumping to the event routine itself. A few critical handlers have larger spacing between them, allowing the entire event routine to be placed directly at the address specified by the EVBA-relative offset generated by hardware. All interrupt sources have autovectored interrupt service routine (ISR) addresses. This allows the interrupt controller to directly specify the ISR address as an address 27 32099I–01/2012 AT32UC3L016/32/64 relative to EVBA. The autovector offset has 14 address bits, giving an offset of maximum 16384 bytes. The target address of the event handler is calculated as (EVBA | event_handler_offset), not (EVBA + event_handler_offset), so EVBA and exception code segments must be set up appropriately. The same mechanisms are used to service all different types of events, including interrupt requests, yielding a uniform event handling scheme. An interrupt controller does the priority handling of the interrupts and provides the autovector offset to the CPU. 4.5.1 System Stack Issues Event handling in AVR32UC uses the system stack pointed to by the system stack pointer, SP_SYS, for pushing and popping R8-R12, LR, status register, and return address. Since event code may be timing-critical, SP_SYS should point to memory addresses in the IRAM section, since the timing of accesses to this memory section is both fast and deterministic. The user must also make sure that the system stack is large enough so that any event is able to push the required registers to stack. If the system stack is full, and an event occurs, the system will enter an UNDEFINED state. 4.5.2 Exceptions and Interrupt Requests When an event other than scall or debug request is received by the core, the following actions are performed atomically: 1. The pending event will not be accepted if it is masked. The I3M, I2M, I1M, I0M, EM, and GM bits in the Status Register are used to mask different events. Not all events can be masked. A few critical events (NMI, Unrecoverable Exception, TLB Multiple Hit, and Bus Error) can not be masked. When an event is accepted, hardware automatically sets the mask bits corresponding to all sources with equal or lower priority. This inhibits acceptance of other events of the same or lower priority, except for the critical events listed above. Software may choose to clear some or all of these bits after saving the necessary state if other priority schemes are desired. It is the event source’s responsability to ensure that their events are left pending until accepted by the CPU. 2. When a request is accepted, the Status Register and Program Counter of the current context is stored to the system stack. If the event is an INT0, INT1, INT2, or INT3, registers R8-R12 and LR are also automatically stored to stack. Storing the Status Register ensures that the core is returned to the previous execution mode when the current event handling is completed. When exceptions occur, both the EM and GM bits are set, and the application may manually enable nested exceptions if desired by clearing the appropriate bit. Each exception handler has a dedicated handler address, and this address uniquely identifies the exception source. 3. The Mode bits are set to reflect the priority of the accepted event, and the correct register file bank is selected. The address of the event handler, as shown in Table 4-4 on page 31, is loaded into the Program Counter. The execution of the event handler routine then continues from the effective address calculated. The rete instruction signals the end of the event. When encountered, the Return Status Register and Return Address Register are popped from the system stack and restored to the Status Register and Program Counter. If the rete instruction returns from INT0, INT1, INT2, or INT3, registers R8-R12 and LR are also popped from the system stack. The restored Status Register contains information allowing the core to resume operation in the previous execution mode. This concludes the event handling. 28 32099I–01/2012 AT32UC3L016/32/64 4.5.3 Supervisor Calls The AVR32 instruction set provides a supervisor mode call instruction. The scall instruction is designed so that privileged routines can be called from any context. This facilitates sharing of code between different execution modes. The scall mechanism is designed so that a minimal execution cycle overhead is experienced when performing supervisor routine calls from timecritical event handlers. The scall instruction behaves differently depending on which mode it is called from. The behaviour is detailed in the instruction set reference. In order to allow the scall routine to return to the correct context, a return from supervisor call instruction, rets, is implemented. In the AVR32UC CPU, scall and rets uses the system stack to store the return address and the status register. 4.5.4 Debug Requests The AVR32 architecture defines a dedicated Debug mode. When a debug request is received by the core, Debug mode is entered. Entry into Debug mode can be masked by the DM bit in the status register. Upon entry into Debug mode, hardware sets the SR.D bit and jumps to the Debug Exception handler. By default, Debug mode executes in the exception context, but with dedicated Return Address Register and Return Status Register. These dedicated registers remove the need for storing this data to the system stack, thereby improving debuggability. The Mode bits in the Status Register can freely be manipulated in Debug mode, to observe registers in all contexts, while retaining full privileges. Debug mode is exited by executing the retd instruction. This returns to the previous context. 4.5.5 Entry Points for Events Several different event handler entry points exist. In AVR32UC, the reset address is 0x80000000. This places the reset address in the boot flash memory area. TLB miss exceptions and scall have a dedicated space relative to EVBA where their event handler can be placed. This speeds up execution by removing the need for a jump instruction placed at the program address jumped to by the event hardware. All other exceptions have a dedicated event routine entry point located relative to EVBA. The handler routine address identifies the exception source directly. AVR32UC uses the ITLB and DTLB protection exceptions to signal a MPU protection violation. ITLB and DTLB miss exceptions are used to signal that an access address did not map to any of the entries in the MPU. TLB multiple hit exception indicates that an access address did map to multiple TLB entries, signalling an error. All interrupt requests have entry points located at an offset relative to EVBA. This autovector offset is specified by an interrupt controller. The programmer must make sure that none of the autovector offsets interfere with the placement of other code. The autovector offset has 14 address bits, giving an offset of maximum 16384 bytes. Special considerations should be made when loading EVBA with a pointer. Due to security considerations, the event handlers should be located in non-writeable flash memory, or optionally in a privileged memory protection region if an MPU is present. If several events occur on the same instruction, they are handled in a prioritized way. The priority ordering is presented in Table 4-4 on page 31. If events occur on several instructions at different locations in the pipeline, the events on the oldest instruction are always handled before any events on any younger instruction, even if the younger instruction has events of higher priority 29 32099I–01/2012 AT32UC3L016/32/64 than the oldest instruction. An instruction B is younger than an instruction A if it was sent down the pipeline later than A. The addresses and priority of simultaneous events are shown in Table 4-4 on page 31. Some of the exceptions are unused in AVR32UC since it has no MMU, coprocessor interface, or floatingpoint unit. 30 32099I–01/2012 AT32UC3L016/32/64 Table 4-4. Priority and Handler Addresses for Events Priority Handler Address Name Event source Stored Return Address 1 0x80000000 Reset External input Undefined 2 Provided by OCD system OCD Stop CPU OCD system First non-completed instruction 3 EVBA+0x00 Unrecoverable exception Internal PC of offending instruction 4 EVBA+0x04 TLB multiple hit MPU PC of offending instruction 5 EVBA+0x08 Bus error data fetch Data bus First non-completed instruction 6 EVBA+0x0C Bus error instruction fetch Data bus First non-completed instruction 7 EVBA+0x10 NMI External input First non-completed instruction 8 Autovectored Interrupt 3 request External input First non-completed instruction 9 Autovectored Interrupt 2 request External input First non-completed instruction 10 Autovectored Interrupt 1 request External input First non-completed instruction 11 Autovectored Interrupt 0 request External input First non-completed instruction 12 EVBA+0x14 Instruction Address CPU PC of offending instruction 13 EVBA+0x50 ITLB Miss MPU PC of offending instruction 14 EVBA+0x18 ITLB Protection MPU PC of offending instruction 15 EVBA+0x1C Breakpoint OCD system First non-completed instruction 16 EVBA+0x20 Illegal Opcode Instruction PC of offending instruction 17 EVBA+0x24 Unimplemented instruction Instruction PC of offending instruction 18 EVBA+0x28 Privilege violation Instruction PC of offending instruction 19 EVBA+0x2C Floating-point UNUSED 20 EVBA+0x30 Coprocessor absent Instruction PC of offending instruction 21 EVBA+0x100 Supervisor call Instruction PC(Supervisor Call) +2 22 EVBA+0x34 Data Address (Read) CPU PC of offending instruction 23 EVBA+0x38 Data Address (Write) CPU PC of offending instruction 24 EVBA+0x60 DTLB Miss (Read) MPU PC of offending instruction 25 EVBA+0x70 DTLB Miss (Write) MPU PC of offending instruction 26 EVBA+0x3C DTLB Protection (Read) MPU PC of offending instruction 27 EVBA+0x40 DTLB Protection (Write) MPU PC of offending instruction 28 EVBA+0x44 DTLB Modified UNUSED 31 32099I–01/2012 AT32UC3L016/32/64 5. Memories 5.1 Embedded Memories • Internal high-speed flash – 64Kbytes (AT32UC3L064) – 32Kbytes (AT32UC3L032) – 16Kbytes (AT32UC3L016) • 0 wait state access at up to 25mhz in worst case conditions • 1 wait state access at up to 50mhz in worst case conditions • Pipelined flash architecture, allowing burst reads from sequential Flash locations, hiding penalty of 1 wait state access • Pipelined flash architecture typically reduces the cycle penalty of 1 wait state operation to only 8% compared to 0 wait state operation • 100 000 write cycles, 15-year data retention capability • Sector lock capabilities, bootloader protection, security bit • 32 fuses, erased during chip erase • User page for data to be preserved during chip erase • Internal high-speed SRAM, single-cycle access at full speed – 16Kbytes (AT32UC3L064, AT32UC3L032) – 8Kbytes (AT32UC3L016) 5.2 Physical Memory Map The system bus is implemented as a bus matrix. All system bus addresses are fixed, and they are never remapped in any way, not even in boot. Note that AVR32 UC CPU uses unsegmented translation, as described in the AVR32 Architecture Manual. The 32-bit physical address space is mapped as follows: Table 5-1. AT32UC3L016/32/64 Physical Memory Map Device Table 5-2. Start Address Size AT32UC3L064 AT32UC3L032 AT32UC3L016 Embedded SRAM 0x00000000 16Kbytes 16Kbytes 8Kbytes Embedded Flash 0x80000000 64Kbytes 32Kbytes 16Kbytes SAU Channels 0x90000000 256 bytes 256 bytes 256 bytes HSB-PB Bridge B 0xFFFE0000 64Kbytes 64Kbytes 64Kbytes HSB-PB Bridge A 0xFFFF0000 64Kbytes 64Kbytes 64Kbytes Flash Memory Parameters Part Number Flash Size (FLASH_PW) Number of pages (FLASH_P) Page size (FLASH_W) AT32UC3L064 64Kbytes 256 256 bytes AT32UC3L032 32Kbytes 128 256 bytes AT32UC3L016 16Kbytes 64 256 bytes 32 32099I–01/2012 AT32UC3L016/32/64 5.3 Peripheral Address Map Table 5-3. Peripheral Address Mapping Address Peripheral Name 0xFFFE0000 FLASHCDW Flash Controller - FLASHCDW 0xFFFE0400 HMATRIX HSB Matrix - HMATRIX 0xFFFE0800 SAU Secure Access Unit - SAU 0xFFFF0000 PDCA Peripheral DMA Controller - PDCA INTC Interrupt controller - INTC 0xFFFF1000 0xFFFF1400 PM Power Manager - PM 0xFFFF1800 SCIF System Control Interface - SCIF AST Asynchronous Timer - AST WDT Watchdog Timer - WDT EIC External Interrupt Controller - EIC 0xFFFF1C00 0xFFFF2000 0xFFFF2400 0xFFFF2800 FREQM Frequency Meter - FREQM 0xFFFF2C00 GPIO 0xFFFF3000 General Purpose Input/Output Controller - GPIO USART0 Universal Synchronous/Asynchronous Receiver/Transmitter - USART0 USART1 Universal Synchronous/Asynchronous Receiver/Transmitter - USART1 USART2 Universal Synchronous/Asynchronous Receiver/Transmitter - USART2 USART3 Universal Synchronous/Asynchronous Receiver/Transmitter - USART3 0xFFFF3400 0xFFFF3800 0xFFFF3C00 0xFFFF4000 SPI Serial Peripheral Interface - SPI 0xFFFF4400 TWIM0 Two-wire Master Interface - TWIM0 33 32099I–01/2012 AT32UC3L016/32/64 Table 5-3. Peripheral Address Mapping 0xFFFF4800 TWIM1 Two-wire Master Interface - TWIM1 TWIS0 Two-wire Slave Interface - TWIS0 TWIS1 Two-wire Slave Interface - TWIS1 PWMA Pulse Width Modulation Controller - PWMA 0xFFFF4C00 0xFFFF5000 0xFFFF5400 0xFFFF5800 TC0 Timer/Counter - TC0 TC1 Timer/Counter - TC1 0xFFFF5C00 0xFFFF6000 ADCIFB ADC Interface - ADCIFB 0xFFFF6400 ACIFB Analog Comparator Interface - ACIFB 0xFFFF6800 CAT Capacitive Touch Module - CAT 0xFFFF6C00 GLOC Glue Logic Controller - GLOC 0xFFFF7000 AW 5.4 aWire - AW CPU Local Bus Mapping Some of the registers in the GPIO module are mapped onto the CPU local bus, in addition to being mapped on the Peripheral Bus. These registers can therefore be reached both by accesses on the Peripheral Bus, and by accesses on the local bus. Mapping these registers on the local bus allows cycle-deterministic toggling of GPIO pins since the CPU and GPIO are the only modules connected to this bus. Also, since the local bus runs at CPU speed, one write or read operation can be performed per clock cycle to the local busmapped GPIO registers. 34 32099I–01/2012 AT32UC3L016/32/64 The following GPIO registers are mapped on the local bus: Table 5-4. Local Bus Mapped GPIO Registers Port Register Mode Local Bus Address Access 0 Output Driver Enable Register (ODER) WRITE 0x40000040 Write-only SET 0x40000044 Write-only CLEAR 0x40000048 Write-only TOGGLE 0x4000004C Write-only WRITE 0x40000050 Write-only SET 0x40000054 Write-only CLEAR 0x40000058 Write-only TOGGLE 0x4000005C Write-only Pin Value Register (PVR) - 0x40000060 Read-only Output Driver Enable Register (ODER) WRITE 0x40000140 Write-only SET 0x40000144 Write-only CLEAR 0x40000148 Write-only TOGGLE 0x4000014C Write-only WRITE 0x40000150 Write-only SET 0x40000154 Write-only CLEAR 0x40000158 Write-only TOGGLE 0x4000015C Write-only - 0x40000160 Read-only Output Value Register (OVR) 1 Output Value Register (OVR) Pin Value Register (PVR) 35 32099I–01/2012 AT32UC3L016/32/64 6. Supply and Startup Considerations 6.1 6.1.1 Supply Considerations Power Supplies The AT32UC3L016/32/64 has several types of power supply pins: •VDDIO: Powers I/O lines. Voltage is 1.8 to 3.3V nominal. •VDDIN: Powers I/O lines and the internal regulator. Voltage is 1.8 to 3.3V nominal. •VDDANA: Powers the ADC. Voltage is 1.8V nominal. •VDDCORE: Powers the core, memories, and peripherals. Voltage is 1.8V nominal. The ground pins GND are common to VDDCORE, VDDIO, and VDDIN. The ground pin for VDDANA is GNDANA. When VDDCORE is not connected to VDDIN, the VDDIN voltage must be higher than 1.98V. Refer to Section 32. on page 770 for power consumption on the various supply pins. For decoupling recommendations for the different power supplies, please refer to the schematic checklist. 6.1.2 Voltage Regulator The AT32UC3L016/32/64 embeds a voltage regulator that converts from 3.3V nominal to 1.8V with a load of up to 60mA. The regulator supplies the output voltage on VDDCORE. The regulator may only be used to drive internal circuitry in the device. VDDCORE should be externally connected to the 1.8V domains. See Section 6.1.3 for regulator connection figures. Adequate output supply decoupling is mandatory for VDDCORE to reduce ripple and avoid oscillations. The best way to achieve this is to use two capacitors in parallell between VDDCORE and GND as close to the device as possible. Please refer to Section 32.8.1 on page 784 for decoupling capacitors values and regulator characteristics. Figure 6-1. Supply Decoupling 3.3V VDDIN C IN3 CIN2 CIN1 1.8V VDDCORE COUT2 6.1.3 1.8V Regulator COUT1 Regulator Connection The AT32UC3L016/32/64 supports three power supply configurations: • 3.3V single supply mode • 1.8V single supply mode • 3.3V supply mode, with 1.8V regulated I/O lines 36 32099I–01/2012 AT32UC3L016/32/64 6.1.3.1 3.3V Single Supply Mode In 3.3V single supply mode the internal regulator is connected to the 3.3V source (VDDIN pin) and its output feeds VDDCORE. Figure 6-2 shows the power schematics to be used for 3.3V single supply mode. All I/O lines will be powered by the same power (VDDIN=VDDIO). Figure 6-2. 3.3V Single Supply Mode + 1.98-3.6V - VDDIN VDDIO I/O Pins I/O Pins VDDCORE OSC32K RC32K AST Wake POR33 SM33 Linear regulator VDDANA GND ADC CPU, Peripherals, Memories, SCIF, BOD, RCSYS, DFLL GNDANA 37 32099I–01/2012 AT32UC3L016/32/64 6.1.3.2 1.8V Single Supply Mode In 1.8V single supply mode the internal regulator is not used, and VDDIO and VDDCORE are powered by a single 1.8V supply as shown in Figure 6-3. All I/O lines will be powered by the same power (VDDIN = VDDIO = VDDCORE). Figure 6-3. 1.8V Single Supply Mode. + 1.62-1.98V - VDDIN VDDIO I/O Pins I/O Pins VDDCORE Linear Regulator VDDANA OSC32K RC32K AST Wake POR33 SM33 ADC GNDANA GND CPU, Peripherals, Memories, SCIF, BOD, RCSYS, DFLL 38 32099I–01/2012 AT32UC3L016/32/64 6.1.3.3 3.3V Supply Mode with 1.8V Regulated I/O Lines In this mode, the internal regulator is connected to the 3.3V source and its output is connected to both VDDCORE and VDDIO as shown in Figure 6-4. This configuration is required in order to use Shutdown mode. Figure 6-4. 3.3V Supply Mode with 1.8V Regulated I/O Lines 1.98-3.6V + - VDDIN VDDIO I/O Pins VDDCORE I/O Pins Linear Regulator VDDANA OSC32K RC32K AST Wake POR33 SM33 ADC GNDANA GND CPU, Peripherals, Memories, SCIF, BOD, RCSYS, DFLL In this mode, some I/O lines are powered by VDDIN while other I/O lines are powered by VDDIO. Refer to Section 3.2 on page 9 for description of power supply for each I/O line. Refer to the Power Manager chapter for a description of what parts of the system are powered in Shutdown mode. Important note: As the regulator has a maximum output current of 60mA, this mode can only be used in applications where the maximum I/O current is known and compatible with the core and peripheral power consumption. Typically, great care must be used to ensure that only a few I/O lines are toggling at the same time and drive very small loads. 39 32099I–01/2012 AT32UC3L016/32/64 6.1.4 Power-up Sequence 6.1.4.1 Maximum Rise Rate To avoid risk of latch-up, the rise rate of the power supplies must not exceed the values described in Table 32-3 on page 771. Recommended order for power supplies is also described in this chapter. 6.1.4.2 Minimum Rise Rate The integrated Power-on Reset (POR33) circuitry monitoring the VDDIN powering supply requires a minimum rise rate for the VDDIN power supply. See Table 32-3 on page 771 for the minimum rise rate value. If the application can not ensure that the minimum rise rate condition for the VDDIN power supply is met, one of the following configurations can be used: • A logic “0” value is applied during power-up on pin PA11 until VDDIN rises above 1.2V. • A logic “0” value is applied during power-up on pin RESET_N until VDDIN rises above 1.2V. 6.2 Startup Considerations This chapter summarizes the boot sequence of the AT32UC3L016/32/64. The behavior after power-up is controlled by the Power Manager. For specific details, refer to the Power Manager chapter. 6.2.1 Starting of Clocks After power-up, the device will be held in a reset state by the Power-on Reset (POR18 and POR33) circuitry for a short time to allow the power to stabilize throughout the device. After reset, the device will use the System RC Oscillator (RCSYS) as clock source. Please refer to Table 32-17 on page 783 for the frequency for this oscillator. On system start-up, the DFLL is disabled. All clocks to all modules are running. No clocks have a divided frequency; all parts of the system receive a clock with the same frequency as the System RC Oscillator. When powering up the device, there may be a delay before the voltage has stabilized, depending on the rise time of the supply used. The CPU can start executing code as soon as the supply is above the POR18 and POR33 thresholds, and before the supply is stable. Before switching to a high-speed clock source, the user should use the BOD to make sure the VDDCORE is above the minimum level (1.62V). 6.2.2 Fetching of Initial Instructions After reset has been released, the AVR32 UC CPU starts fetching instructions from the reset address, which is 0x80000000. This address points to the first address in the internal flash. The code read from the internal flash is free to configure the clock system and clock sources . Please refer to the Power Manager and SCIF chapters for details. 40 32099I–01/2012 AT32UC3L016/32/64 7. Peripheral DMA Controller (PDCA) Rev: 1.2.2.1 7.1 Features • • • • • • 7.2 Multiple channels Generates transfers between memories and peripherals such as USART and SPI Two address pointers/counters per channel allowing double buffering Performance monitors to measure average and maximum transfer latency Optional synchronizing of data transfers with extenal peripheral events Ring buffer functionality Overview The Peripheral DMA Controller (PDCA) transfers data between on-chip peripheral modules such as USART, SPI and memories (those memories may be on- and off-chip memories). Using the PDCA avoids CPU intervention for data transfers, improving the performance of the microcontroller. The PDCA can transfer data from memory to a peripheral or from a peripheral to memory. The PDCA consists of multiple DMA channels. Each channel has: • A Peripheral Select Register • A 32-bit memory pointer • A 16-bit transfer counter • A 32-bit memory pointer reload value • A 16-bit transfer counter reload value The PDCA communicates with the peripheral modules over a set of handshake interfaces. The peripheral signals the PDCA when it is ready to receive or transmit data. The PDCA acknowledges the request when the transmission has started. When a transmit buffer is empty or a receive buffer is full, an optional interrupt request can be generated. 41 32099I–01/2012 AT32UC3L016/32/64 7.3 Block Diagram Figure 7-1. PDCA Block Diagram Peripheral 0 Memory HSB to PB Bridge HSB Peripheral Bus HSB High Speed Bus Matrix HSB Interrupt Controller IRQ Peripheral 2 ... Peripheral DMA Controller (PDCA) Peripheral 1 Peripheral (n-1) Handshake Interfaces 7.4 Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 7.4.1 Power Management If the CPU enters a sleep mode that disables the PDCA clocks, the PDCA will stop functioning and resume operation after the system wakes up from sleep mode. 7.4.2 Clocks The PDCA has two bus clocks connected: One High Speed Bus clock (CLK_PDCA_HSB) and one Peripheral Bus clock (CLK_PDCA_PB). These clocks are generated by the Power Manager. Both clocks are enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the PDCA before disabling the clocks, to avoid freezing the PDCA in an undefined state. 7.4.3 Interrupts The PDCA interrupt request lines are connected to the interrupt controller. Using the PDCA interrupts requires the interrupt controller to be programmed first. 42 32099I–01/2012 AT32UC3L016/32/64 7.4.4 7.5 7.5.1 Peripheral Events The PDCA peripheral events are connected via the Peripheral Event System. Refer to the Peripheral Event System chapter for details. Functional Description Basic Operation The PDCA consists of multiple independent PDCA channels, each capable of handling DMA requests in parallel. Each PDCA channels contains a set of configuration registers which must be configured to start a DMA transfer. In this section the steps necessary to configure one PDCA channel is outlined. The peripheral to transfer data to or from must be configured correctly in the Peripheral Select Register (PSR). This is performed by writing the Peripheral Identity (PID) value for the corresponding peripheral to the PID field in the PSR register. The PID also encodes the transfer direction, i.e. memory to peripheral or peripheral to memory. See Section 7.5.6. The transfer size must be written to the Transfer Size field in the Mode Register (MR.SIZE). The size must match the data size produced or consumed by the selected peripheral. See Section 7.5.7. The memory address to transfer to or from, depending on the PSR, must be written to the Memory Address Register (MAR). For each transfer the memory address is increased by either a one, two or four, depending on the size set in MR. See Section 7.5.2. The number of data items to transfer is written to the TCR register. If the PDCA channel is enabled, a transfer will start immediately after writing a non-zero value to TCR or the reload version of TCR, TCRR. After each transfer the TCR value is decreased by one. Both MAR and TCR can be read while the PDCA channel is active to monitor the DMA progress. See Section 7.5.3. The channel must be enabled for a transfer to start. A channel is enable by writing a one to the EN bit in the Control Register (CR). 7.5.2 Memory Pointer Each channel has a 32-bit Memory Address Register (MAR). This register holds the memory address for the next transfer to be performed. The register is automatically updated after each transfer. The address will be increased by either one, two or four depending on the size of the DMA transfer (byte, halfword or word). The MAR can be read at any time during transfer. 7.5.3 Transfer Counter Each channel has a 16-bit Transfer Counter Register (TCR). This register must be written with the number of transfers to be performed. The TCR register should contain the number of data items to be transferred independently of the transfer size. The TCR can be read at any time during transfer to see the number of remaining transfers. 7.5.4 Reload Registers Both the MAR and the TCR have a reload register, respectively Memory Address Reload Register (MARR) and Transfer Counter Reload Register (TCRR). These registers provide the possibility for the PDCA to work on two memory buffers for each channel. When one buffer has completed, MAR and TCR will be reloaded with the values in MARR and TCRR. The reload logic is always enabled and will trigger if the TCR reaches zero while TCRR holds a non-zero value. After reload, the MARR and TCRR registers are cleared. 43 32099I–01/2012 AT32UC3L016/32/64 If TCR is zero when writing to TCRR, the TCR and MAR are automatically updated with the value written in TCRR and MARR. 7.5.5 Ring Buffer When Ring Buffer mode is enabled the TCRR and MARR registers will not be cleared when TCR and MAR registers reload. This allows the PDCA to read or write to the same memory region over and over again until the transfer is actively stopped by the user. Ring Buffer mode is enabled by writing a one to the Ring Buffer bit in the Mode Register (MR.RING). 7.5.6 Peripheral Selection The Peripheral Select Register (PSR) decides which peripheral should be connected to the PDCA channel. A peripheral is selected by writing the corresponding Peripheral Identity (PID) to the PID field in the PSR register. Writing the PID will both select the direction of the transfer (memory to peripheral or peripheral to memory), which handshake interface to use, and the address of the peripheral holding register. Refer to the Peripheral Identity (PID) table in the Module Configuration section for the peripheral PID values. 7.5.7 Transfer Size The transfer size can be set individually for each channel to be either byte, halfword or word (8bit, 16-bit or 32-bit respectively). Transfer size is set by writing the desired value to the Transfer Size field in the Mode Register (MR.SIZE). When the PDCA moves data between peripherals and memory, data is automatically sized and aligned. When memory is accessed, the size specified in MR.SIZE and system alignment is used. When a peripheral register is accessed the data to be transferred is converted to a word where bit n in the data corresponds to bit n in the peripheral register. If the transfer size is byte or halfword, bits greater than 8 and16 respectively are set to zero. Refer to the Module Configuration section for information regarding what peripheral registers are used for the different peripherals and then to the peripheral specific chapter for information about the size option available for the different registers. 7.5.8 Enabling and Disabling Each DMA channel is enabled by writing a one to the Transfer Enable bit in the Control Register (CR.TEN) and disabled by writing a one to the Transfer Disable bit (CR.TDIS). The current status can be read from the Status Register (SR). While the PDCA channel is enabled all DMA request will be handled as long the TCR and TCRR is not zero. 7.5.9 Interrupts Interrupts can be enabled by writing a one to the corresponding bit in the Interrupt Enable Register (IER) and disabled by writing a one to the corresponding bit in the Interrupt Disable Register (IDR). The Interrupt Mask Register (IMR) can be read to see whether an interrupt is enabled or not. The current status of an interrupt source can be read through the Interrupt Status Register (ISR). The PDCA has three interrupt sources: • Reload Counter Zero - The TCRR register is zero. • Transfer Finished - Both the TCR and TCRR registers are zero. • Transfer Error - An error has occurred in accessing memory. 44 32099I–01/2012 AT32UC3L016/32/64 7.5.10 Priority If more than one PDCA channel is requesting transfer at a given time, the PDCA channels are prioritized by their channel number. Channels with lower numbers have priority over channels with higher numbers, giving channel zero the highest priority. 7.5.11 Error Handling If the Memory Address Register (MAR) is set to point to an invalid location in memory, an error will occur when the PDCA tries to perform a transfer. When an error occurs, the Transfer Error bit in the Interrupt Status Register (ISR.TERR) will be set and the DMA channel that caused the error will be stopped. In order to restart the channel, the user must program the Memory Address Register to a valid address and then write a one to the Error Clear bit in the Control Register (CR.ECLR). If the Transfer Error interrupt is enabled, an interrupt request will be generated when a transfer error occurs. 7.5.12 7.6 Peripheral Event Trigger Peripheral events can be used to trigger PDCA channel transfers. Peripheral Event synchronizations are enabled by writing a one to the Event Trigger bit in the Mode Register (MR.ETRIG). When set, all DMA requests will be blocked until a peripheral event is received. For each peripheral event received, only one data item is transferred. If no DMA requests are pending when a peripheral event is received, the PDCA will start a transfer as soon as a peripheral event is detected. If multiple events are received while the PDCA channel is busy transferring data, an overflow condition will be signaled in the Peripheral Event System. Refer to the Peripheral Event System chapter for more information. Performance Monitors Up to two performance monitors allow the user to measure the activity and stall cycles for PDCA transfers. To monitor a PDCA channel, the corresponding channel number must be written to one of the MON0/1CH fields in the Performance Control Register (PCONTROL) and a one must be written to the corresponding CH0/1EN bit in the same register. Due to performance monitor hardware resource sharing, the two monitor channels should NOT be programmed to monitor the same PDCA channel. This may result in UNDEFINED performance monitor behavior. 7.6.1 Measuring mechanisms Three different parameters can be measured by each channel: • The number of data transfer cycles since last channel reset, both for read and write • The number of stall cycles since last channel reset, both for read and write • The maximum latency since last channel reset, both for read and write These measurements can be extracted by software and used to generate indicators for bus latency, bus load, and maximum bus latency. Each of the counters has a fixed width, and may therefore overflow. When an overflow is encountered in either the Performance Channel Data Read/Write Cycle registers (PRDATA0/1 and PWDATA0/1) or the Performance Channel Read/Write Stall Cycles registers (PRSTALL0/1 and PWSTALL0/1) of a channel, all registers in the channel are reset. This behavior is altered if the Channel Overflow Freeze bit is one in the Performance Control register (PCONTROL.CH0/1OVF). If this bit is one, the channel registers are frozen when either DATA or STALL reaches its maximum value. This simplifies one-shot readout of the counter values. 45 32099I–01/2012 AT32UC3L016/32/64 The registers can also be manually reset by writing a one to the Channel Reset bit in the PCONTROL register (PCONTROL.CH0/1RES). The Performance Channel Read/Write Latency registers (PRLAT0/1 and PWLAT0/1) are saturating when their maximum count value is reached. The PRLAT0/1 and PWLAT0/1 registers can only be reset by writing a one to the corresponding reset bit in PCONTROL (PCONTROL.CH0/1RES). A counter is enabled by writing a one to the Channel Enable bit in the Performance Control Register (PCONTROL.CH0/1EN). 46 32099I–01/2012 AT32UC3L016/32/64 7.7 User Interface 7.7.1 Memory Map Overview Table 7-1. PDCA Register Memory Map Address Range Contents 0x000 - 0x03F DMA channel 0 configuration registers 0x040 - 0x07F DMA channel 1 configuration registers ... ... (0x000 - 0x03F)+m*0x040 DMA channel m configuration registers 0x800-0x830 Performance Monitor registers 0x834 Version register The channels are mapped as shown in Table 7-1. Each channel has a set of configuration registers, shown in Table 7-2, where n is the channel number. 7.7.2 Channel Memory Map Table 7-2. PDCA Channel Configuration Registers Offset Register Register Name Access Reset 0x000 + n*0x040 Memory Address Register MAR Read/Write 0x00000000 0x004 + n*0x040 Peripheral Select Register PSR Read/Write - (1) 0x008 + n*0x040 Transfer Counter Register TCR Read/Write 0x00000000 0x00C + n*0x040 Memory Address Reload Register MARR Read/Write 0x00000000 0x010 + n*0x040 Transfer Counter Reload Register TCRR Read/Write 0x00000000 0x014 + n*0x040 Control Register CR Write-only 0x00000000 0x018 + n*0x040 Mode Register MR Read/Write 0x00000000 0x01C + n*0x040 Status Register SR Read-only 0x00000000 0x020 + n*0x040 Interrupt Enable Register IER Write-only 0x00000000 0x024 + n*0x040 Interrupt Disable Register IDR Write-only 0x00000000 0x028 + n*0x040 Interrupt Mask Register IMR Read-only 0x00000000 0x02C + n*0x040 Interrupt Status Register ISR Read-only 0x00000000 Note: 1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter. 47 32099I–01/2012 AT32UC3L016/32/64 7.7.3 Performance Monitor Memory Map PDCA Performance Monitor Registers(1) Table 7-3. Offset Register Register Name Access Reset 0x800 Performance Control Register PCONTROL Read/Write 0x00000000 0x804 Channel0 Read Data Cycles PRDATA0 Read-only 0x00000000 0x808 Channel0 Read Stall Cycles PRSTALL0 Read-only 0x00000000 0x80C Channel0 Read Max Latency PRLAT0 Read-only 0x00000000 0x810 Channel0 Write Data Cycles PWDATA0 Read-only 0x00000000 0x814 Channel0 Write Stall Cycles PWSTALL0 Read-only 0x00000000 0x818 Channel0 Write Max Latency PWLAT0 Read-only 0x00000000 0x81C Channel1 Read Data Cycles PRDATA1 Read-only 0x00000000 0x820 Channel1 Read Stall Cycles PRSTALL1 Read-only 0x00000000 0x824 Channel1 Read Max Latency PRLAT1 Read-only 0x00000000 0x828 Channel1 Write Data Cycles PWDATA1 Read-only 0x00000000 0x82C Channel1 Write Stall Cycles PWSTALL1 Read-only 0x00000000 0x830 Channel1 Write Max Latency PWLAT1 Read-only 0x00000000 Note: 7.7.4 Version Register Memory Map Table 7-4. Note: 1. The number of performance monitors is device specific. If the device has only one performance monitor, the Channel1 registers are not available. Please refer to the Module Configuration section at the end of this chapter for the number of performance monitors on this device. PDCA Version Register Memory Map Offset Register Register Name Access Reset 0x834 Version Register VERSION Read-only - (1) 1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter. 48 32099I–01/2012 AT32UC3L016/32/64 7.7.5 Name: Memory Address Register MAR Access Type: Read/Write Offset: 0x000 + n*0x040 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 18 17 16 11 10 9 8 3 2 1 0 MADDR[31:24] 23 22 21 20 19 MADDR[23:16] 15 14 13 12 MADDR[15:8] 7 6 5 4 MADDR[7:0] • MADDR: Memory Address Address of memory buffer. MADDR should be programmed to point to the start of the memory buffer when configuring the PDCA. During transfer, MADDR will point to the next memory location to be read/written. 49 32099I–01/2012 AT32UC3L016/32/64 7.7.6 Name: Peripheral Select Register PSR Access Type: Read/Write Offset: 0x004 + n*0x040 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 PID • PID: Peripheral Identifier The Peripheral Identifier selects which peripheral should be connected to the DMA channel. Writing a PID will select both which handshake interface to use, the direction of the transfer and also the address of the Receive/Transfer Holding Register for the peripheral. See the Module Configuration section of PDCA for details. The width of the PID field is device specific and dependent on the number of peripheral modules in the device. 50 32099I–01/2012 AT32UC3L016/32/64 7.7.7 Name: Transfer Counter Register TCR Access Type: Read/Write Offset: 0x008 + n*0x040 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 TCV[15:8] 7 6 5 4 TCV[7:0] • TCV: Transfer Counter Value Number of data items to be transferred by the PDCA. TCV must be programmed with the total number of transfers to be made. During transfer, TCV contains the number of remaining transfers to be done. 51 32099I–01/2012 AT32UC3L016/32/64 7.7.8 Name: Memory Address Reload Register MARR Access Type: Read/Write Offset: 0x00C + n*0x040 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 MARV[31:24] 23 22 21 20 MARV[23:16] 15 14 13 12 MARV[15:8] 7 6 5 4 MARV[7:0] • MARV: Memory Address Reload Value Reload Value for the MAR register. This value will be loaded into MAR when TCR reaches zero if the TCRR register has a nonzero value. 52 32099I–01/2012 AT32UC3L016/32/64 7.7.9 Name: Transfer Counter Reload Register TCRR Access Type: Read/Write Offset: 0x010 + n*0x040 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 TCRV[15:8] 7 6 5 4 TCRV[7:0] • TCRV: Transfer Counter Reload Value Reload value for the TCR register. When TCR reaches zero, it will be reloaded with TCRV if TCRV has a positive value. If TCRV is zero, no more transfers will be performed for the channel. When TCR is reloaded, the TCRR register is cleared. 53 32099I–01/2012 AT32UC3L016/32/64 7.7.10 Name: Control Register CR Access Type: Write-only Offset: 0x014 + n*0x040 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - ECLR 7 6 5 4 3 2 1 0 - - - - - - TDIS TEN • ECLR: Transfer Error Clear Writing a zero to this bit has no effect. Writing a one to this bit will clear the Transfer Error bit in the Status Register (SR.TERR). Clearing the SR.TERR bit will allow the channel to transmit data. The memory address must first be set to point to a valid location. • TDIS: Transfer Disable Writing a zero to this bit has no effect. Writing a one to this bit will disable transfer for the DMA channel. • TEN: Transfer Enable Writing a zero to this bit has no effect. Writing a one to this bit will enable transfer for the DMA channel. 54 32099I–01/2012 AT32UC3L016/32/64 7.7.11 Name: Mode Register MR Access Type: Read/Write Offset: 0x018 + n*0x040 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - RING ETRIG SIZE • RING: Ring Buffer 0:The Ring buffer functionality is disabled. 1:The Ring buffer functionality is enabled. When enabled, the reload registers, MARR and TCRR will not be cleared after reload. • ETRIG: Event Trigger 0:Start transfer when the peripheral selected in Peripheral Select Register (PSR) requests a transfer. 1:Start transfer only when or after a peripheral event is received. • SIZE: Size of Transfer Table 7-5. Size of Transfer SIZE Size of Transfer 0 Byte 1 Halfword 2 Word 3 Reserved 55 32099I–01/2012 AT32UC3L016/32/64 7.7.12 Name: Status Register SR Access Type: Read-only Offset: 0x01C + n*0x040 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - TEN • TEN: Transfer Enabled This bit is cleared when the TDIS bit in CR is written to one. This bit is set when the TEN bit in CR is written to one. 0: Transfer is disabled for the DMA channel. 1: Transfer is enabled for the DMA channel. 56 32099I–01/2012 AT32UC3L016/32/64 7.7.13 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x020 + n*0x040 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - TERR TRC RCZ Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 57 32099I–01/2012 AT32UC3L016/32/64 7.7.14 Name: Interrupt Disable Register IDR Access Type: Write-only Offset: 0x024 + n*0x040 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - TERR TRC RCZ Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 58 32099I–01/2012 AT32UC3L016/32/64 7.7.15 Name: Interrupt Mask Register IMR Access Type: Read-only Offset: 0x028 + n*0x040 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - TERR TRC RCZ 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. A bit in this register is cleared when the corresponding bit in IDR is written to one. A bit in this register is set when the corresponding bit in IER is written to one. 59 32099I–01/2012 AT32UC3L016/32/64 7.7.16 Name: Interrupt Status Register ISR Access Type: Read-only Offset: 0x02C + n*0x040 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - TERR TRC RCZ • TERR: Transfer Error This bit is cleared when no transfer errors have occurred since the last write to CR.ECLR. This bit is set when one or more transfer errors has occurred since reset or the last write to CR.ECLR. • TRC: Transfer Complete This bit is cleared when the TCR and/or the TCRR holds a non-zero value. This bit is set when both the TCR and the TCRR are zero. • RCZ: Reload Counter Zero This bit is cleared when the TCRR holds a non-zero value. This bit is set when TCRR is zero. 60 32099I–01/2012 AT32UC3L016/32/64 7.7.17 Name: Performance Control Register PCONTROL Access Type: Read/Write Offset: 0x800 Reset Value: 0x00000000 31 30 29 28 - - 23 22 - - 15 14 13 12 - - - 7 6 - - 27 26 25 24 18 17 16 11 10 9 8 - - - CH1RES CH0RES 5 4 3 2 1 0 CH1OF CH0OF - - CH1EN CH0EN MON1CH 21 20 19 MON0CH • MON1CH: Performance Monitor Channel 1 • MON0CH: Performance Monitor Channel 0 The PDCA channel number to monitor with counter n Due to performance monitor hardware resource sharing, the two performance monitor channels should NOT be programmed to monitor the same PDCA channel. This may result in UNDEFINED monitor behavior. • CH1RES: Performance Channel 1 Counter Reset Writing a zero to this bit has no effect. Writing a one to this bit will reset the counter in the channel specified in MON1CH. This bit always reads as zero. • CH0RES: Performance Channel 0 Counter Reset Writing a zero to this bit has no effect. Writing a one to this bit will reset the counter in the channel specified in MON0CH. This bit always reads as zero. • CH1OF: Channel 1 Overflow Freeze 0: The performance channel registers are reset if DATA or STALL overflows. 1: All performance channel registers are frozen just before DATA or STALL overflows. • CH1OF: Channel 0 Overflow Freeze 0: The performance channel registers are reset if DATA or STALL overflows. 1: All performance channel registers are frozen just before DATA or STALL overflows. • CH1EN: Performance Channel 1 Enable 0: Performance channel 1 is disabled. 1: Performance channel 1 is enabled. • CH0EN: Performance Channel 0 Enable 0: Performance channel 0 is disabled. 1: Performance channel 0 is enabled. 61 32099I–01/2012 AT32UC3L016/32/64 7.7.18 Name: Performance Channel 0 Read Data Cycles PRDATA0 Access Type: Read-only Offset: 0x804 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 DATA[31:24] 23 22 21 20 DATA[23:16] 15 14 13 12 DATA[15:8] 7 6 5 4 DATA[7:0] • DATA: Data Cycles Counted Since Last Reset Clock cycles are counted using the CLK_PDCA_HSB clock 62 32099I–01/2012 AT32UC3L016/32/64 7.7.19 Name: Performance Channel 0 Read Stall Cycles PRSTALL0 Access Type: Read-only Offset: 0x808 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 STALL[31:24] 23 22 21 20 STALL[23:16] 15 14 13 12 STALL[15:8] 7 6 5 4 STALL[7:0] • STALL: Stall Cycles Counted Since Last Reset Clock cycles are counted using the CLK_PDCA_HSB clock 63 32099I–01/2012 AT32UC3L016/32/64 7.7.20 Name: Performance Channel 0 Read Max Latency PRLAT0 Access Type: Read/Write Offset: 0x80C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 LAT[15:8] 7 6 5 4 LAT[7:0] • LAT: Maximum Transfer Initiation Cycles Counted Since Last Reset Clock cycles are counted using the CLK_PDCA_HSB clock This counter is saturating. The register is reset only when PCONTROL.CH0RES is written to one. 64 32099I–01/2012 AT32UC3L016/32/64 7.7.21 Name: Performance Channel 0 Write Data Cycles PWDATA0 Access Type: Read-only Offset: 0x810 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 DATA[31:24] 23 22 21 20 DATA[23:16] 15 14 13 12 DATA[15:8] 7 6 5 4 DATA[7:0] • DATA: Data Cycles Counted Since Last Reset Clock cycles are counted using the CLK_PDCA_HSB clock 65 32099I–01/2012 AT32UC3L016/32/64 7.7.22 Name: Performance Channel 0 Write Stall Cycles PWSTALL0 Access Type: Read-only Offset: 0x814 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 STALL[31:24] 23 22 21 20 STALL[23:16] 15 14 13 12 STALL[15:8] 7 6 5 4 STALL[7:0] • STALL: Stall Cycles Counted Since Last Reset Clock cycles are counted using the CLK_PDCA_HSB clock 66 32099I–01/2012 AT32UC3L016/32/64 7.7.23 Name: Performance Channel 0 Write Max Latency PWLAT0 Access Type: Read/Write Offset: 0x818 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 LAT[15:8] 7 6 5 4 LAT[7:0] • LAT: Maximum Transfer Initiation Cycles Counted Since Last Reset Clock cycles are counted using the CLK_PDCA_HSB clock This counter is saturating. The register is reset only when PCONTROL.CH0RES is written to one. 67 32099I–01/2012 AT32UC3L016/32/64 7.7.24 Name: Performance Channel 1 Read Data Cycles PRDATA1 Access Type: Read-only Offset: 0x81C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 DATA[31:24] 23 22 21 20 DATA[23:16] 15 14 13 12 DATA[15:8] 7 6 5 4 DATA[7:0] • DATA: Data Cycles Counted Since Last Reset Clock cycles are counted using the CLK_PDCA_HSB clock 68 32099I–01/2012 AT32UC3L016/32/64 7.7.25 Name: Performance Channel 1 Read Stall Cycles PRSTALL1 Access Type: Read-only Offset: 0x820 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 STALL[31:24] 23 22 21 20 STALL[23:16] 15 14 13 12 STALL[15:8] 7 6 5 4 STALL[7:0] • STALL: Stall Cycles Counted Since Last Reset Clock cycles are counted using the CLK_PDCA_HSB clock 69 32099I–01/2012 AT32UC3L016/32/64 7.7.26 Name: Performance Channel 1 Read Max Latency PRLAT1 Access Type: Read/Write Offset: 0x824 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 LAT[15:8] 7 6 5 4 LAT[7:0] • LAT: Maximum Transfer Initiation Cycles Counted Since Last Reset Clock cycles are counted using the CLK_PDCA_HSB clock This counter is saturating. The register is reset only when PCONTROL.CH1RES is written to one. 70 32099I–01/2012 AT32UC3L016/32/64 7.7.27 Name: Performance Channel 1 Write Data Cycles PWDATA1 Access Type: Read-only Offset: 0x828 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 DATA[31:24] 23 22 21 20 DATA[23:16] 15 14 13 12 DATA[15:8] 7 6 5 4 DATA[7:0] • DATA: Data Cycles Counted Since Last Reset Clock cycles are counted using the CLK_PDCA_HSB clock 71 32099I–01/2012 AT32UC3L016/32/64 7.7.28 Name: Performance Channel 1 Write Stall Cycles PWSTALL1 Access Type: Read-only Offset: 0x82C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 STALL[31:24] 23 22 21 20 STALL[23:16] 15 14 13 12 STALL[15:8] 7 6 5 4 STALL[7:0] • STALL: Stall Cycles Counted Since Last Reset Clock cycles are counted using the CLK_PDCA_HSB clock 72 32099I–01/2012 AT32UC3L016/32/64 7.7.29 Name: Performance Channel 1 Write Max Latency PWLAT1 Access Type: Read/Write Offset: 0x830 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 LAT[15:8] 7 6 5 4 LAT[7:0] • LAT: Maximum Transfer Initiation Cycles Counted Since Last Reset Clock cycles are counted using the CLK_PDCA_HSB clock This counter is saturating. The register is reset only when PCONTROL.CH1RES is written to one. 73 32099I–01/2012 AT32UC3L016/32/64 7.7.30 Name: PDCA Version Register VERSION Access Type: Read-only Offset: 0x834 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant Number Reserved. No functionality associated. • VERSION: Version Number Version number of the module. No functionality associated. 74 32099I–01/2012 AT32UC3L016/32/64 7.8 Module Configuration The specific configuration for each PDCA instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 7-6. PDCA Configuration Feature PDCA Number of channels 12 Number of performance monitors 1 Table 7-7. Module Clock Name Module name Clock Name Description CLK_PDCA_HSB Clock for the PDCA HSB interface CLK_PDCA_PB Clock for the PDCA PB interface PDCA Table 7-8. Register Reset Values Register Reset Value PSR CH 0 0 PSR CH 1 1 PSR CH 2 2 PSR CH 3 3 PSR CH 4 4 PSR CH 5 5 PSR CH 6 6 PSR CH 7 7 PSR CH 8 8 PSR CH 9 9 PSR CH 10 10 PSR CH 11 11 VERSION 122 The table below defines the valid Peripheral Identifiers (PIDs). The direction is specified as observed from the memory, so RX means transfers from peripheral to memory and TX means from memory to peripheral. Table 7-9. Peripheral Identity Values PID Direction Peripheral Instance Peripheral Register 0 RX USART0 RHR 1 RX USART1 RHR 2 RX USART2 RHR 3 RX USART3 RHR 75 32099I–01/2012 AT32UC3L016/32/64 Table 7-9. Peripheral Identity Values PID Direction Peripheral Instance Peripheral Register 4 RX SPI RDR 5 RX TWIM0 RHR 6 RX TWIM1 RHR 7 RX TWIS0 RHR 8 RX TWIS1 RHR 9 RX ADCIFB LCDR 10 RX AW RHR 11 RX CAT ACOUNT 12 TX USART0 THR 13 TX USART1 THR 14 TX USART2 THR 15 TX USART3 THR 16 TX SPI TDR 17 TX TWIM0 THR 18 TX TWIM1 THR 19 TX TWIS0 THR 20 TX TWIS1 THR 21 TX AW THR 22 TX CAT MBLEN 76 32099I–01/2012 AT32UC3L016/32/64 8. Flash Controller (FLASHCDW) Rev: 1.0.2.0 8.1 Features • • • • • • • • • • • • • 8.2 Controls on-chip flash memory Supports 0 and 1 wait state bus access Buffers reducing penalty of wait state in sequential code or loops Allows interleaved burst reads for systems with one wait state, outputting one 32-bit word per clock cycle for sequential reads Secure State for supporting FlashVault technology 32-bit HSB interface for reads from flash and writes to page buffer 32-bit PB interface for issuing commands to and configuration of the controller Flash memory is divided into 16 regions can be individually protected or unprotected Additional protection of the Boot Loader pages Supports reads and writes of general-purpose Non Volatile Memory (NVM) bits Supports reads and writes of additional NVM pages Supports device protection through a security bit Dedicated command for chip-erase, first erasing all on-chip volatile memories before erasing flash and clearing security bit Overview The Flash Controller (FLASHCDW) interfaces the on-chip flash memory with the 32-bit internal HSB bus. The controller manages the reading, writing, erasing, locking, and unlocking sequences. 8.3 Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 8.3.1 Power Management If the CPU enters a sleep mode that disables clocks used by the FLASHCDW, the FLASHCDW will stop functioning and resume operation after the system wakes up from sleep mode. 8.3.2 Clocks The FLASHCDW has two bus clocks connected: One High Speed Bus clock (CLK_FLASHCDW_HSB) and one Peripheral Bus clock (CLK_FLASHCDW_PB). These clocks are generated by the Power Manager. Both clocks are enabled at reset, and can be disabled by writing to the Power Manager. The user has to ensure that CLK_FLASHCDW_HSB is not turned off before reading the flash or writing the pagebuffer and that CLK_FLASHCDW_PB is not turned off before accessing the FLASHCDW configuration and control registers. Failing to do so may deadlock the bus. 8.3.3 Interrupts The FLASHCDW interrupt request lines are connected to the interrupt controller. Using the FLASHCDW interrupts requires the interrupt controller to be programmed first. 77 32099I–01/2012 AT32UC3L016/32/64 8.3.4 8.4 8.4.1 Debug Operation When an external debugger forces the CPU into debug mode, the FLASHCDW continues normal operation. If the FLASHCDW is configured in a way that requires it to be periodically serviced by the CPU through interrupts or similar, improper operation or data loss may result during debugging. Functional Description Bus Interfaces The FLASHCDW has two bus interfaces, one High Speed Bus (HSB) interface for reads from the flash memory and writes to the page buffer, and one Peripheral Bus (PB) interface for issuing commands and reading status from the controller. 8.4.2 Memory Organization The flash memory is divided into a set of pages. A page is the basic unit addressed when programming the flash. A page consists of several words. The pages are grouped into 16 regions of equal size. Each of these regions can be locked by a dedicated fuse bit, protecting it from accidental modification. • p pages (FLASH_P) • w bytes in each page and in the page buffer (FLASH_W) • pw bytes in total (FLASH_PW) • f general-purpose fuse bits (FLASH_F), used as region lock bits and for other device-specific purposes • 1 security fuse bit • 1 User page 8.4.3 User Page The User page is an additional page, outside the regular flash array, that can be used to store various data, such as calibration data and serial numbers. This page is not erased by regular chip erase. The User page can only be written and erased by a special set of commands. Read accesses to the User page are performed just as any other read accesses to the flash. The address map of the User page is given in Figure 8-1 on page 80. 8.4.4 Read Operations The on-chip flash memory is typically used for storing instructions to be executed by the CPU. The CPU will address instructions using the HSB bus, and the FLASHCDW will access the flash memory and return the addressed 32-bit word. In systems where the HSB clock period is slower than the access time of the flash memory, the FLASHCDW can operate in 0 wait state mode, and output one 32-bit word on the bus per clock cycle. If the clock frequency allows, the user should use 0 wait state mode, because this gives the highest performance as no stall cycles are encountered. The FLASHCDW can also operate in systems where the HSB bus clock period is faster than the access speed of the flash memory. Wait state support and a read granularity of 64 bits ensure efficiency in such systems. Performance for systems with high clock frequency is increased since the internal read word width of the flash memory is 64 bits. When a 32-bit word is to be addressed, the word itself and 78 32099I–01/2012 AT32UC3L016/32/64 also the other word in the same 64-bit location is read. The first word is output on the bus, and the other word is put into an internal buffer. If a read to a sequential address is to be performed in the next cycle, the buffered word is output on the bus, while the next 64-bit location is read from the flash memory. Thus, latency in 1 wait state mode is hidden for sequential fetches. The programmer can select the wait states required by writing to the FWS field in the Flash Control Register (FCR). It is the responsibility of the programmer to select a number of wait states compatible with the clock frequency and timing characteristics of the flash memory. In 0ws mode, no wait states are encountered on any flash read operations. In 1 ws mode, one stall cycle is encountered on the first access in a single or burst transfer. In 1 ws mode, if the first access in a burst access is to an address that is not 64-bit aligned, an additional stall cycle is also encountered when reading the second word in the burst. All subsequent words in the burst are accessed without any stall cycles. The Flash Controller provides two sets of buffers that can be enabled in order to speed up instruction fetching. These buffers can be enabled by writing a one to the FCR.SEQBUF and FCR.BRBUF bits. The SEQBUF bit enables buffering hardware optimizing sequential instruction fetches. The BRBUF bit enables buffering hardware optimizing tight inner loops. These buffers are never used when the flash is in 0 wait state mode. Usually, both these buffers should be enabled when operating in 1 wait state mode. Some users requiring absolute cycle determinism may want to keep the buffers disabled. The Flash Controller address space is displayed in Figure 8-1. The memory space between address pw and the User page is reserved, and reading addresses in this space returns an undefined result. The User page is permanently mapped to an offset of 0x00800000 from the start address of the flash memory. Table 8-1. User Page Addresses Memory type Start address, byte sized Size Main array 0 pw bytes User 0x00800000 w bytes 79 32099I–01/2012 AT32UC3L016/32/64 Figure 8-1. Memory Map for the Flash Memories Offset from base address Reserved User Page Reserved 0x0080 0000 Flash data array pw 0 Flash base address Flash with User Page All addresses are byte addresses 8.4.5 High Speed Read Mode The flash provides a High Speed Read Mode, offering slightly higher flash read speed at the cost of higher power consumption. Two dedicated commands, High Speed Read Mode Enable (HSEN) and High Speed Read Mode Disable (HSDIS) control the speed mode. The High Speed Mode (HSMODE) bit in the Flash Status Register (FSR) shows which mode the flash is in. After reset, the High Speed Mode is disabled, and must be manually enabled if the user wants to. Refer to the Electrical Characteristics chapter at the end of this datasheet for details on the maximum clock frequencies in Normal and High Speed Read Mode. 80 32099I–01/2012 AT32UC3L016/32/64 Figure 8-2. High Speed Mode Frequency 1 wait state 0 wait state Frequency limit for 0 wait state operation Speed mode H al m h ig or N 8.4.6 Quick Page Read A dedicated command, Quick Page Read (QPR), is provided to read all words in an addressed page. All bits in all words in this page are AND’ed together, returning a 1-bit result. This result is placed in the Quick Page Read Result (QPRR) bit in Flash Status Register (FSR). The QPR command is useful to check that a page is in an erased state. The QPR instruction is much faster than performing the erased-page check using a regular software subroutine. 8.4.7 Quick User Page Read A dedicated command, Quick User Page Read (QPRUP), is provided to read all words in the user page. All bits in all words in this page are AND’ed together, returning a 1-bit result. This result is placed in the Quick Page Read Result (QPRR) bit in Flash Status Register (FSR). The QPRUP command is useful to check that a page is in an erased state. The QPRUP instruction is much faster than performing the erased-page check using a regular software subroutine. 8.4.8 Page Buffer Operations The flash memory has a write and erase granularity of one page; data is written and erased in chunks of one page. When programming a page, the user must first write the new data into the Page Buffer. The contents of the entire Page Buffer is copied into the desired page in flash memory when the user issues the Write Page command, Refer to Section 8.5.1 on page 83. In order to program data into flash page Y, write the desired data to locations Y0 to Y31 in the regular flash memory map. Writing to an address A in the flash memory map will not update the flash memory, but will instead update location A%32 in the page buffer. The PAGEN field in the Flash Command (FCMD) register will at the same time be updated with the value A/32. 81 32099I–01/2012 AT32UC3L016/32/64 Figure 8-3. Mapping from Page Buffer to Flash Flash All locations are doubleword locations Page Buffer 64-bit data 31 27 23 19 15 11 7 3 30 26 22 18 14 10 6 2 29 25 21 17 13 9 5 1 28 24 20 16 12 8 4 0 Z31 Z27 Z23 Z19 Z15 Z11 Z7 Z3 Y31 Y27 Y23 Y19 Y15 Y11 Y7 Y3 X31 X27 X23 X19 X15 X11 X7 X3 Z30 Z29 Z26 Z25 Z22 Z21 Z18 Z17 Page Z Z14 Z13 Z10 Z9 Z6 Z5 Z2 Z1 Y30 Y29 Y26 Y25 Y22 Y21 Y18 Y17 Page Y Y14 Y13 Y10 Y9 Y6 Y5 Y2 Y1 X30 X29 X26 X25 X22 X21 X18 X17 Page X X14 X13 X10 X9 X6 X5 X2 X1 Z28 Z24 Z20 Z16 Z12 Z8 Z4 Z0 Y28 Y24 Y20 Y16 Y12 Y8 Y4 Y0 X28 X24 X20 X16 X12 X8 X4 X0 Internally, the flash memory stores data in 64-bit doublewords. Therefore, the native data size of the Page Buffer is also a 64-bit doubleword. All locations shown in Figure 8-3 are therefore doubleword locations. Since the HSB bus only has a 32-bit data width, two 32-bit HSB transfers must be performed to write a 64-bit doubleword into the Page Buffer. The FLASHCDW has logic to combine two 32-bit HSB transfers into a 64-bit data before writing this 64-bit data into the Page Buffer. This logic requires the word with the low address to be written to the HSB bus before the word with the high address. To exemplify, to write a 64-bit value to doubleword X0 residing in page X, first write a 32-bit word to the byte address pointing to address X0, thereafter write a word to the byte address pointing to address (X0+4). The page buffer is word-addressable and should only be written with aligned word transfers, never with byte or halfword transfers. The page buffer can not be read. The page buffer is also used for writes to the User page. Page buffer write operations are performed with 4 wait states. Any accesses attempted to the FLASHCDW on the HSB bus during these cycles will be automatically stalled. Writing to the page buffer can only change page buffer bits from one to zero, i.e. writing 0xAAAAAAAA to a page buffer location that has the value 0x00000000 will not change the page buffer value. The only way to change a bit from zero to one is to erase the entire page buffer with the Clear Page Buffer command. 82 32099I–01/2012 AT32UC3L016/32/64 The page buffer is not automatically reset after a page write. The programmer should do this manually by issuing the Clear Page Buffer flash command. This can be done after a page write, or before the page buffer is loaded with data to be stored to the flash page. 8.5 Flash Commands The FLASHCDW offers a command set to manage programming of the flash memory, locking and unlocking of regions, and full flash erasing. See Section 8.8.2 for a complete list of commands. To run a command, the CMD field in the Flash Command Register (FCMD) has to be written with the command number. As soon as the FCMD register is written, the FRDY bit in the Flash Status Register (FSR) is automatically cleared. Once the current command is complete, the FSR.FRDY bit is automatically set. If an interrupt has been enabled by writing a one to FCR.FRDY, the interrupt request line of the Flash Controller is activated. All flash commands except for Quick Page Read (QPR) and Quick User Page Read (QPRUP) will generate an interrupt request upon completion if FCR.FRDY is one. Any HSB bus transfers attempting to read flash memory when the FLASHCDW is busy executing a flash command will be stalled, and allowed to continue when the flash command is complete. After a command has been written to FCMD, the programming algorithm should wait until the command has been executed before attempting to read instructions or data from the flash or writing to the page buffer, as the flash will be busy. The waiting can be performed either by polling the Flash Status Register (FSR) or by waiting for the flash ready interrupt. The command written to FCMD is initiated on the first clock cycle where the HSB bus interface in FLASHCDW is IDLE. The user must make sure that the access pattern to the FLASHCDW HSB interface contains an IDLE cycle so that the command is allowed to start. Make sure that no bus masters such as DMA controllers are performing endless burst transfers from the flash. Also, make sure that the CPU does not perform endless burst transfers from flash. This is done by letting the CPU enter sleep mode after writing to FCMD, or by polling FSR for command completion. This polling will result in an access pattern with IDLE HSB cycles. All the commands are protected by the same keyword, which has to be written in the eight highest bits of the FCMD register. Writing FCMD with data that does not contain the correct key and/or with an invalid command has no effect on the flash memory; however, the PROGE bit is set in the Flash Status Register (FSR). This bit is automatically cleared by a read access to the FSR register. Writing a command to FCMD while another command is being executed has no effect on the flash memory; however, the PROGE bit is set in the Flash Status Register (FSR). This bit is automatically cleared by a read access to the FSR register. If the current command writes or erases a page in a locked region, or a page protected by the BOOTPROT fuses, the command has no effect on the flash memory; however, the LOCKE bit is set in the FSR register. This bit is automatically cleared by a read access to the FSR register. 8.5.1 Write/Erase Page Operation Flash technology requires that an erase must be done before programming. The entire flash can be erased by an Erase All command. Alternatively, pages can be individually erased by the Erase Page command. The User page can be written and erased using the mechanisms described in this chapter. 83 32099I–01/2012 AT32UC3L016/32/64 After programming, the page can be locked to prevent miscellaneous write or erase sequences. Locking is performed on a per-region basis, so locking a region locks all pages inside the region. Additional protection is provided for the lowermost address space of the flash. This address space is allocated for the Boot Loader, and is protected both by the lock bit(s) corresponding to this address space, and the BOOTPROT[2:0] fuses. Data to be written is stored in an internal buffer called the page buffer. The page buffer contains w words. The page buffer wraps around within the internal memory area address space and appears to be repeated by the number of pages in it. Writing of 8-bit and 16-bit data to the page buffer is not allowed and may lead to unpredictable data corruption. Data must be written to the page buffer before the programming command is written to the Flash Command Register (FCMD). The sequence is as follows: • Reset the page buffer with the Clear Page Buffer command. • Fill the page buffer with the desired contents as described in Section 8.4.8 on page 81. • Programming starts as soon as the programming key and the programming command are written to the Flash Command Register. The PAGEN field in the Flash Command Register (FCMD) must contain the address of the page to write. PAGEN is automatically updated when writing to the page buffer, but can also be written to directly. The FRDY bit in the Flash Status Register (FSR) is automatically cleared when the page write operation starts. • When programming is completed, the FRDY bit in the Flash Status Register (FSR) is set. If an interrupt was enabled by writing FCR.FRDY to one, an interrupt request is generated. Two errors can be detected in the FSR register after a programming sequence: • Programming Error: A bad keyword and/or an invalid command have been written in the FCMD register. • Lock Error: Can have two different causes: – The page to be programmed belongs to a locked region. A command must be executed to unlock the corresponding region before programming can start. – A bus master without secure status attempted to program a page requiring secure privileges. 8.5.2 Erase All Operation The entire memory is erased if the Erase All command (EA) is written to the Flash Command Register (FCMD). Erase All erases all bits in the flash array. The User page is not erased. All flash memory locations, the general-purpose fuse bits, and the security bit are erased (reset to 0xFF) after an Erase All. The EA command also ensures that all volatile memories, such as register file and RAMs, are erased before the security bit is erased. Erase All operation is allowed only if no regions are locked, and the BOOTPROT fuses are configured with a BOOTPROT region size of 0. Thus, if at least one region is locked, the bit LOCKE in FSR is set and the command is cancelled. If the LOCKE bit in FCR is one, an interrupt request is set generated. When the command is complete, the FRDY bit in the Flash Status Register (FSR) is set. If an interrupt has been enabled by writing FCR.FRDY to one, an interrupt request is generated. Two errors can be detected in the FSR register after issuing the command: 84 32099I–01/2012 AT32UC3L016/32/64 • Programming Error: A bad keyword and/or an invalid command have been written in the FCMD register. • Lock Error: At least one lock region is protected, or BOOTPROT is different from 0. The erase command has been aborted and no page has been erased. A “Unlock region containing given page” (UP) command must be executed to unlock any locked regions. 8.5.3 Region Lock Bits The flash memory has p pages, and these pages are grouped into 16 lock regions, each region containing p/16 pages. Each region has a dedicated lock bit preventing writing and erasing pages in the region. After production, the device may have some regions locked. These locked regions are reserved for a boot or default application. Locked regions can be unlocked to be erased and then programmed with another application or other data. To lock or unlock a region, the commands Lock Region Containing Page (LP) and Unlock Region Containing Page (UP) are provided. Writing one of these commands, together with the number of the page whose region should be locked/unlocked, performs the desired operation. One error can be detected in the FSR register after issuing the command: • Programming Error: A bad keyword and/or an invalid command have been written in the FCMD register. The lock bits are implemented using the lowest 16 general-purpose fuse bits. This means that lock bits can also be set/cleared using the commands for writing/erasing general-purpose fuse bits, see Section 8.6. The general-purpose bit being in an erased (1) state means that the region is unlocked. The lowermost pages in the flash can additionally be protected by the BOOTPROT fuses, see Section 8.6. 8.6 General-purpose Fuse Bits The flash memory has a number of general-purpose fuse bits that the application programmer can use freely. The fuse bits can be written and erased using dedicated commands, and read 85 32099I–01/2012 AT32UC3L016/32/64 through a dedicated Peripheral Bus address. Some of the general-purpose fuse bits are reserved for special purposes, and should not be used for other functions: Table 8-2. General-purpose Fuses with Special Functions GeneralPurpose fuse number Name Usage 15:0 LOCK Region lock bits. EPFL External Privileged Fetch Lock. Used to prevent the CPU from fetching instructions from external memories when in privileged mode. This bit can only be changed when the security bit is cleared. The address range corresponding to external memories is device-specific, and not known to the Flash Controller. This fuse bit is simply routed out of the CPU or bus system, the Flash Controller does not treat this fuse in any special way, except that it can not be altered when the security bit is set. If the security bit is set, only an external JTAG or aWire Chip Erase can clear EPFL. No internal commands can alter EPFL if the security bit is set. When the fuse is erased (i.e. "1"), the CPU can execute instructions fetched from external memories. When the fuse is programmed (i.e. "0"), instructions can not be executed from external memories. This fuse has no effect in devices with no External Memory Interface (EBI). BOOTPROT Used to select one of eight different bootloader sizes. Pages included in the bootloader area can not be erased or programmed except by a JTAG or aWire chip erase. BOOTPROT can only be changed when the security bit is cleared. If the security bit is set, only an external JTAG or aWire Chip Erase can clear BOOTPROT, and thereby allow the pages protected by BOOTPROT to be programmed. No internal commands can alter BOOTPROT or the pages protected by BOOTPROT if the security bit is set. SECURE Used to configure secure state and secure state debug capabilities. Decoded into SSE and SSDE signals as shown in Table 8-5. Refer to the AVR32 Architecture Manual and the AVR32UC Technical Reference Manual for more details on SSE and SSDE. UPROT If programmed (i.e. “0”), the JTAG USER PROTECTION feature is enabled. If this fuse is programmed some HSB addresses will be accessible by JTAG access even if the flash security fuse is programmed. Refer to the JTAG documentation for more information on this functionality. This bit can only be changed when the security bit is cleared. 16 19:17 21:20 22 The BOOTPROT fuses protects the following address space for the Boot Loader: 86 32099I–01/2012 AT32UC3L016/32/64 Table 8-4. Boot Loader Area Specified by BOOTPROT BOOTPROT Pages protected by BOOTPROT Size of protected memory 7 None 0 6 0-1 512 byte 5 0-3 1Kbyte 4 0-7 2Kbyte 3 0-15 4Kbyte 2 0-31 8Kbyte 1 0-63 16Kbyte 0 0-127 32Kbyte The SECURE fuses have the following functionality: Table 8-5. Secure State Configuration SECURE Functionality SSE SSDE 00 Secure state disabled 0 0 01 Secure enabled, secure state debug enabled 1 1 10 Secure enabled, secure state debug disabled 1 0 11 Secure state disabled 0 0 To erase or write a general-purpose fuse bit, the commands Write General-Purpose Fuse Bit (WGPB) and Erase General-Purpose Fuse Bit (EGPB) are provided. Writing one of these commands, together with the number of the fuse to write/erase, performs the desired operation. An entire General-Purpose Fuse byte can be written at a time by using the Program GP Fuse Byte (PGPFB) instruction. A PGPFB to GP fuse byte 2 is not allowed if the flash is locked by the security bit. The PFB command is issued with a parameter in the PAGEN field: • PAGEN[2:0] - byte to write • PAGEN[10:3] - Fuse value to write All general-purpose fuses can be erased by the Erase All General-Purpose fuses (EAGP) command. An EAGP command is not allowed if the flash is locked by the security bit. Two errors can be detected in the FSR register after issuing these commands: • Programming Error: A bad keyword and/or an invalid command have been written in the FCMD register. • Lock Error: – A write or erase of the BOOTPROT or EPFL or UPROT fuse bits was attempted while the flash is locked by the security bit. – A write or erase of the SECURE fuse bits was attempted when SECURE mode was enabled. The lock bits are implemented using the lowest 16 general-purpose fuse bits. This means that the 16 lowest general-purpose fuse bits can also be written/erased using the commands for locking/unlocking regions, see Section 8.5.3. 87 32099I–01/2012 AT32UC3L016/32/64 8.7 Security Bit The security bit allows the entire device to be locked from external JTAG, aWire, or other debug access for code security. The security bit can be written by a dedicated command, Set Security Bit (SSB). Once set, the only way to clear the security bit is through the JTAG or aWire Chip Erase command. Once the security bit is set, the following Flash Controller commands will be unavailable and return a lock error if attempted: • Write General-Purpose Fuse Bit (WGPB) to BOOTPROT or EPFL fuses • Erase General-Purpose Fuse Bit (EGPB) to BOOTPROT or EPFL fuses • Program General-Purpose Fuse Byte (PGPFB) of fuse byte 2 • Erase All General-Purpose Fuses (EAGPF) One error can be detected in the FSR register after issuing the command: • Programming Error: A bad keyword and/or an invalid command have been written in the FCMD register. 88 32099I–01/2012 AT32UC3L016/32/64 8.8 User Interface Table 8-6. Note: FLASHCDW Register Memory Map Offset Register Register Name Access Reset 0x00 Flash Control Register FCR Read/Write 0x00000000 0x04 Flash Command Register FCMD Read/Write 0x00000000 0x08 Flash Status Register FSR Read-only -(1) 0x0C Flash Parameter Register FPR Read-only -(3) 0x10 Flash Version Register FVR Read-only -(3) 0x14 Flash General Purpose Fuse Register Hi FGPFRHI Read-only -(2) 0x18 Flash General Purpose Fuse Register Lo FGPFRLO Read-only -(2) 1. The value of the Lock bits depend on their programmed state. All other bits in FSR are 0. 2. All bits in FGPRHI/LO are dependent on the programmed state of the fuses they map to. Any bits in these registers not mapped to a fuse read as 0. 3. The reset values for these registers are device specific. Please refer to the Module Configuration section at the end of this chapter. 89 32099I–01/2012 AT32UC3L016/32/64 8.8.1 Name: Flash Control Register FCR Access Type: Read/Write Offset: 0x00 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - BRBUF SEQBUF - 7 6 5 4 3 2 1 0 - FWS - - PROGE LOCKE - FRDY • BRBUF: Branch Target Instruction Buffer Enable 0: The Branch Target Instruction Buffer is disabled. 1: The Branch Target Instruction Buffer is enabled. • SEQBUF: Sequential Instruction Fetch Buffer Enable 0: The Sequential Instruction Fetch Buffer is disabled. 1: The Sequential Instruction Fetch Buffer is enabled. • FWS: Flash Wait State 0: The flash is read with 0 wait states. 1: The flash is read with 1 wait state. • PROGE: Programming Error Interrupt Enable 0: Programming Error does not generate an interrupt request. 1: Programming Error generates an interrupt request. • LOCKE: Lock Error Interrupt Enable 0: Lock Error does not generate an interrupt request. 1: Lock Error generates an interrupt request. • FRDY: Flash Ready Interrupt Enable 0: Flash Ready does not generate an interrupt request. 1: Flash Ready generates an interrupt request. 90 32099I–01/2012 AT32UC3L016/32/64 8.8.2 Name: Flash Command Register FCMD Access Type: Read/Write Offset: 0x04 Reset Value: 0x00000000 The FCMD can not be written if the flash is in the process of performing a flash command. Doing so will cause the FCR write to be ignored, and the PROGE bit in FSR to be set. 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 KEY 23 22 21 20 PAGEN [15:8] 15 14 13 12 PAGEN [7:0] 7 6 - - 5 4 CMD • KEY: Write protection key This field should be written with the value 0xA5 to enable the command defined by the bits of the register. If the field is written with a different value, the write is not performed and no action is started. This field always reads as 0. • PAGEN: Page number The PAGEN field is used to address a page or fuse bit for certain operations. In order to simplify programming, the PAGEN field is automatically updated every time the page buffer is written to. For every page buffer write, the PAGEN field is updated with the page number of the address being written to. Hardware automatically masks writes to the PAGEN field so that only bits representing valid page numbers can be written, all other bits in PAGEN are always 0. As an example, in a flash with 1024 pages (page 0 - page 1023), bits 15:10 will always be 0. Table 8-7. Semantic of PAGEN field in different commands Command PAGEN description No operation Not used Write Page The number of the page to write Clear Page Buffer Not used Lock region containing given Page Page number whose region should be locked Unlock region containing given Page Page number whose region should be unlocked Erase All Not used Write General-Purpose Fuse Bit GPFUSE # Erase General-Purpose Fuse Bit GPFUSE # Set Security Bit Not used 91 32099I–01/2012 AT32UC3L016/32/64 Table 8-7. Semantic of PAGEN field in different commands Command PAGEN description Program GP Fuse Byte WriteData[7:0], ByteAddress[2:0] Erase All GP Fuses Not used Quick Page Read Page number Write User Page Not used Erase User Page Not used Quick Page Read User Page Not used High Speed Mode Enable Not used High Speed Mode Disable Not used • CMD: Command This field defines the flash command. Issuing any unused command will cause the Programming Error bit in FSR to be set, and the corresponding interrupt to be requested if the PROGE bit in FCR is one. Table 8-8. Set of commands Command Value Mnemonic No operation 0 NOP Write Page 1 WP Erase Page 2 EP Clear Page Buffer 3 CPB Lock region containing given Page 4 LP Unlock region containing given Page 5 UP Erase All 6 EA Write General-Purpose Fuse Bit 7 WGPB Erase General-Purpose Fuse Bit 8 EGPB Set Security Bit 9 SSB Program GP Fuse Byte 10 PGPFB Erase All GPFuses 11 EAGPF Quick Page Read 12 QPR Write User Page 13 WUP Erase User Page 14 EUP Quick Page Read User Page 15 QPRUP High Speed Mode Enable 16 HSEN High Speed Mode Disable 17 HSDIS RESERVED 16-31 92 32099I–01/2012 AT32UC3L016/32/64 8.8.3 Name: Flash Status Register FSR Access Type: Read-only Offset: 0x08 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 LOCK15 LOCK14 LOCK13 LOCK12 LOCK11 LOCK10 LOCK9 LOCK8 23 22 21 20 19 18 17 16 LOCK7 LOCK6 LOCK5 LOCK4 LOCK3 LOCK2 LOCK1 LOCK0 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - HSMODE QPRR SECURITY PROGE LOCKE - FRDY • LOCKx: Lock Region x Lock Status 0: The corresponding lock region is not locked. 1: The corresponding lock region is locked. • HSMODE: High-Speed Mode 0: High-speed mode disabled. 1: High-speed mode enabled. • QPRR: Quick Page Read Result 0: The result is zero, i.e. the page is not erased. 1: The result is one, i.e. the page is erased. • SECURITY: Security Bit Status 0: The security bit is inactive. 1: The security bit is active. • PROGE: Programming Error Status Automatically cleared when FSR is read. 0: No invalid commands and no bad keywords were written in the Flash Command Register FCMD. 1: An invalid command and/or a bad keyword was/were written in the Flash Command Register FCMD. • LOCKE: Lock Error Status Automatically cleared when FSR is read. 0: No programming of at least one locked lock region has happened since the last read of FSR. 1: Programming of at least one locked lock region has happened since the last read of FSR. • FRDY: Flash Ready Status 0: The Flash Controller is busy and the application must wait before running a new command. 1: The Flash Controller is ready to run a new command. 93 32099I–01/2012 AT32UC3L016/32/64 8.8.4 Name: Flash Parameter Register FPR Access Type: Read-only Offset: 0x0C Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - 7 6 5 4 3 - - - - PSZ 2 1 0 FSZ • PSZ: Page Size The size of each flash page. Table 8-9. Flash Page Size PSZ Page Size 0 32 Byte 1 64 Byte 2 128 Byte 3 256 Byte 4 512 Byte 5 1024 Byte 6 2048 Byte 7 4096 Byte 94 32099I–01/2012 AT32UC3L016/32/64 • FSZ: Flash Size The size of the flash. Not all device families will provide all flash sizes indicated in the table. Table 8-10. Flash Size FSZ Flash Size FSZ Flash Size 0 4 Kbyte 8 192 Kbyte 1 8 Kbyte 9 256 Kbyte 2 16 Kbyte 10 384 Kbyte 3 32 Kbyte 11 512 Kbyte 4 48 Kbyte 12 768 Kbyte 5 64 Kbyte 13 1024 Kbyte 6 96 Kbyte 14 2048 Kbyte 7 128 Kbyte 15 Reserved 95 32099I–01/2012 AT32UC3L016/32/64 8.8.5 Name: Flash Version Register FVR Access Type: Read-only Offset: 0x10 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant Number Reserved. No functionality associated. • VERSION: Version Number Version number of the module. No functionality associated. 96 32099I–01/2012 AT32UC3L016/32/64 8.8.6 Name: Flash General Purpose Fuse Register High FGPFRHI Access Type: Read-only Offset: 0x14 Reset Value: - 31 30 29 28 27 26 25 24 GPF63 GPF62 GPF61 GPF60 GPF59 GPF58 GPF57 GPF56 23 22 21 20 19 18 17 16 GPF55 GPF54 GPF53 GPF52 GPF51 GPF50 GPF49 GPF48 15 14 13 12 11 10 9 8 GPF47 GPF46 GPF45 GPF44 GPF43 GPF42 GPF41 GPF40 7 6 5 4 3 2 1 0 GPF39 GPF38 GPF37 GPF36 GPF35 GPF34 GPF33 GPF32 This register is only used in systems with more than 32 GP fuses. • GPFxx: General Purpose Fuse xx 0: The fuse has a written/programmed state. 1: The fuse has an erased state. 97 32099I–01/2012 AT32UC3L016/32/64 8.8.7 Name: Flash General Purpose Fuse Register Low FGPFRLO Access Type: Read-only Offset: 0x18 Reset Value: - 31 30 29 28 27 26 25 24 GPF31 GPF30 GPF29 GPF28 GPF27 GPF26 GPF25 GPF24 23 22 21 20 19 18 17 16 GPF23 GPF22 GPF21 GPF20 GPF19 GPF18 GPF17 GPF16 15 14 13 12 11 10 9 8 GPF15 GPF14 GPF13 GPF12 GPF11 GPF10 GPF09 GPF08 7 6 5 4 3 2 1 0 GPF07 GPF06 GPF05 GPF04 GPF03 GPF02 GPF01 GPF00 • GPFxx: General Purpose Fuse xx 0: The fuse has a written/programmed state. 1: The fuse has an erased state. 98 32099I–01/2012 AT32UC3L016/32/64 8.9 Fuse Settings The flash contains 32 general purpose fuses. These 32 fuses can be found in the Flash General Purpose Fuse Register Low (FGPFRLO). The Flash General Purpose Fuse Register High (FGPFRHI) is not used. In addition to the general purpose fuses, parts of the flash user page can have a defined meaning outside of the flash controller and will also be described in this section. Note that when writing to the user page the values do not get loaded by the other modules on the device until a chip reset occurs. The general purpose fuses are erased by a JTAG or aWire chip erase. 99 32099I–01/2012 AT32UC3L016/32/64 8.9.1 Flash General Purpose Fuse Register Low (FGPFRLO) 31 30 29 BODEN 28 27 BODHYST 23 22 21 BODLEVEL[0] UPROT 15 14 25 24 17 16 BODLEVEL[5:1] 20 19 SECURE 13 26 18 BOOTPROT 12 EPFL 11 10 9 8 3 2 1 0 LOCK[15:8] 7 6 5 4 LOCK[7:0] • BODEN: Brown Out Detector Enable BODEN Description 00 BOD disabled 01 BOD enabled, BOD reset enabled 10 BOD enabled, BOD reset disabled 11 BOD disabled • BODHYST: Brown Out Detector Hysteresis 0: The Brown out detector hysteresis is disabled 1: The Brown out detector hysteresis is enabled • BODLEVEL: Brown Out Detector Trigger Level This controls the voltage trigger level for the Brown out detector. Refer to ”Electrical Characteristics” on page 770. • UPROT, SECURE, BOOTPROT, EPFL, LOCK These are Flash Controller fuses and are described in the FLASHCDW section. 8.9.1.1 Default Fuse Value The devices are shipped with the FGPFRLO register value:0xE075FFFF: • BODEN fuses set to 11. BOD is disabled. • BODHYST fuse set to 1. The BOD hysteresis is enabled. • BODLEVEL fuses set to 000000. This is the minimum voltage trigger level for BOD. This level is lower than the POR level, so when BOD is enabled, it will never trigger with this default value. • UPROT fuse set to 1. • SECURE fuse set to 11. • BOOTPROT fuses set to 010. The bootloader protected size is 8Kbytes. • EPFL fuse set to 1. External privileged fetch is not locked. • LOCK fuses set to 1111111111111111. No region locked. After the JTAG or aWire chip erase command, the FGPFR register value is 0xFFFFFFFF. 100 32099I–01/2012 AT32UC3L016/32/64 8.9.2 First Word of the User Page (Address 0x80800000) 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - WDTAUTO • WDTAUTO: WatchDog Timer Auto Enable at Startup 0: The WDT is automatically enabled at startup. 1: The WDT is not automatically enabled at startup. Please refer to the WDT chapter for detail about timeout settings when the WDT is automatically enabled. 8.9.2.1 Default user page first word value The devices are shipped with the user page erased (all bits 1): • WDTAUTO set to 1, WDT disabled. 101 32099I–01/2012 AT32UC3L016/32/64 8.9.3 Second Word of the User Page (Address 0x80800004) 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 SSADRR[15:8] 23 22 21 20 SSADRR[7:0] 15 14 13 12 SSADRF[15:8] 7 6 5 4 SSADRF[7:0] • SSADRR: Secure State End Address for the RAM • SSADRF: Secure State End Address for the Flash 8.9.3.1 Default user page second word value The devices are shipped with the User page erased (all bits 1). 102 32099I–01/2012 AT32UC3L016/32/64 8.10 Bootloader Configuration The bootloader uses two words in the flash User page to store its configuration: • Configuration word 1 at address 0x808000FC is read first at boot time to know if it should start the ISP process inconditionally and whether it should use the configuration word 2 where further configuration is stored. • Configuration word 2 at address 0x808000F8 stores the I/O conditions that determine which of the ISP and the application to start at the end of the boot process. Please refer to the bootloader documentation for more information. 8.11 Serial Number Each device has a unique 120 bits serial number readable from address 0x8080010C to 0x8080011A. 8.12 Module Configuration The specific configuration for each FLASHCDW instance is listed in the following tables. The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 8-11. Module Configuration Feature AT32UC3L064 AT32UC3L032 AT32UC3L016 Flash size 64Kbytes 32Kbytes 16Kbytes Number of pages 256 128 64 Page size 256 bytes 256 bytes 256 bytes Table 8-12. Module Name Module Clock Name Clock Name Description CLK_FLASHCDW_HSB Clock for the FLASHCDW HSB interface CLK_FLASHCDW_PB Clock for the FLASHCDW PB interface FLASHCDW Table 8-13. Register Reset Values Register Reset Value FVR 0x00000102 FPR 0x00000305 103 32099I–01/2012 AT32UC3L016/32/64 9. Secure Access Unit (SAU) Rev: 1.1.0.3 9.1 Features • Remaps registers in memory regions protected by the MPU to regions not protected by the MPU • Programmable physical address for each channel • Two modes of operation: Locked and Open – In Locked Mode, access to a channel must be preceded by an unlock action • An unlocked channel remains open only for a specific amount of time, if no access is performed during this time, the channel is relocked • Only one channel can be open at a time, opening a channel while another one is open locks the first one • Access to a locked channel is denied, a bus error and optionally an interrupt is returned • If a channel is relocked due to an unlock timeout, an interrupt can optionally be generated – In Open Mode, all channels are permanently unlocked 9.2 Overview In many systems, erroneous access to peripherals can lead to catastrophic failure. An example of such a peripheral is the Pulse Width Modulator (PWM) used to control electric motors. The PWM outputs a pulse train that controls the motor. If the control registers of the PWM module are inadvertently updated with wrong values, the motor can start operating out of control, possibly causing damage to the application and the surrounding environment. However, sometimes the PWM control registers must be updated with new values, for example when modifying the pulse train to accelerate the motor. A mechanism must be used to protect the PWM control registers from inadvertent access caused by for example: • Errors in the software code • Transient errors in the CPU caused by for example electrical noise altering the execution path of the program To improve the security in a computer system, the AVR32UC implements a Memory Protection Unit (MPU). The MPU can be set up to limit the accesses that can be performed to specific memory addresses. The MPU divides the memory space into regions, and assigns a set of access restrictions on each region. Access restrictions can for example be read/write if the CPU is in supervisor mode, and read-only if the CPU is in application mode. The regions can be of different size, but each region is usually quite large, e.g. protecting 1 kilobyte of address space or more. Furthermore, access to each region is often controlled by the execution state of the CPU, i.e. supervisor or application mode. Such a simple control mechanism is often too inflexible (too coarse-grained chunks) and with too much overhead (often requiring system calls to access protected memory locations) for simple or real-time systems such as embedded microcontrollers. Usually, the Secure Access Unit (SAU) is used together with the MPU to provide the required security and integrity. The MPU is set up to protect regions of memory, while the SAU is set up to provide a secure channel into specific memory locations that are protected by the MPU. These specific locations can be thought of as fine-grained overrides of the general coarsegrained protection provided by the MPU. 104 32099I–01/2012 AT32UC3L016/32/64 9.3 Block Diagram Figure 9-1 presents the SAU integrated in an example system with a CPU, some memories, some peripherals, and a bus system. The SAU is connected to both the Peripheral Bus (PB) and the High Speed Bus (HSB). Configuration of the SAU is done via the PB, while memory accesses are done via the HSB. The SAU receives an access on its HSB slave interface, remaps it, checks that the channel is unlocked, and if so, initiates a transfer on its HSB master interface to the remapped address. The thin arrows in Figure 9-1 exemplifies control flow when using the SAU. The CPU wants to read the RX Buffer in the USART. The MPU has been configured to protect all registers in the USART from user mode access, while the SAU has been configured to remap the RX Buffer into a memory space that is not protected by the MPU. This unprotected memory space is mapped into the SAU HSB slave space. When the CPU reads the appropriate address in the SAU, the SAU will perform an access to the desired RX buffer register in the USART, and thereafter return the read results to the CPU. The return data flow will follow the opposite direction of the control flow arrows in Figure 9-1. Figure 9-1. SAU Block Diagram CPU MPU Flash RAM Bus master Bus slave Bus slave High Speed Bus Bus slave Bus master Bus slave Bus bridge SAU Channel SAU Interrupt request USART PWM Peripheral Bus SAU Configuration 105 32099I–01/2012 AT32UC3L016/32/64 9.4 Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 9.4.1 Power Management If the CPU enters a sleep mode that disables clocks used by the SAU, the SAU will stop functioning and resume operation after the system wakes up from sleep mode. 9.4.2 Clocks The SAU has two bus clocks connected: One High Speed Bus clock (CLK_SAU_HSB) and one Peripheral Bus clock (CLK_SAU_PB). These clocks are generated by the Power Manager. Both clocks are enabled at reset, and can be disabled by writing to the Power Manager. The user has to ensure that CLK_SAU_HSB is not turned off before accessing the SAU. Likewise, the user must ensure that no bus access is pending in the SAU before disabling CLK_SAU_HSB. Failing to do so may deadlock the High Speed Bus. 9.4.3 Interrupt The SAU interrupt request line is connected to the interrupt controller. Using the SAU interrupt requires the interrupt controller to be programmed first. 9.4.4 9.5 Debug Operation When an external debugger forces the CPU into debug mode, the SAU continues normal operation. If the SAU is configured in a way that requires it to be periodically serviced by the CPU through interrupts or similar, improper operation or data loss may result during debugging. Functional Description 9.5.1 Enabling the SAU The SAU is enabled by writing a one to the Enable (EN) bit in the Control Register (CR). This will set the SAU Enabled (EN) bit in the Status Register (SR). 9.5.2 Configuring the SAU Channels The SAU has a set of channels, mapped in the HSB memory space. These channels can be configured by a Remap Target Register (RTR), located at the same memory address. When the SAU is in normal mode, the SAU channel is addressed, and when the SAU is in setup mode, the RTR can be addressed. Before the SAU can be used, the channels must be configured and enabled. To configure a channel, the corresponding RTR must be programmed with the Remap Target Address. To do this, make sure the SAU is in setup mode by writing a one to the Setup Mode Enable (SEN) bit in CR. This makes sure that a write to the RTR address accesses the RTR, not the SAU channel. Thereafter, the RTR is written with the address to remap to, typically the address of a specific PB register. When all channels have been configured, return to normal mode by writing a one to the Setup Mode Disable (SDIS) in CR. The channels can now be enabled by writing ones to the corresponding bits in the Channel Enable Registers (CERH/L). The SAU is only able to remap addresses above 0xFFFC0000. 106 32099I–01/2012 AT32UC3L016/32/64 9.5.2.1 9.5.3 Protecting SAU configuration registers In order to prevent the SAU configuration registers to be changed by malicious or runaway code, they should be protected by the MPU as soon as they have been configured. Maximum security is provided in the system if program memory does not contain any code to unprotect the configuration registers in the MPU. This guarantees that runaway code can not accidentally unprotect and thereafter change the SAU configuration registers. Lock Mechanism The SAU can be configured to use two different access mechanisms: Open and Locked. In Open Mode, SAU channels can be accessed freely after they have been configured and enabled. In order to prevent accidental accesses to remapped addresses, it is possible to configure the SAU in Locked Mode. Writing a one to the Open Mode bit in the CONFIG register (CONFIG.OPEN) will enable Open Mode. Writing a zero to CONFIG.OPEN will enable Locked Mode. When using Locked Mode, the lock mechanism must be configured by writing a user defined key value to the Unlock Key (UKEY) field in the Configuration Register (CONFIG). The number of CLK_SAU_HSB cycles the channel remains unlocked must be written to the Unlock Number of Clock Cycles (UCYC) field in CONFIG. Access control to the SAU channels is enabled by means of the Unlock Register (UR), which resides in the same address space as the SAU channels. Before a channel can be accessed, the unlock register must be written with th correct key and channel number (single write access). Access to the channel is then permitted for the next CONFIG.UCYC clock cycles, or until a successful access to the unlocked channel has been made. Only one channel can be unlocked at a time. If any other channel is unlocked at the time of writing UR, this channel will be automatically locked before the channel addressed by the UR write is unlocked. An attempted access to a locked channel will be aborted, and the Channel Access Unsuccessful bit (SR.CAU) will be set. Any pending errors bits in SR must be cleared before it is possible to access UR. The following SR bits are defined as error bits: EXP, CAU, URREAD, URKEY, URES, MBERROR, RTRADR. If any of these bits are set while writing to UR, the write is aborted and the Unlock Register Error Status (URES) bit in SR is set. 9.5.4 Normal Operation The following sequence must be used in order to access a SAU channel in normal operation (CR.SEN=0): 1. If not in Open Mode, write the unlock key to UR.KEY and the channel number to UR.CHANNEL. 2. Perform the read or write operation to the SAU channel. If not in Open Mode, this must be done within CONFIG.UCYC clock cycles of unlocking the channel. The SAU will use its HSB master interface to remap the access to the target address pointed to by the corresponding RTR. 3. To confirm that the access was successful, wait for the IDLE transfer status bit (SR.IDLE) to indicate the operation is completed. Then check SR for possible error conditions. The SAU can be configured to generate interrupt requests or a Bus Error Exception if the access failed. 107 32099I–01/2012 AT32UC3L016/32/64 9.5.4.1 Operation example Figure 9-2 shows a typical memory map, consisting of some memories, some simple peripherals, and a SAU with multiple channels and an Unlock Register (UR). Imagine that the MPU has been set up to disallow all accesses from the CPU to the grey modules. Thus the CPU has no way of accessing for example the Transmit Holding register in the UART, present on address X on the bus. Note that the SAU RTRs are not protected by the MPU, thus the RTRs can be accessed. If for example RTR0 is configured to point to address X, an access to RTR0 will be remapped by the SAU to address X according to the algorithm presented above. By programming the SAU RTRs, specific addresses in modules that have generally been protected by the MPU can be performed. Figure 9-2. Example Memory Map for a System with SAU SAU CONFIG UART Receive Holding Transmit Holding Baudrate Control Address X SAU CHANNEL 9.5.5 ... UR RTR62 Channel RTR1 1 RTR0 Address Z Interrupts The SAU can generate an interrupt request to signal different events. All events that can generate an interrupt request have dedicated bits in the Status Register (SR). An interrupt request will be generated if the corresponding bit in the Interrupt Mask Register (IMR) is set. Bits in IMR are set by writing a one to the corresponding bit in the Interrupt Enable Register (IER), and cleared by writing a one to the corresponding bit in the Interrupt Disable Register (IDR). The interrupt request remains active until the corresponding bit in SR is cleared by writing a one to the corresponding bit in the Interrupt Clear Register (ICR). The following SR bits are used for signalling the result of SAU accesses: • RTR Address Error (RTRADR) is set if an illegal address is written to the RTRs. Only addresses in the range 0xFFFC0000-0xFFFFFFFF are allowed. • Master Interface Bus Error (MBERROR) is set if any of the conditions listed in Section 9.5.7 occurred. 108 32099I–01/2012 AT32UC3L016/32/64 • Unlock Register Error Status (URES) is set if an attempt was made to unlock a channel by writing to the Unlock Register while one or more error bits in SR were set (see Section 9.5.6). The unlock operation was aborted. • Unlock Register Key Error (URKEY) is set if the Unlock Register was attempted written with an invalid key. • Unlock Register Read (URREAD) is set if the Unlock Register was attempted read. • Channel Access Unsuccessful (CAU) is set if the channel access was unsuccessful. • Channel Access Successful (CAS) is set if the channel access was successful. • Channel Unlock Expired (EXP) is set if the channel lock expired, with no channel being accessed after the channel was unlocked. 9.5.6 Error bits If error bits are set when attempting to unlock a channel, SR.URES will be set. The following SR bits are considered error bits: • EXP • CAU • URREAD • URKEY • URES • MBERROR • RTRADR 9.5.7 Bus Error Responses By writing a one to the Bus Error Response Enable bit (CR.BERREN), serious access errors will be configured to return a bus error to the CPU. This will cause the CPU to execute its Bus Error Data Fetch exception routine. The conditions that can generate a bus error response are: • Reading the Unlock Register • Trying to access a locked channel • The SAU HSB master receiving a bus error response from its addressed slave 9.5.8 Disabling the SAU To disable the SAU, the user must first ensure that no SAU bus operations are pending. This can be done by checking that the SR.IDLE bit is set. The SAU may then be disabled by writing a one to the Disable (DIS) bit in CR. 109 32099I–01/2012 AT32UC3L016/32/64 9.6 User Interface The following addresses are used by SAU channel configuration registers. All offsets are relative to the SAU’s PB base address. Table 9-1. SAU Configuration Register Memory Map Offset Register Register Name Access Reset 0x00 Control Register CR Write-only 0x00000000 0x04 Configuration Register CONFIG Write-only 0x00000000 0x08 Channel Enable Register High CERH Read/Write 0x00000000 0x0C Channel Enable Register Low CERL Read/Write 0x00000000 0x10 Status Register SR Read-only 0x00000400 0x14 Interrupt Enable Register IER Write-only 0x00000000 0x18 Interrupt Disable Register IDR Write-only 0x00000000 0x1C Interrupt Mask Register IMR Read-only 0x00000000 0x20 Interrupt Clear Register ICR Write-only 0x00000000 0x24 Parameter Register PARAMETER Read-only -(1) 0x28 Version Register VERSION Read-only -(1) 1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter. Note: The following addresses are used by SAU channel registers. All offsets are relative to the SAU’s HSB base address. The number of channels implemented is device specific, refer to the Module Configuration section at the end of this chapter. Table 9-2. SAU Channel Register Memory Map Offset Register Register Name Access Reset 0x00 Remap Target Register 0 RTR0 Read/Write N/A 0x04 Remap Target Register 1 RTR1 Read/Write N/A 0x08 Remap Target Register 2 RTR2 Read/Write N/A ... ... ... ... ... 0x04*n Remap Target Register n RTRn Read/Write N/A 0xFC Unlock Register UR Write-only N/A 110 32099I–01/2012 AT32UC3L016/32/64 9.6.1 Name: Control Register CR Access Type: Write-only Offset: 0x00 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - BERRDIS BERREN SDIS SEN DIS EN • BERRDIS: Bus Error Response Disable Writing a zero to this bit has no effect. Writing a one to this bit disables Bus Error Response from the SAU. • BERREN: Bus Error Response Enable Writing a zero to this bit has no effect. Writing a one to this bit enables Bus Error Response from the SAU. • SDIS: Setup Mode Disable Writing a zero to this bit has no effect. Writing a one to this bit exits setup mode. • SEN: Setup Mode Enable Writing a zero to this bit has no effect. Writing a one to this bit enters setup mode. • DIS: SAU Disable Writing a zero to this bit has no effect. Writing a one to this bit disables the SAU. • EN: SAU Enable Writing a zero to this bit has no effect. Writing a one to this bit enables the SAU. 111 32099I–01/2012 AT32UC3L016/32/64 9.6.2 Name: Configuration Register CONFIG Access Type: Write-only Offset: 0x04 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - OPEN 15 14 13 12 11 10 9 8 3 2 1 0 UCYC 7 6 5 4 UKEY • OPEN: Open Mode Enable Writing a zero to this bit disables open mode. Writing a one to this bit enables open mode. • UCYC: Unlock Number of Clock Cycles Once a channel has been unlocked, it remains unlocked for this amount of CLK_SAU_HSB clock cycles or until one access to a channel has been made. • UKEY: Unlock Key The value in this field must be written to UR.KEY to unlock a channel. 112 32099I–01/2012 AT32UC3L016/32/64 9.6.3 Name: Channel Enable Register High CERH Access Type: Read/Write Offset: 0x08 Reset Value: 0x00000000 31 30 29 28 - 23 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 CERH[30:24] 22 21 20 CERH[23:16] 15 14 13 12 CERH[15:8] 7 6 5 4 CERH[7:0] • CERH[n]: Channel Enable Register High 0: Channel (n+32) is not enabled. 1: Channel (n+32) is enabled. 113 32099I–01/2012 AT32UC3L016/32/64 9.6.4 Name: Channel Enable Register Low CERL Access Type: Read/Write Offset: 0x0C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 CERL[31:24] 23 22 21 20 CERL[23:16] 15 14 13 12 CERL[15:8] 7 6 5 4 CERL[7:0] • CERL[n]: Channel Enable Register Low 0: Channel n is not enabled. 1: Channel n is enabled. 114 32099I–01/2012 AT32UC3L016/32/64 9.6.5 Name: Status Register SR Access Type: Read-only Offset: 0x10 Reset Value: 0x00000400 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - IDLE SEN EN 7 6 5 4 3 2 1 0 RTRADR MBERROR URES URKEY URREAD CAU CAS EXP • IDLE • • • • • • • This bit is cleared when a read or write operation to the SAU channel is started. This bit is set when the operation is completed and no SAU bus operations are pending. SEN: SAU Setup Mode Enable This bit is cleared when the SAU exits setup mode. This bit is set when the SAU enters setup mode. EN: SAU Enabled This bit is cleared when the SAU is disabled. This bit is set when the SAU is enabled. RTRADR: RTR Address Error This bit is cleared when the corresponding bit in ICR is written to one. This bit is set if, in the configuration phase, an RTR was written with an illegal address, i.e. the upper 16 bits in the address were different from 0xFFFC, 0xFFFD, 0xFFFE or 0xFFFF. MBERROR: Master Interface Bus Error This bit is cleared when the corresponding bit in ICR is written to one. This bit is set if a channel access generated a transfer on the master interface that received a bus error response from the addressed slave. URES: Unlock Register Error Status This bit is cleared when the corresponding bit in ICR is written to one. This bit is set if an attempt was made to unlock a channel by writing to the Unlock Register while one or more error bits were set in SR. The unlock operation was aborted. URKEY: Unlock Register Key Error This bit is cleared when the corresponding bit in ICR is written to one. This bit is set if the Unlock Register was attempted written with an invalid key. URREAD: Unlock Register Read This bit is cleared when the corresponding bit in ICR is written to one. This bit is set if the Unlock Register was read. 115 32099I–01/2012 AT32UC3L016/32/64 • CAU: Channel Access Unsuccessful This bit is cleared when the corresponding bit in ICR is written to one. This bit is set if channel access was unsuccessful, i.e. an access was attempted to a locked or disabled channel. • CAS: Channel Access Successful This bit is cleared when the corresponding bit in ICR is written to one. This bit is set if channel access successful, i.e. one access was made after the channel was unlocked. • EXP: Channel Unlock Expired This bit is cleared when the corresponding bit in ICR is written to one. This bit is set if channel unlock has expired, i.e. no access being made after the channel was unlocked. 116 32099I–01/2012 AT32UC3L016/32/64 9.6.6 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x14 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 RTRADR MBERROR URES URKEY URREAD CAU CAS EXP Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 117 32099I–01/2012 AT32UC3L016/32/64 9.6.7 Name: Interrupt Disable Register IDR Access Type: Write-only Offset: 0x18 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 RTRADR MBERROR URES URKEY URREAD CAU CAS EXP Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 118 32099I–01/2012 AT32UC3L016/32/64 9.6.8 Name: Interrupt Mask Register IMR Access Type: Read-only Offset: 0x1C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 RTRADR MBERROR URES URKEY URREAD CAU CAS EXP 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. A bit in this register is cleared when the corresponding bit in IDR is written to one. A bit in this register is set when the corresponding bit in IER is written to one. 119 32099I–01/2012 AT32UC3L016/32/64 9.6.9 Name: Interrupt Clear Register ICR Access Type: Write-only Offset: 0x20 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 RTRADR MBERROR URES URKEY URREAD CAU CAS EXP Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in SR and any corresponding interrupt request. 120 32099I–01/2012 AT32UC3L016/32/64 9.6.10 Name: Parameter Register PARAMETER Access Type: Read-only Offset: 0x24 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - CHANNELS • CHANNELS: Number of channels implemented. 121 32099I–01/2012 AT32UC3L016/32/64 9.6.11 Name: Version Register VERSION Access Type: Write-only Offset: 0x28 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 1 0 VARIANT 11 10 VERSION[11:8] 3 2 VERSION[7:0] • VARIANT: Variant Number Reserved. No functionality associated. • VERSION: Version Number Version number of the module. No functionality associated. 122 32099I–01/2012 AT32UC3L016/32/64 9.6.12 Name: Remap Target Register n RTRn Access Type: Read/Write Offset: n*4 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 RTR[31:24] 23 22 21 20 RTR[23:16] 15 14 13 12 RTR[15:8] 7 6 5 4 RTR[7:0] • RTR: Remap Target Address for Channel n RTR[31:16] must have one of the following values, any other value will result in UNDEFINED behavior: 0xFFFC 0xFFFD 0xFFFE 0xFFFF RTR[1:0] must be written to 00, any other value will result in UNDEFINED behavior. 123 32099I–01/2012 AT32UC3L016/32/64 9.6.13 Name: Unlock Register UR Access Type : Write-only Offset: 0xFC Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 KEY 7 6 - - 5 4 CHANNEL • KEY: Unlock Key The correct key must be written in order to unlock a channel. The key value written must correspond to the key value defined in CONFIG.UKEY. • CHANNEL: Channel Number Number of the channel to unlock. 124 32099I–01/2012 AT32UC3L016/32/64 9.7 Module Configuration The specific configuration for each SAU instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 9-3. Module configuration Feature SAU SAU Channels 16 Table 9-4. Module clock name Module name Clock name Description SAU CLK_SAU_HSB Clock for the SAU HSB interface SAU CLK_SAU_PB Clock for the SAU PB interface Table 9-5. Register Reset Values Register Reset Value VERSION 0x00000110 PARAMETER 0x00000010 125 32099I–01/2012 AT32UC3L016/32/64 10. HSB Bus Matrix (HMATRIXB) Rev: 1.3.0.3 10.1 Features • • • • • • • • • 10.2 User Interface on peripheral bus Configurable number of masters (up to 16) Configurable number of slaves (up to 16) One decoder for each master Programmable arbitration for each slave – Round-Robin – Fixed priority Programmable default master for each slave – No default master – Last accessed default master – Fixed default master One cycle latency for the first access of a burst Zero cycle latency for default master One special function register for each slave (not dedicated) Overview The Bus Matrix implements a multi-layer bus structure, that enables parallel access paths between multiple High Speed Bus (HSB) masters and slaves in a system, thus increasing the overall bandwidth. The Bus Matrix interconnects up to 16 HSB Masters to up to 16 HSB Slaves. The normal latency to connect a master to a slave is one cycle except for the default master of the accessed slave which is connected directly (zero cycle latency). The Bus Matrix provides 16 Special Function Registers (SFR) that allow the Bus Matrix to support application specific features. 10.3 Product Dependencies In order to configure this module by accessing the user registers, other parts of the system must be configured correctly, as described below. 10.3.1 Clocks The clock for the HMATRIX bus interface (CLK_HMATRIX) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. 10.4 10.4.1 Functional Description Special Bus Granting Mechanism The Bus Matrix provides some speculative bus granting techniques in order to anticipate access requests from some masters. This mechanism reduces latency at first access of a burst or single transfer. This bus granting mechanism sets a different default master for every slave. At the end of the current access, if no other request is pending, the slave remains connected to its associated default master. A slave can be associated with three kinds of default masters: no default master, last access master, and fixed default master. 126 32099I–01/2012 AT32UC3L016/32/64 To change from one kind of default master to another, the Bus Matrix user interface provides the Slave Configuration Registers, one for each slave, that set a default master for each slave. The Slave Configuration Register contains two fields: DEFMSTR_TYPE and FIXED_DEFMSTR. The 2-bit DEFMSTR_TYPE field selects the default master type (no default, last access master, fixed default master), whereas the 4-bit FIXED_DEFMSTR field selects a fixed default master provided that DEFMSTR_TYPE is set to fixed default master. Please refer to the Bus Matrix user interface description. 10.4.1.1 No Default Master At the end of the current access, if no other request is pending, the slave is disconnected from all masters. No Default Master suits low-power mode. 10.4.1.2 Last Access Master At the end of the current access, if no other request is pending, the slave remains connected to the last master that performed an access request. 10.4.1.3 Fixed Default Master At the end of the current access, if no other request is pending, the slave connects to its fixed default master. Unlike last access master, the fixed master does not change unless the user modifies it by a software action (field FIXED_DEFMSTR of the related SCFG). 10.4.2 Arbitration The Bus Matrix provides an arbitration mechanism that reduces latency when conflict cases occur, i.e. when two or more masters try to access the same slave at the same time. One arbiter per HSB slave is provided, thus arbitrating each slave differently. The Bus Matrix provides the user with the possibility of choosing between 2 arbitration types for each slave: 1. Round-Robin Arbitration (default) 2. Fixed Priority Arbitration This is selected by the ARBT field in the Slave Configuration Registers (SCFG). Each algorithm may be complemented by selecting a default master configuration for each slave. When a re-arbitration must be done, specific conditions apply. This is described in “Arbitration Rules” . 10.4.2.1 Arbitration Rules Each arbiter has the ability to arbitrate between two or more different master requests. In order to avoid burst breaking and also to provide the maximum throughput for slave interfaces, arbitration may only take place during the following cycles: 1. Idle Cycles: When a slave is not connected to any master or is connected to a master which is not currently accessing it. 2. Single Cycles: When a slave is currently doing a single access. 3. End of Burst Cycles: When the current cycle is the last cycle of a burst transfer. For defined length burst, predicted end of burst matches the size of the transfer but is managed differently for undefined length burst. This is described below. 4. Slot Cycle Limit: When the slot cycle counter has reached the limit value indicating that the current master access is too long and must be broken. This is described below. 127 32099I–01/2012 AT32UC3L016/32/64 • Undefined Length Burst Arbitration In order to avoid long slave handling during undefined length bursts (INCR), the Bus Matrix provides specific logic in order to re-arbitrate before the end of the INCR transfer. A predicted end of burst is used as a defined length burst transfer and can be selected among the following five possibilities: 1. Infinite: No predicted end of burst is generated and therefore INCR burst transfer will never be broken. 2. One beat bursts: Predicted end of burst is generated at each single transfer inside the INCP transfer. 3. Four beat bursts: Predicted end of burst is generated at the end of each four beat boundary inside INCR transfer. 4. Eight beat bursts: Predicted end of burst is generated at the end of each eight beat boundary inside INCR transfer. 5. Sixteen beat bursts: Predicted end of burst is generated at the end of each sixteen beat boundary inside INCR transfer. This selection can be done through the ULBT field in the Master Configuration Registers (MCFG). • Slot Cycle Limit Arbitration The Bus Matrix contains specific logic to break long accesses, such as very long bursts on a very slow slave (e.g., an external low speed memory). At the beginning of the burst access, a counter is loaded with the value previously written in the SLOT_CYCLE field of the related Slave Configuration Register (SCFG) and decreased at each clock cycle. When the counter reaches zero, the arbiter has the ability to re-arbitrate at the end of the current byte, halfword, or word transfer. 10.4.2.2 Round-Robin Arbitration This algorithm allows the Bus Matrix arbiters to dispatch the requests from different masters to the same slave in a round-robin manner. If two or more master requests arise at the same time, the master with the lowest number is first serviced, then the others are serviced in a round-robin manner. There are three round-robin algorithms implemented: 1. Round-Robin arbitration without default master 2. Round-Robin arbitration with last default master 3. Round-Robin arbitration with fixed default master • Round-Robin Arbitration without Default Master This is the main algorithm used by Bus Matrix arbiters. It allows the Bus Matrix to dispatch requests from different masters to the same slave in a pure round-robin manner. At the end of the current access, if no other request is pending, the slave is disconnected from all masters. This configuration incurs one latency cycle for the first access of a burst. Arbitration without default master can be used for masters that perform significant bursts. • Round-Robin Arbitration with Last Default Master This is a biased round-robin algorithm used by Bus Matrix arbiters. It allows the Bus Matrix to remove the one latency cycle for the last master that accessed the slave. At the end of the cur128 32099I–01/2012 AT32UC3L016/32/64 rent transfer, if no other master request is pending, the slave remains connected to the last master that performed the access. Other non privileged masters still get one latency cycle if they want to access the same slave. This technique can be used for masters that mainly perform single accesses. • Round-Robin Arbitration with Fixed Default Master This is another biased round-robin algorithm. It allows the Bus Matrix arbiters to remove the one latency cycle for the fixed default master per slave. At the end of the current access, the slave remains connected to its fixed default master. Every request attempted by this fixed default master will not cause any latency whereas other non privileged masters will still get one latency cycle. This technique can be used for masters that mainly perform single accesses. 10.4.2.3 Fixed Priority Arbitration This algorithm allows the Bus Matrix arbiters to dispatch the requests from different masters to the same slave by using the fixed priority defined by the user. If two or more master requests are active at the same time, the master with the highest priority number is serviced first. If two or more master requests with the same priority are active at the same time, the master with the highest number is serviced first. For each slave, the priority of each master may be defined through the Priority Registers for Slaves (PRAS and PRBS). 10.4.3 Slave and Master assignation The index number assigned to Bus Matrix slaves and masters are described in the Module Configuration section at the end of this chapter. 129 32099I–01/2012 AT32UC3L016/32/64 10.5 User Interface Table 10-1. HMATRIX Register Memory Map Offset Register Name Access Reset Value 0x0000 Master Configuration Register 0 MCFG0 Read/Write 0x00000002 0x0004 Master Configuration Register 1 MCFG1 Read/Write 0x00000002 0x0008 Master Configuration Register 2 MCFG2 Read/Write 0x00000002 0x000C Master Configuration Register 3 MCFG3 Read/Write 0x00000002 0x0010 Master Configuration Register 4 MCFG4 Read/Write 0x00000002 0x0014 Master Configuration Register 5 MCFG5 Read/Write 0x00000002 0x0018 Master Configuration Register 6 MCFG6 Read/Write 0x00000002 0x001C Master Configuration Register 7 MCFG7 Read/Write 0x00000002 0x0020 Master Configuration Register 8 MCFG8 Read/Write 0x00000002 0x0024 Master Configuration Register 9 MCFG9 Read/Write 0x00000002 0x0028 Master Configuration Register 10 MCFG10 Read/Write 0x00000002 0x002C Master Configuration Register 11 MCFG11 Read/Write 0x00000002 0x0030 Master Configuration Register 12 MCFG12 Read/Write 0x00000002 0x0034 Master Configuration Register 13 MCFG13 Read/Write 0x00000002 0x0038 Master Configuration Register 14 MCFG14 Read/Write 0x00000002 0x003C Master Configuration Register 15 MCFG15 Read/Write 0x00000002 0x0040 Slave Configuration Register 0 SCFG0 Read/Write 0x00000010 0x0044 Slave Configuration Register 1 SCFG1 Read/Write 0x00000010 0x0048 Slave Configuration Register 2 SCFG2 Read/Write 0x00000010 0x004C Slave Configuration Register 3 SCFG3 Read/Write 0x00000010 0x0050 Slave Configuration Register 4 SCFG4 Read/Write 0x00000010 0x0054 Slave Configuration Register 5 SCFG5 Read/Write 0x00000010 0x0058 Slave Configuration Register 6 SCFG6 Read/Write 0x00000010 0x005C Slave Configuration Register 7 SCFG7 Read/Write 0x00000010 0x0060 Slave Configuration Register 8 SCFG8 Read/Write 0x00000010 0x0064 Slave Configuration Register 9 SCFG9 Read/Write 0x00000010 0x0068 Slave Configuration Register 10 SCFG10 Read/Write 0x00000010 0x006C Slave Configuration Register 11 SCFG11 Read/Write 0x00000010 0x0070 Slave Configuration Register 12 SCFG12 Read/Write 0x00000010 0x0074 Slave Configuration Register 13 SCFG13 Read/Write 0x00000010 0x0078 Slave Configuration Register 14 SCFG14 Read/Write 0x00000010 0x007C Slave Configuration Register 15 SCFG15 Read/Write 0x00000010 0x0080 Priority Register A for Slave 0 PRAS0 Read/Write 0x00000000 0x0084 Priority Register B for Slave 0 PRBS0 Read/Write 0x00000000 0x0088 Priority Register A for Slave 1 PRAS1 Read/Write 0x00000000 130 32099I–01/2012 AT32UC3L016/32/64 Table 10-1. HMATRIX Register Memory Map (Continued) Offset Register Name Access Reset Value 0x008C Priority Register B for Slave 1 PRBS1 Read/Write 0x00000000 0x0090 Priority Register A for Slave 2 PRAS2 Read/Write 0x00000000 0x0094 Priority Register B for Slave 2 PRBS2 Read/Write 0x00000000 0x0098 Priority Register A for Slave 3 PRAS3 Read/Write 0x00000000 0x009C Priority Register B for Slave 3 PRBS3 Read/Write 0x00000000 0x00A0 Priority Register A for Slave 4 PRAS4 Read/Write 0x00000000 0x00A4 Priority Register B for Slave 4 PRBS4 Read/Write 0x00000000 0x00A8 Priority Register A for Slave 5 PRAS5 Read/Write 0x00000000 0x00AC Priority Register B for Slave 5 PRBS5 Read/Write 0x00000000 0x00B0 Priority Register A for Slave 6 PRAS6 Read/Write 0x00000000 0x00B4 Priority Register B for Slave 6 PRBS6 Read/Write 0x00000000 0x00B8 Priority Register A for Slave 7 PRAS7 Read/Write 0x00000000 0x00BC Priority Register B for Slave 7 PRBS7 Read/Write 0x00000000 0x00C0 Priority Register A for Slave 8 PRAS8 Read/Write 0x00000000 0x00C4 Priority Register B for Slave 8 PRBS8 Read/Write 0x00000000 0x00C8 Priority Register A for Slave 9 PRAS9 Read/Write 0x00000000 0x00CC Priority Register B for Slave 9 PRBS9 Read/Write 0x00000000 0x00D0 Priority Register A for Slave 10 PRAS10 Read/Write 0x00000000 0x00D4 Priority Register B for Slave 10 PRBS10 Read/Write 0x00000000 0x00D8 Priority Register A for Slave 11 PRAS11 Read/Write 0x00000000 0x00DC Priority Register B for Slave 11 PRBS11 Read/Write 0x00000000 0x00E0 Priority Register A for Slave 12 PRAS12 Read/Write 0x00000000 0x00E4 Priority Register B for Slave 12 PRBS12 Read/Write 0x00000000 0x00E8 Priority Register A for Slave 13 PRAS13 Read/Write 0x00000000 0x00EC Priority Register B for Slave 13 PRBS13 Read/Write 0x00000000 0x00F0 Priority Register A for Slave 14 PRAS14 Read/Write 0x00000000 0x00F4 Priority Register B for Slave 14 PRBS14 Read/Write 0x00000000 0x00F8 Priority Register A for Slave 15 PRAS15 Read/Write 0x00000000 0x00FC Priority Register B for Slave 15 PRBS15 Read/Write 0x00000000 0x0110 Special Function Register 0 SFR0 Read/Write – 0x0114 Special Function Register 1 SFR1 Read/Write – 0x0118 Special Function Register 2 SFR2 Read/Write – 0x011C Special Function Register 3 SFR3 Read/Write – 0x0120 Special Function Register 4 SFR4 Read/Write – 0x0124 Special Function Register 5 SFR5 Read/Write – 0x0128 Special Function Register 6 SFR6 Read/Write – 131 32099I–01/2012 AT32UC3L016/32/64 Table 10-1. HMATRIX Register Memory Map (Continued) Offset Register Name Access Reset Value 0x012C Special Function Register 7 SFR7 Read/Write – 0x0130 Special Function Register 8 SFR8 Read/Write – 0x0134 Special Function Register 9 SFR9 Read/Write – 0x0138 Special Function Register 10 SFR10 Read/Write – 0x013C Special Function Register 11 SFR11 Read/Write – 0x0140 Special Function Register 12 SFR12 Read/Write – 0x0144 Special Function Register 13 SFR13 Read/Write – 0x0148 Special Function Register 14 SFR14 Read/Write – 0x014C Special Function Register 15 SFR15 Read/Write – 132 32099I–01/2012 AT32UC3L016/32/64 10.5.1 Name: Master Configuration Registers MCFG0...MCFG15 Access Type: Read/Write Offset: 0x00 - 0x3C Reset Value: 0x00000002 31 30 29 28 27 26 25 24 – – – – – – – – 23 22 21 20 19 18 17 16 – – – – – – – – 15 14 13 12 11 10 9 8 – – – – – – – – 7 6 5 4 3 2 1 0 – – – – – ULBT • ULBT: Undefined Length Burst Type Table 10-2. Undefined Length Burst Type ULBT Undefined Length Burst Type Description 000 Inifinite Length Burst No predicted end of burst is generated and therefore INCR bursts coming from this master cannot be broken. 001 Single-Access The undefined length burst is treated as a succession of single accesses, allowing rearbitration at each beat of the INCR burst. 010 4 Beat Burst The undefined length burst is split into a four-beat burst, allowing re-arbitration at each four-beat burst end. 011 8 Beat Burst The undefined length burst is split into an eight-beat burst, allowing re-arbitration at each eight-beat burst end. 100 16 Beat Burst The undefined length burst is split into a sixteen-beat burst, allowing re-arbitration at each sixteen-beat burst end. 133 32099I–01/2012 AT32UC3L016/32/64 10.5.2 Name: Slave Configuration Registers SCFG0...SCFG15 Access Type: Read/Write Offset: 0x40 - 0x7C Reset Value: 0x00000010 31 30 29 28 27 26 25 24 – – – – – – – ARBT 23 22 21 20 19 18 17 16 – – 15 14 13 12 11 10 9 8 – – – – – – – – 7 6 5 4 3 2 1 0 FIXED_DEFMSTR DEFMSTR_TYPE SLOT_CYCLE • ARBT: Arbitration Type 0: Round-Robin Arbitration 1: Fixed Priority Arbitration • FIXED_DEFMSTR: Fixed Default Master This is the number of the Default Master for this slave. Only used if DEFMSTR_TYPE is 2. Specifying the number of a master which is not connected to the selected slave is equivalent to setting DEFMSTR_TYPE to 0. • DEFMSTR_TYPE: Default Master Type 0: No Default Master At the end of the current slave access, if no other master request is pending, the slave is disconnected from all masters. This results in a one cycle latency for the first access of a burst transfer or for a single access. 1: Last Default Master At the end of the current slave access, if no other master request is pending, the slave stays connected to the last master having accessed it. This results in not having one cycle latency when the last master tries to access the slave again. 2: Fixed Default Master At the end of the current slave access, if no other master request is pending, the slave connects to the fixed master the number that has been written in the FIXED_DEFMSTR field. This results in not having one cycle latency when the fixed master tries to access the slave again. • SLOT_CYCLE: Maximum Number of Allowed Cycles for a Burst When the SLOT_CYCLE limit is reached for a burst, it may be broken by another master trying to access this slave. This limit has been placed to avoid locking a very slow slave when very long bursts are used. This limit must not be very small. Unreasonably small values break every burst and the Bus Matrix arbitrates without performing any data transfer. 16 cycles is a reasonable value for SLOT_CYCLE. 134 32099I–01/2012 AT32UC3L016/32/64 10.5.3 Bus Matrix Priority Registers A For Slaves Register Name: PRAS0...PRAS15 Access Type: Read/Write Offset: - Reset Value: 0x00000000 31 30 - - 23 22 - - 15 14 - - 7 6 - - 29 28 M7PR 21 20 M5PR 13 12 M3PR 5 4 M1PR 27 26 - - 19 18 - - 11 10 - - 3 2 - - 25 24 M6PR 17 16 M4PR 9 8 M2PR 1 0 M0PR • MxPR: Master x Priority Fixed priority of Master x for accessing the selected slave. The higher the number, the higher the priority. 135 32099I–01/2012 AT32UC3L016/32/64 10.5.4 Name: Priority Registers B For Slaves PRBS0...PRBS15 Access Type: Read/Write Offset: - Reset Value: 0x00000000 31 30 - - 23 22 - - 15 14 - - 7 6 - - 29 28 M15PR 21 20 M13PR 13 12 M11PR 5 4 M9PR 27 26 - - 19 18 - - 11 10 - - 3 2 - - 25 24 M14PR 17 16 M12PR 9 8 M10PR 1 0 M8PR • MxPR: Master x Priority Fixed priority of Master x for accessing the selected slave. The higher the number, the higher the priority. 136 32099I–01/2012 AT32UC3L016/32/64 10.5.5 Name: Special Function Registers SFR0...SFR15 Access Type: Read/Write Offset: 0x110 - 0x14C Reset Value: - 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 SFR 23 22 21 20 SFR 15 14 13 12 SFR 7 6 5 4 SFR • SFR: Special Function Register Fields Those registers are not a HMATRIX specific register. The field of those will be defined where they are used. 137 32099I–01/2012 AT32UC3L016/32/64 10.6 Module Configuration The specific configuration for each HMATRIX instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 10-3. 10.6.1 HMATRIX Clocks Clock Name Description CLK_HMATRIX Clock for the HMATRIX bus interface Bus Matrix Connections The bus matrix has the several masters and slaves. Each master has its own bus and its own decoder, thus allowing a different memory mapping per master. The master number in the table below can be used to index the HMATRIX control registers. For example, HMATRIX MCFG0 register is associated with the CPU Data master interface. Table 10-4. High Speed Bus Masters Master 0 CPU Data Master 1 CPU Instruction Master 2 CPU SAB Master 3 SAU Master 4 PDCA Each slave has its own arbiter, thus allowing a different arbitration per slave. The slave number in the table below can be used to index the HMATRIX control registers. For example, SCFG3 is associated with the Internal SRAM Slave Interface. Accesses to unused areas returns an error result to the master requesting such an access. Table 10-5. High Speed Bus Slaves Slave 0 Internal Flash Slave 1 HSB-PB Bridge A Slave 2 HSB-PB Bridge B Slave 3 Internal SRAM Slave 4 SAU 138 32099I–01/2012 AT32UC3L016/32/64 Figure 10-1. HMatrix Master / Slave Connections HMATRIX MASTERS CPU Data 0 CPU Instruction 1 CPU SAB 2 SAU 3 PDCA 4 Internal Flash HSB-PB Bridge A HSB-PB Bridge B Internal SRAM SAU HMATRIX SLAVES 0 1 2 3 4 139 32099I–01/2012 AT32UC3L016/32/64 11. Interrupt Controller (INTC) Rev: 1.0.2.5 11.1 Features • Autovectored low latency interrupt service with programmable priority – 4 priority levels for regular, maskable interrupts – One Non-Maskable Interrupt • Up to 64 groups of interrupts with up to 32 interrupt requests in each group 11.2 Overview The INTC collects interrupt requests from the peripherals, prioritizes them, and delivers an interrupt request and an autovector to the CPU. The AVR32 architecture supports 4 priority levels for regular, maskable interrupts, and a Non-Maskable Interrupt (NMI). The INTC supports up to 64 groups of interrupts. Each group can have up to 32 interrupt request lines, these lines are connected to the peripherals. Each group has an Interrupt Priority Register (IPR) and an Interrupt Request Register (IRR). The IPRs are used to assign a priority level and an autovector to each group, and the IRRs are used to identify the active interrupt request within each group. If a group has only one interrupt request line, an active interrupt group uniquely identifies the active interrupt request line, and the corresponding IRR is not needed. The INTC also provides one Interrupt Cause Register (ICR) per priority level. These registers identify the group that has a pending interrupt of the corresponding priority level. If several groups have a pending interrupt of the same level, the group with the lowest number takes priority. 11.3 Block Diagram Figure 11-1 gives an overview of the INTC. The grey boxes represent registers that can be accessed via the user interface. The interrupt requests from the peripherals (IREQn) and the NMI are input on the left side of the figure. Signals to and from the CPU are on the right side of the figure. 140 32099I–01/2012 AT32UC3L016/32/64 Figure 11-1. INTC Block Diagram Interrupt Controller CPU NMIREQ Masks OR IRRn GrpReqN IREQ63 IREQ34 IREQ33 IREQ32 OR GrpReq1 INT_level, offset IPRn . . . Request Masking ValReq1 INT_level, offset IPR1 . . . INTLEVEL Prioritizer . . . ValReqN SREG Masks I[3-0]M GM AUTOVECTOR IRR1 IREQ31 IREQ2 IREQ1 IREQ0 OR GrpReq0 ValReq0 IPR0 INT_level, offset IRR0 IRR Registers 11.4 IPR Registers ICR Registers Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 11.4.1 Power Management If the CPU enters a sleep mode that disables CLK_SYNC, the INTC will stop functioning and resume operation after the system wakes up from sleep mode. 11.4.2 Clocks The clock for the INTC bus interface (CLK_INTC) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. The INTC sampling logic runs on a clock which is stopped in any of the sleep modes where the system RC oscillator is not running. This clock is referred to as CLK_SYNC. This clock is enabled at reset, and only turned off in sleep modes where the system RC oscillator is stopped. 11.4.3 11.5 Debug Operation When an external debugger forces the CPU into debug mode, the INTC continues normal operation. Functional Description All of the incoming interrupt requests (IREQs) are sampled into the corresponding Interrupt Request Register (IRR). The IRRs must be accessed to identify which IREQ within a group that is active. If several IREQs within the same group are active, the interrupt service routine must prioritize between them. All of the input lines in each group are logically ORed together to form the GrpReqN lines, indicating if there is a pending interrupt in the corresponding group. The Request Masking hardware maps each of the GrpReq lines to a priority level from INT0 to INT3 by associating each group with the Interrupt Level (INTLEVEL) field in the corresponding 141 32099I–01/2012 AT32UC3L016/32/64 Interrupt Priority Register (IPR). The GrpReq inputs are then masked by the mask bits from the CPU status register. Any interrupt group that has a pending interrupt of a priority level that is not masked by the CPU status register, gets its corresponding ValReq line asserted. Masking of the interrupt requests is done based on five interrupt mask bits of the CPU status register, namely Interrupt Level 3 Mask (I3M) to Interrupt Level 0 Mask (I0M), and Global Interrupt Mask (GM). An interrupt request is masked if either the GM or the corresponding interrupt level mask bit is set. The Prioritizer hardware uses the ValReq lines and the INTLEVEL field in the IPRs to select the pending interrupt of the highest priority. If an NMI interrupt request is pending, it automatically gets the highest priority of any pending interrupt. If several interrupt groups of the highest pending interrupt level have pending interrupts, the interrupt group with the lowest number is selected. The INTLEVEL and handler autovector offset (AUTOVECTOR) of the selected interrupt are transmitted to the CPU for interrupt handling and context switching. The CPU does not need to know which interrupt is requesting handling, but only the level and the offset of the handler address. The IRR registers contain the interrupt request lines of the groups and can be read via user interface registers for checking which interrupts of the group are actually active. The delay through the INTC from the peripheral interrupt request is set until the interrupt request to the CPU is set is three cycles of CLK_SYNC. 11.5.1 Non-Maskable Interrupts A NMI request has priority over all other interrupt requests. NMI has a dedicated exception vector address defined by the AVR32 architecture, so AUTOVECTOR is undefined when INTLEVEL indicates that an NMI is pending. 11.5.2 CPU Response When the CPU receives an interrupt request it checks if any other exceptions are pending. If no exceptions of higher priority are pending, interrupt handling is initiated. When initiating interrupt handling, the corresponding interrupt mask bit is set automatically for this and lower levels in status register. E.g, if an interrupt of level 3 is approved for handling, the interrupt mask bits I3M, I2M, I1M, and I0M are set in status register. If an interrupt of level 1 is approved, the masking bits I1M and I0M are set in status register. The handler address is calculated by logical OR of the AUTOVECTOR to the CPU system register Exception Vector Base Address (EVBA). The CPU will then jump to the calculated address and start executing the interrupt handler. Setting the interrupt mask bits prevents the interrupts from the same and lower levels to be passed through the interrupt controller. Setting of the same level mask bit prevents also multiple requests of the same interrupt to happen. It is the responsibility of the handler software to clear the interrupt request that caused the interrupt before returning from the interrupt handler. If the conditions that caused the interrupt are not cleared, the interrupt request remains active. 11.5.3 Clearing an Interrupt Request Clearing of the interrupt request is done by writing to registers in the corresponding peripheral module, which then clears the corresponding NMIREQ/IREQ signal. The recommended way of clearing an interrupt request is a store operation to the controlling peripheral register, followed by a dummy load operation from the same register. This causes a 142 32099I–01/2012 AT32UC3L016/32/64 pipeline stall, which prevents the interrupt from accidentally re-triggering in case the handler is exited and the interrupt mask is cleared before the interrupt request is cleared. 143 32099I–01/2012 AT32UC3L016/32/64 11.6 User Interface Table 11-1. INTC Register Memory Map Offset Register Register Name Access Reset 0x000 Interrupt Priority Register 0 IPR0 Read/Write 0x00000000 0x004 Interrupt Priority Register 1 IPR1 Read/Write 0x00000000 ... ... ... ... ... 0x0FC Interrupt Priority Register 63 IPR63 Read/Write 0x00000000 0x100 Interrupt Request Register 0 IRR0 Read-only N/A 0x104 Interrupt Request Register 1 IRR1 Read-only N/A ... ... ... ... ... 0x1FC Interrupt Request Register 63 IRR63 Read-only N/A 0x200 Interrupt Cause Register 3 ICR3 Read-only N/A 0x204 Interrupt Cause Register 2 ICR2 Read-only N/A 0x208 Interrupt Cause Register 1 ICR1 Read-only N/A 0x20C Interrupt Cause Register 0 ICR0 Read-only N/A 144 32099I–01/2012 AT32UC3L016/32/64 11.6.1 Name: Interrupt Priority Registers IPR0...IPR63 Access Type: Read/Write Offset: 0x000 - 0x0FC Reset Value: 0x00000000 31 30 INTLEVEL 29 28 27 26 25 24 - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - 7 6 1 0 AUTOVECTOR[13:8] 5 4 3 2 AUTOVECTOR[7:0] • INTLEVEL: Interrupt Level Indicates the EVBA-relative offset of the interrupt handler of the corresponding group: 00: INT0: Lowest priority 01: INT1 10: INT2 11: INT3: Highest priority • AUTOVECTOR: Autovector Address Handler offset is used to give the address of the interrupt handler. The least significant bit should be written to zero to give halfword alignment. 145 32099I–01/2012 AT32UC3L016/32/64 11.6.2 Name: Interrupt Request Registers IRR0...IRR63 Access Type: Read-only Offset: 0x0FF - 0x1FC Reset Value: N/A 31 30 29 28 27 26 25 24 IRR[32*x+31] IRR[32*x+30] IRR[32*x+29] IRR[32*x+28] IRR[32*x+27] IRR[32*x+26] IRR[32*x+25] IRR[32*x+24] 23 22 21 20 19 18 17 16 IRR[32*x+23] IRR[32*x+22] IRR[32*x+21] IRR[32*x+20] IRR[32*x+19] IRR[32*x+18] IRR[32*x+17] IRR[32*x+16] 15 14 13 12 11 10 9 8 IRR[32*x+15] IRR[32*x+14] IRR[32*x+13] IRR[32*x+12] IRR[32*x+11] IRR[32*x+10] IRR[32*x+9] IRR[32*x+8] 7 6 5 4 3 2 1 0 IRR[32*x+7] IRR[32*x+6] IRR[32*x+5] IRR[32*x+4] IRR[32*x+3] IRR[32*x+2] IRR[32*x+1] IRR[32*x+0] • IRR: Interrupt Request line This bit is cleared when no interrupt request is pending on this input request line. This bit is set when an interrupt request is pending on this input request line. The are 64 IRRs, one for each group. Each IRR has 32 bits, one for each possible interrupt request, for a total of 2048 possible input lines. The IRRs are read by the software interrupt handler in order to determine which interrupt request is pending. The IRRs are sampled continuously, and are read-only. 146 32099I–01/2012 AT32UC3L016/32/64 11.6.3 Name: Interrupt Cause Registers ICR0...ICR3 Access Type: Read-only Offset: 0x200 - 0x20C Reset Value: N/A 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - CAUSE • CAUSE: Interrupt Group Causing Interrupt of Priority n ICRn identifies the group with the highest priority that has a pending interrupt of level n. This value is only defined when at least one interrupt of level n is pending. 147 32099I–01/2012 AT32UC3L016/32/64 11.7 Module Configuration The specific configuration for each INTC instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 11-2. 11.7.1 INTC Clock Name Module Name Clock Name Description INTC CLK_INTC Clock for the INTC bus interface Interrupt Request Signal Map The Interrupt Controller supports up to 64 groups of interrupt requests. Each group can have up to 32 interrupt request signals. All interrupt signals in the same group share the same autovector address and priority level. The table below shows how the interrupt request signals are connected to the INTC. Table 11-3. Interrupt Request Signal Map Group Line Module Signal 0 0 AVR32 UC3 Core SYSREG COMPARE 0 AVR32 UC3 Core OCD DCEMU_DIRTY 1 AVR32 UC3 Core OCD DCCPU_READ 2 0 Flash Controller 3 0 Secure Access Unit 0 Peripheral DMA Controller PDCA 0 1 Peripheral DMA Controller PDCA 1 2 Peripheral DMA Controller PDCA 2 3 Peripheral DMA Controller PDCA 3 0 Peripheral DMA Controller PDCA 4 1 Peripheral DMA Controller PDCA 5 2 Peripheral DMA Controller PDCA 6 3 Peripheral DMA Controller PDCA 7 0 Peripheral DMA Controller PDCA 8 1 Peripheral DMA Controller PDCA 9 2 Peripheral DMA Controller PDCA 10 3 Peripheral DMA Controller PDCA 11 7 0 Power Manager 8 0 System Control Interface 9 0 Asynchronous Timer 1 FLASHCDW SAU 4 5 6 PM SCIF AST ALARM 148 32099I–01/2012 AT32UC3L016/32/64 Table 11-3. Interrupt Request Signal Map 0 Asynchronous Timer AST PER 1 Asynchronous Timer AST OVF 2 Asynchronous Timer AST READY 3 Asynchronous Timer AST CLKREADY 0 External Interrupt Controller EIC 1 1 External Interrupt Controller EIC 2 2 External Interrupt Controller EIC 3 3 External Interrupt Controller EIC 4 12 0 External Interrupt Controller EIC 5 13 0 Frequency Meter FREQM 0 General Purpose Input/Output Controller GPIO 0 1 General Purpose Input/Output Controller GPIO 1 2 General Purpose Input/Output Controller GPIO 2 3 General Purpose Input/Output Controller GPIO 3 4 General Purpose Input/Output Controller GPIO 4 5 General Purpose Input/Output Controller GPIO 5 15 0 Universal Synchronous/Asynchronous Receiver/Transmitter USART0 16 0 Universal Synchronous/Asynchronous Receiver/Transmitter USART1 17 0 Universal Synchronous/Asynchronous Receiver/Transmitter USART2 18 0 Universal Synchronous/Asynchronous Receiver/Transmitter USART3 19 0 Serial Peripheral Interface SPI 20 0 Two-wire Master Interface TWIM0 21 0 Two-wire Master Interface TWIM1 22 0 Two-wire Slave Interface TWIS0 23 0 Two-wire Slave Interface TWIS1 24 0 Pulse Width Modulation Controller PWMA 0 Timer/Counter TC00 1 Timer/Counter TC01 2 Timer/Counter TC02 0 Timer/Counter TC10 1 Timer/Counter TC11 2 Timer/Counter TC12 0 ADC Interface ADCIFB 10 11 14 25 26 27 149 32099I–01/2012 AT32UC3L016/32/64 Table 11-3. Interrupt Request Signal Map 28 0 Analog Comparator Interface ACIFB 29 0 Capacitive Touch Module CAT 30 0 aWire AW 150 32099I–01/2012 AT32UC3L016/32/64 12. Power Manager (PM) Rev: 4.1.1.4 12.1 Features • • • • • • • 12.2 Generates clocks and resets for digital logic On-the-fly frequency change of CPU, HSB and PBx clocks Sleep modes allow simple disabling of logic clocks and clock sources Module-level clock gating through maskable peripheral clocks Wake-up from internal or external interrupts Automatic identification of reset sources Supports advanced Shutdown sleep mode Overview The Power Manager (PM) provides synchronous clocks used to clock the main digital logic in the device, namely the CPU, and the modules and peripherals connected to the High Speed Bus (HSB) and the Peripheral Buses (PBx). The PM contains advanced power-saving features, allowing the user to optimize the power consumption for an application. The synchronous clocks are divided into a number of clock domains, one for the CPU and HSB, and one for each PBx. The clocks can run at different speeds, allowing the user to save power by running peripherals relatively slow, whilst maintaining high CPU performance. The clocks can be independently changed on-the-fly, without halting any peripherals. The user may adjust CPU and memory speeds according to the dynamic application load, without disturbing or re-configuring active peripherals. Each module has a separate clock, enabling the user to save power by switching off clocks to inactive modules. Clocks and oscillators can be automatically switched off during idle periods by the CPU sleep instruction. The system will return to normal operation when interrupts occur. To achieve minimal power usage, a special sleep mode, called Shutdown is available, where power on all internal logic (CPU, peripherals) and most of the I/O lines is removed, reducing current leakage. Only a small amount of logic, including the 32KHz crystal oscillator (OSC32K) and the AST remain powered. The Power Manager also contains a Reset Controller, which collects all possible reset sources, generates hard and soft resets, and allows the reset source to be identified by software. 151 32099I–01/2012 AT32UC3L016/32/64 12.3 Block Diagram Figure 12-1. PM Block Diagram Main Clock Sources Synchronous Clock Generator Synchronous clocks CPU, HSB, PBx Interrupts Sleep Controller Sleep Instruction Reset Controller Resets Reset Sources Power-on Reset Detector(s) External Reset Pin 12.4 I/O Lines Description Table 12-1. I/O Lines Description Name Description Type Active Level RESET_N Reset Input Low 12.5 12.5.1 Product Dependencies Interrupt The PM interrupt line is connected to one of the interrupt controllers internal sources. Using the PM interrupt requires the interrupt controller to be configured first. 12.5.2 Clock Implementation In AT32UC3L016/32/64, the HSB shares source clock with the CPU. Write attempts to the HSB Clock Select register (HSBSEL) will be ignored, and it will always read the same as the CPU Clock Select register (CPUSEL). The PM bus interface clock (CLK_PM) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. If disabled it can only be reenabled by a reset. 12.5.3 Power Considerations The Shutdown mode is only available for the “3.3V supply mode, with 1.8V regulated I/O lines“ power configuration. 152 32099I–01/2012 AT32UC3L016/32/64 12.6 Functional Description 12.6.1 Synchronous Clocks The System RC Oscillator (RCSYS) and a selection of other clock sources can provide the source for the main clock, which is the origin for the synchronous CPU/HSB and PBx module clocks. For details about the other main clock sources, please refer to the Main Clock Control (MCCTRL) register description. The synchronous clocks can run of the main clock and all the 8bit prescaler settings as long as fCPU ≥ fPBx,. The synchronous clock source can be changed onthe fly, according to variations in application load. The clock domains can be shut down in sleep mode, as described in Section 12.6.3. The module clocks in every synchronous clock domain can be individually masked to minimize power consumption in inactive modules. Figure 12-2. Synchronous Clock Generation Sleep Controller Sleep Instruction 0 Main Clock Main Clock Sources 1 Prescaler CPUDIV MCSEL Mask CPU Clocks HSB Clocks CPUMASK PBx Clocks CPUSEL 12.6.1.1 Selecting the main clock source The common main clock can be connected to RCSYS or a selection of other clock sources. For details about the other main clock sources, please refer to the MCCTRL register description. By default, the main clock will be connected to RCSYS. The user can connect the main clock to another source by writing to the Main Clock Select (MCCTRL.MCSEL) field. The user must first assure that the source is enabled and ready in order to avoid a deadlock. Care should also be taken so that the new synchronous clock frequencies do not exceed the maximum frequency for each clock domain. 12.6.1.2 Selecting synchronous clock division ratio The main clock feeds an 8-bit prescaler, which can be used to generate the synchronous clocks. By default, the synchronous clocks run on the undivided main clock. The user can select a prescaler division for the CPU clock by writing a one to the CPU Division bit in the CPU Clock Select register (CPUSEL.CPUDIV), and a value to the CPU Clock Select field (CPUSEL.CPUSEL), resulting in a CPU clock frequency: fCPU = fmain / 2(CPUSEL+1) 153 32099I–01/2012 AT32UC3L016/32/64 Similarly, the PBx clocks can be divided by writing their respective Clock Select (PBxSEL) registers to get the divided PBx frequency: fPBx = fmain / 2(PBSEL+1) The PBx clock frequency can not exceed the CPU clock frequency. The user must select a PBxSEL.PBSEL value greater than or equal to the CPUSEL.CPUSEL value, so that fCPU ≥ fPBx. If the user selects division factors that will result in fCPU< fPBx, the Power Manager will automatically change the PBxSEL.PBSEL/PBDIV values to ensure correct operation (fCPU ≥ fPBx). The HSB clock will always be forced to the same division as the CPU clock. To ensure correct operation, the frequencies must never exceed the specified maximum frequency for each clock domain. For modules connected to the HSB bus, the PB clock frequency must be the same as the CPU clock frequency. 12.6.1.3 Clock Ready flag There is a slight delay from CPUSEL and PBxSEL being written to the new clock setting taking effect. During this interval, the Clock Ready bit in the Status Register (SR.CKRDY) will read as zero. When the clock settings change is completed, the bit will read as one. The Clock Select registers (CPUSEL, PBxSEL) must not be written to while SR.CKRDY is zero, or the system may become unstable or hang. The Clock Ready bit in the Interrupt Status Register (ISR.CKRDY) is set on a SR.CKRDY zeroto-one transition. If the Clock Ready bit in the Interrupt Mask Register (IMR.CKRDY) is set, an interrupt request is generated. IMR.CKRDY is set by writing a one to the corresponding bit in the Interrupt Enable Register (IER.CKRDY). 12.6.2 Peripheral Clock Masking By default, the clocks for all modules are enabled, regardless of which modules are actually being used. It is possible to disable the clock for a module in the CPU, HSB, or PBx clock domain by writing a zero to the corresponding bit in the corresponding Clock Mask (CPUMASK/HSBMASK/PBxMASK) register. When a module is not clocked, it will cease operation, and its registers cannot be read nor written. The module can be re-enabled later by writing a one to the corresponding mask bit. A module may be connected to several clock domains, in which case it will have several mask bits. The Maskable Module Clocks table in the Clock Mask register description contains a list of implemented maskable clocks. 12.6.2.1 12.6.3 Cautionary note Note that clocks should only be switched off if it is certain that the module will not be used. Switching off the clock for the Flash Controller will cause a problem if the CPU needs to read from the flash. Switching off the clock to the Power Manager, which contains the mask registers, or the corresponding PBx bridge, will make it impossible to write to the mask registers again. In this case, they can only be re-enabled by a system reset. Sleep Modes In normal operation, all clock domains are active, allowing software execution and peripheral operation. When the CPU is idle, it is possible to switch it and other (optional) clock domains off to save power. This is done by the sleep instruction, which takes the sleep mode index number from Table 12-2 on page 155 as argument. 154 32099I–01/2012 AT32UC3L016/32/64 12.6.3.1 Entering and exiting sleep modes The sleep instruction will halt the CPU and all modules belonging to the stopped clock domains. The modules will be halted regardless of the bit settings in the mask registers. Clock sources can also be switched off to save power. Some of these have a relatively long start-up time, and are only switched off when very low power consumption is required. The CPU and affected modules are restarted when the sleep mode is exited. This occurs when an interrupt triggers. Note that even if an interrupt is enabled in sleep mode, it may not trigger if the source module is not clocked. 12.6.3.2 Supported sleep modes The following sleep modes are supported. These are detailed in Table 12-2 on page 155. • Idle: The CPU is stopped, the rest of the device is operational. • Frozen: The CPU and HSB modules are stopped, peripherals are operational. • Standby: All synchronous clocks are stopped, and the clock sources are running, allowing for a quick wake-up to normal mode. • Stop: As Standby, but oscillators, and other clock sources are also stopped. 32KHz Oscillator OSC32K(2), RCSYS, AST, and WDT will remain operational. • DeepStop: All synchronous clocks and clock sources are stopped. Bandgap voltage reference and BOD are turned off. OSC32K(2) and RCSYS remain operational. • Static: All clock sources, including RCSYS are stopped. Bandgap voltage reference and BOD are turned off. OSC32K(2) remains operational. • Shutdown: All clock sources, including RCSYS are stopped. Bandgap voltage reference, BOD detector, and Voltage regulator are turned off. OSC32K(2) remains operational. This mode can only be used in the “3.3V supply mode, with 1.8V regulated I/O lines“ configuration (described in Power Considerations chapter). Refer to Section 12.6.4 for more details. Table 12-2. Sleep Modes Clock Sources(3), SYSTIMER(4) OSC32K Sleep Mode CPU HSB PBx, GCLK 0 Idle Stop Run Run Run Run Run On Normal mode 1 Frozen Stop Stop Run Run Run Run On Normal mode 2 Standby Stop Stop Stop Run Run Run On Normal mode 3 Stop Stop Stop Stop Stop Run Run On Low power mode 4 DeepStop Stop Stop Stop Stop Run Run Off Low power mode 5 Static Stop Stop Stop Stop Run Stop Off Low power mode 6 Shutdown Stop Stop Stop Stop Run Stop Off Off (1) Index Notes: (2) RCSYS BOD & Bandgap Voltage Regulator 1. The sleep mode index is used as argument for the sleep instruction. 2. OSC32K will only remain operational if pre-enabled. 3. Clock sources other than those specifically listed in the table. 4. SYSTIMER is the clock for the CPU COUNT and COMPARE registers. The internal voltage regulator is also adjusted according to the sleep mode in order to reduce its power consumption. 155 32099I–01/2012 AT32UC3L016/32/64 12.6.3.3 Waking from sleep modes There are two types of wake-up sources from sleep mode, synchronous and asynchronous. Synchronous wake-up sources are all non-masked interrupts. Asynchronous wake-up sources are AST, WDT, external interrupts from EIC, external reset, external wake pin (WAKE_N), and all asynchronous wake-ups enabled in the Asynchronous Wake Up Enable (AWEN) register. The valid wake-up sources for each sleep mode are detailed in Table 12-3 on page 156. In Shutdown the only wake-up sources are external reset, external wake-up pin or AST. See Section 12.6.4.3 on page 158. Table 12-3. Index(1) Wake-up Sources Sleep Mode Wake-up Sources 0 Idle Synchronous, Asynchronous 1 Frozen Synchronous(2), Asynchronous 2 Standby Asynchronous 3 Stop Asynchronous 4 DeepStop Asynchronous 5 Static Asynchronous(3) 6 Shutdown External reset, External wake-up pin Notes: 1. The sleep mode index is used as argument for the sleep instruction. 2. Only PB modules operational, as HSB module clocks are stopped. 3. WDT only available if clocked from pre-enabled OSC32K. 12.6.3.4 SleepWalking In all sleep modes where the PBx clocks are stopped, except for Shutdown mode, the device can partially wake up if a PBx module asynchronously discovers that it needs its clock. Only the requested clocks and clock sources needed will be started, all other clocks will remain masked to zero. E.g. if the main clock source is OSC0, only OSC0 will be started even if other clock sources were enabled in normal mode. Generic clocks can also be started in a similar way. The state where only requested clocks are running is referred to as SleepWalking. The time spent to start the requested clock is mostly limited by the startup time of the given clock source. This allows PBx modules to handle incoming requests, while still keeping the power consumption at a minimum. When the device is SleepWalking any asynchronous wake-up can wake the device up at any time without stopping the requested PBx clock. All requests to start clocks can be masked by writing to the Peripheral Power Control Register (PPCR), all requests are enabled at reset. During SleepWalking the interrupt controller clock will be running. If an interrupt is pending when entering SleepWalking, it will wake the whole device up. 12.6.3.5 Precautions when entering sleep mode Modules communicating with external circuits should normally be disabled before entering a sleep mode that will stop the module operation. This will prevent erratic behavior caused by entering or exiting sleep modes. Please refer to the relevant module documentation for recommended actions. 156 32099I–01/2012 AT32UC3L016/32/64 Communication between the synchronous clock domains is disturbed when entering and exiting sleep modes. Bus transactions over clock domains affected by the sleep mode are therefore not recommended. The system may hang if the bus clocks are stopped during a bus transaction. The CPU is automatically stopped in a safe state to ensure that all CPU bus operations are complete when the sleep mode goes into effect. Thus, when entering Idle mode, no further action is necessary. When entering a sleep mode (except Idle mode), all HSB masters must be stopped before entering the sleep mode. In order to let potential PBx write operations complete, the user should let the CPU perform a PBx register read operation before issuing the sleep instruction. This will stall the CPU until pending PBx operations have completed. The Shutdown sleep mode requires extra care. Please refer to Section 12.6.4. 12.6.4 Shutdown Sleep Mode 12.6.4.1 Description The Shutdown sleep mode is available only when the device is used in the “3.3V supply mode, with 1.8V regulated I/O lines“ configuration (refer to the Power Considerations chapter). In this configuration, the voltage regulator supplies both VDDCORE and VDDIO power supplies. When the device enters Shutdown mode, the regulator is turned off and only the following logic is kept powered by VDDIN: – OSC32K using alternate pinout PA13/PA20 – AST core logic (internal counter and alarm detection logic) – Backup Registers – I/O lines PA11, PA13, PA20, PA21, PB04, PB05, and PB10 – RESET_N line The table below lists I/O line functionality that remains operational during Shutdown sleep mode. If no special function is used the I/O line will keep its setting when entering the sleep mode Table 12-4. I/O Lines Usage During Shutdown Mode Pin Possible Usage During Shutdown Sleep Mode PA11 WAKE_N signal (active low wake-up) PA13 XIN32_2 (OSC32K using alternate pinout) PA20 XOUT32_2 (OSC32K using alternate pinout) PA21 PB04 PB05 PB10 RESET_N 12.6.4.2 Reset pin Entering Shutdown sleep mode Before entering the Shutdown sleep mode, a few actions are required: – All modules should normally be disabled before entering Shutdown sleep mode (see Section 12.6.3.5) 157 32099I–01/2012 AT32UC3L016/32/64 – The POR33 must be masked to avoid spurious resets when the power is back. This must also be done when POR33 is disabled, as POR33 will be enabled automatically when the device wakes up from Shutdown mode. Disable the POR33 by writing a one to the POR33MASK bit in the SCIF.VREGCR register. Due to internal synchronisation, this bit must be read as a one before the sleep instruction is executed by the CPU. Refer to the System Control Interface (SCIF) chapter for more details. – The 32KHz RC oscillator (RC32K) must be running and stable. This is done by writing a one to the EN bit in the SCIF.RC32KCR register. Due to internal synchronisation, this bit must be read as a one to ensure that the oscillator is stable before the sleep instruction is executed by the CPU. As soon as the Shutdown sleep mode is entered, all CPU and peripherals are reset to ensure a consistent state. POR33 and RC32K are automatically disabled to save extra power. 12.6.4.3 Leaving Shutdown sleep mode Exiting Shutdown sleep mode can be done by the events described in Table 12-5. Table 12-5. Events That Can Wake up the Device from Shutdown Mode Source How PA11 (WAKE_N) Pulling-down PA11 will wake up the device RESET_N Pulling-down RESET_N pin will wake up the device The device is kept under reset until RESET_N is tied high again AST OSC32K must be set-up to use alternate pinout (XIN32_2 and XOUT32_2) Refer to the SCIF Chapter AST must be configured to use the clock from OSC32K AST must be configured to allow alarm, periodic, or overflow wake-up When a wake-up event occurs, the regulator is turned on and the device will wait for VDDCORE to be valid before starting. The Sleep Reset bit in the Reset Cause register (RCAUSE.SLEEP) is then set, allowing software running on the device to distinguish between the first power-up and a wake-up from Shutdown mode. 12.6.4.4 12.6.5 Special consideration regarding waking up from Shutdown sleep mode using the WAKE_N pin By default, the WAKE_N pin will only wake the device up if it is pulled low after entering Shutdown mode. If the WAKE_N is pulled low before the Shutdown mode is entered, it will not wake the device from the Shutdown sleep mode. In order to wake the device by pulling WAKE_N low before entering Shutdown mode, the user has to write a one to the bit corresponding to the WAKEN wake-up source in the AWEN register. In this scenario, the CPU execution will proceed with the next instruction, and the RCAUSE register content will not be altered. Divided PB Clocks The clock generator in the Power Manager provides divided PBx clocks for use by peripherals that require a prescaled PBx clock. This is described in the documentation for the relevant modules. The divided clocks are directly maskable, and are stopped in sleep modes where the PBx clocks are stopped. 158 32099I–01/2012 AT32UC3L016/32/64 12.6.6 Reset Controller The Reset Controller collects the various reset sources in the system and generates hard and soft resets for the digital logic. The device contains a Power-on Reset (POR) detector, which keeps the system reset until power is stable. This eliminates the need for external reset circuitry to guarantee stable operation when powering up the device. It is also possible to reset the device by pulling the RESET_N pin low. This pin has an internal pull-up, and does not need to be driven externally during normal operation. Table 12-6 on page 159 lists these and other reset sources supported by the Reset Controller. Figure 12-3. Reset Controller Block Diagram RCAUSE RESET_N Power-on Reset Detector(s) CPU, HSB, PBx Brown-out Detector Reset Controller OCD, AST, WDT, Clock Generator SM33 Detector JTAG AWIRE OCD Watchdog Reset In addition to the listed reset types, the JTAG & aWire can keep parts of the device statically reset. See JTAG and aWire documentation for details. Table 12-6. Reset Description Reset Source Description Power-on Reset Supply voltage below the Power-on Reset detector threshold voltage VPOT External Reset RESET_N pin asserted Brown-out Reset VDDCORE supply voltage below the Brown-out detector threshold voltage 159 32099I–01/2012 AT32UC3L016/32/64 Reset Source Description SM33 Reset Internal regulator supply voltage below the SM33 threshold voltage. This generates a Power-on Reset. Watchdog Timer See Watchdog Timer documentation OCD See On-Chip Debug documentation Depending on the reset source, when a reset occurs, some parts of the device are not always reset. Only the Power-on Reset (POR) will force a whole device reset. Refer to the table in the Module Configuration section at the end of this chapter for further details. The latest reset cause can be read in the RCAUSE register, and can be read during the applications boot sequence in order to determine proper action. 12.6.6.1 Power-on Reset Detector The Power-on Reset 1.8V (POR18) detector monitors the VDDCORE supply pin and generates a Power-on Reset (POR) when the device is powered on. The POR is active until the VDDCORE voltage is above the power-on threshold level (VPOT). The POR will be re-generated if the voltage drops below the power-on threshold level. See Electrical Characteristics for parametric details. The Power-on Reset 3.3V (POR33) detector monitors the internal regulator supply pin and generates a Power-on Reset (POR) when the device is powered on. The POR is active until the internal regulator supply voltage is above the regulator power-on threshold level (VPOT). The POR will be re-generated if the voltage drops below the regulator power-on threshold level. See Electrical Characteristics for parametric details. 12.6.6.2 12.6.7 External Reset The external reset detector monitors the RESET_N pin state. By default, a low level on this pin will generate a reset. Clock Failure Detector This mechanism automatically switches the main clock source to the safe RCSYS clock when the main clock source fails. This may happen when an external crystal is selected as a source for the main clock and the crystal is not mounted on the board. The main clock is compared with RCSYS, and if no rising edge of the main clock is detected during one RCSYS period, the clock is considered to have failed. The detector is enabled by writing a one to the Clock Failure Detection Enable bit in the Clock Failure Detector Control Register (CFDCTRL.CFDEN). As soon as the detector is enabled, the clock failure detector will monitor the divided main clock. Note that the detector does not monitor the main clock if RCSYS is the source of the main clock, or if the main clock is temporarily not available (startup-time after a wake-up, switching timing etc.), or in sleep mode where the main clock is driven by the RCSYS (Stop and DeepStop mode). When a clock failure is detected, the main clock automatically switches to the RCSYS clock and the Clock Failure Detected (CFD) interrupt is generated if enabled. The MCCTRL register is also changed by hardware to indicate that the main clock comes from RCSYS. 12.6.8 Interrupts The PM has a number of interrupt sources: • AE - Access Error, 160 32099I–01/2012 AT32UC3L016/32/64 – A lock protected register is written to without first being unlocked. • CKRDY - Clock Ready: – New Clock Select settings in the CPUSEL/PBxSEL registers have taken effect. (A zero-to-one transition on SR.CKRDY is detected). • CFD - Clock Failure Detected: – The system detects that the main clock is not running. The Interrupt Status Register contains one bit for each interrupt source. A bit in this register is set on a zero-to-one transition of the corresponding bit in the Status Register (SR), and cleared by writing a one to the corresponding bit in the Interrupt Clear Register (ICR). The interrupt sources will generate an interrupt request if the corresponding bit in the Interrupt Mask Register is set. The interrupt sources are ORed together to form one interrupt request. The Power Manager will generate an interrupt request if at least one of the bits in the Interrupt Mask Register (IMR) is set. Bits in IMR are set by writing a one to the corresponding bit in the Interrupt Enable Register (IER), and cleared by writing a one to the corresponding bit in the Interrupt Disable Register (IDR). The interrupt request remains active until the corresponding bit in the Interrupt Status Register (ISR) is cleared by writing a one to the corresponding bit in the Interrupt Clear Register (ICR). Because all the interrupt sources are ORed together, the interrupt request from the Power Manager will remain active until all the bits in ISR are cleared. 161 32099I–01/2012 AT32UC3L016/32/64 12.7 User Interface Table 12-7. PM Register Memory Map Offset Register Register Name Access Reset 0x000 Main Clock Control MCCTRL Read/Write 0x00000000 0x004 CPU Clock Select CPUSEL Read/Write 0x00000000 0x008 HSB Clock Select HSBSEL Read-only 0x00000000 0x00C PBA Clock Select PBASEL Read/Write 0x00000000 0x010 PBB Clock Select PBBSEL Read/Write 0x00000000 0x014 - 0x01C Reserved 0x020 CPU Mask CPUMASK Read/Write 0x00010003 0x024 HSB Mask HSBMASK Read/Write 0x0000007F 0x028 PBA Mask PBAMASK Read/Write 0x03FFFFFF 0x02C PBB Mask PBBMASK Read/Write 0x00000007 0x030- 0x03C Reserved 0x040 PBA Divided Mask PBADIVMASK Read/Write 0x0000007F 0x044 - 0x050 Reserved 0x054 Clock Failure Detector Control CFDCTRL Read/Write 0x00000000 0x058 Unlock Register UNLOCK Write-only 0x00000000 0x05C - 0x0BC Reserved 0x0C0 Interrupt Enable Register IER Write-only 0x00000000 0x0C4 Interrupt Disable Register IDR Write-only 0x00000000 0x0C8 Interrupt Mask Register IMR Read-only 0x00000000 0x0CC Interrupt Status Register ISR Read-only 0x00000000 0x0D0 Interrupt Clear Register ICR Write-only 0x00000000 0x0D4 Status Register SR Read-only 0x00000020 0x0D8 - 0x15C Reserved 0x160 Peripheral Power Control Register PPCR Read/Write 0x000001FA 0x164 - 0x17C Reserved 0x180 Reset Cause Register RCAUSE Read-only -(2) 0x184 Wake Cause Register WCAUSE Read-only -(3) 0x188 Asynchronous Wake Up Enable Register AWEN Read/Write 0x00000000 0x18C - 0x3F4 Reserved 0x3F8 Configuration Register CONFIG Read-only 0x00000043 0x3FC Version Register VERSION Read-only -(1) Note: 1. The reset value is device specific. Please refer to the Module Configuration section at the end of this chapter. 2. Latest Reset Source. 3. Latest Wake Source. 162 32099I–01/2012 AT32UC3L016/32/64 12.7.1 Main Clock Control Name: MCCTRL Access Type: Read/Write Offset: 0x000 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - MCSEL • MCSEL: Main Clock Select Table 12-8. Note: Main clocks in AT32UC3L016/32/64 MCSEL[2:0] Main clock source 0 System RC oscillator (RCSYS) 1 Oscillator0 (OSC0) 2 DFLL 3 120MHz RC oscillator (RC120M)(1) 1. If the 120MHz RC oscillator is selected as main clock source, it must be divided by at least 4 before being used as clock source for the CPU. This division is selected by writing to the CPUSEL and CPUDIV bits in the CPUSEL register, before switching to RC120M as main clock source. Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 163 32099I–01/2012 AT32UC3L016/32/64 12.7.2 CPU Clock Select Name: CPUSEL Access Type: Read/Write Offset: 0x004 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 CPUDIV - - - - CPUSEL • CPUDIV, CPUSEL: CPU Division and Clock Select CPUDIV = 0: CPU clock equals main clock. CPUDIV = 1: CPU clock equals main clock divided by 2(CPUSEL+1). Note that if CPUDIV is written to 0, CPUSEL should also be written to 0 to ensure correct operation. Also note that writing this register clears SR.CKRDY. The register must not be re-written until CKRDY is set. Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 164 32099I–01/2012 AT32UC3L016/32/64 12.7.3 HSB Clock Select Name: HSBSEL Access Type: Read Offset: 0x008 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 HSBDIV - - - - HSBSEL This register is read-only and its content is always equal to CPUSEL. 165 32099I–01/2012 AT32UC3L016/32/64 12.7.4 PBx Clock Select Name: PBxSEL Access Type: Read/Write Offset: 0x00C-0x010 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 PBDIV - - - - PBSEL • PBDIV, PBSEL: PBx Division and Clock Select PBDIV = 0: PBx clock equals main clock. PBDIV = 1: PBx clock equals main clock divided by 2(PBSEL+1). Note that if PBDIV is written to 0, PBSEL should also be written to 0 to ensure correct operation. Also note that writing this register clears SR.CKRDY. The register must not be re-written until SR.CKRDY is set. Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 166 32099I–01/2012 AT32UC3L016/32/64 12.7.5 Clock Mask Name: CPUMASK/HSBMASK/PBAMASK/PBBMASK Access Type: Read/Write Offset: 0x020-0x02C Reset Value: - 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 MASK[31:24] 23 22 21 20 MASK[23:16] 15 14 13 12 MASK[15:8] 7 6 5 4 MASK[7:0] • MASK: Clock Mask If bit n is cleared, the clock for module n is stopped. If bit n is set, the clock for module n is enabled according to the current power mode. The number of implemented bits in each mask register, as well as which module clock is controlled by each bit, is shown in Table 12-9. Table 12-9. Maskable Module Clocks in AT32UC3L016/32/64. Bit CPUMASK HSBMASK PBAMASK PBBMASK 0 OCD PDCA PDCA FLASHCDW 1 Reserved FLASHCDW INTC HMATRIX 2 - SAU PM SAU 3 - PBB bridge SCIF - 4 - PBA bridge AST - 5 - Peripheral Event System WDT - 6 - - EIC - 7 - - FREQM - 8 - - GPIO - 9 - - USART0 - 10 - - USART1 - 11 - - USART2 - 167 32099I–01/2012 AT32UC3L016/32/64 Table 12-9. Maskable Module Clocks in AT32UC3L016/32/64. Bit CPUMASK HSBMASK PBAMASK PBBMASK 12 - - USART3 - 13 - - SPI - 14 - - TWIM0 - 15 - - TWIM1 - 16 SYSTIMER - TWIS0 - 17 - - TWIS1 - 18 - - PWMA - 19 - - TC0 - 20 - - TC1 - 21 - - ADCIFB - 22 - - ACIFB - 23 - - CAT - 24 - - GLOC - 25 - - AW - 31:26 - - - - Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 168 32099I–01/2012 AT32UC3L016/32/64 12.7.6 PBA Divided Mask Name: PBADIVMASK Access Type: Read/Write Offset: 0x040 Reset Value: 0x0000007F 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - MASK[6:0] • MASK: Clock Mask If bit n is written to zero, the clock divided by 2(n+1) is stopped. If bit n is written to one, the clock divided by 2(n+1) is enabled according to the current power mode. Table 12-10 shows what clocks are affected by the different MASK bits. Table 12-10. Divided Clock Mask Bit USART0 USART1 USART2 USART3 TC0 TC1 0 - - - - TIMER_CLOCK2 TIMER_CLOCK2 1 - - - - - - 2 CLK_USART/ DIV CLK_USART/ DIV CLK_USART/ DIV CLK_USART/ DIV TIMER_CLOCK3 TIMER_CLOCK3 3 - - - - - - 4 - - - - TIMER_CLOCK4 TIMER_CLOCK4 5 - - - - - - 6 - - - - TIMER_CLOCK5 TIMER_CLOCK5 Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 169 32099I–01/2012 AT32UC3L016/32/64 12.7.7 Clock Failure Detector Control Register Name: CFDCTRL Access Type: Read/Write Offset: 0x054 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 SFV - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - CFDEN • SFV: Store Final Value 0: The register is read/write 1: The register is read-only, to protect against further accidental writes. • CFDEN: Clock Failure Detection Enable 0: Clock Failure Detector is disabled 1: Clock Failure Detector is enabled Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 170 32099I–01/2012 AT32UC3L016/32/64 12.7.8 Unlock Register Name: UNLOCK Access Type: Write-only Offset: 0x058 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 KEY 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - 7 6 5 4 3 2 ADDR[9:8] 1 0 ADDR[7:0] To unlock a write protected register, first write to the UNLOCK register with the address of the register to unlock in the ADDR field and 0xAA in the KEY field. Then, in the next PB access write to the register specified in the ADDR field. • KEY: Unlock Key Write this bit field to 0xAA to enable unlock. • ADDR: Unlock Address Write the address of the register to unlock to this field. 171 32099I–01/2012 AT32UC3L016/32/64 12.7.9 Interrupt Enable Register Name: IER Access Type: Write-only Offset: 0x0C0 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 AE - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - CKRDY - - - - CFD Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 172 32099I–01/2012 AT32UC3L016/32/64 12.7.10 Interrupt Disable Register Name: IDR Access Type: Write-only Offset: 0x0C4 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 AE - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - CKRDY - - - - CFD Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 173 32099I–01/2012 AT32UC3L016/32/64 12.7.11 Interrupt Mask Register Name: IMR Access Type: Read-only Offset: 0x0C8 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 AE - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - CKRDY - - - - CFD 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. This bit is cleared when the corresponding bit in IDR is written to one. This bit is set when the corresponding bit in IER is written to one. 174 32099I–01/2012 AT32UC3L016/32/64 12.7.12 Interrupt Status Register Name: ISR Access Type: Read-only Offset: 0x0CC Reset Value: 0x00000000 31 30 29 28 27 26 25 24 AE - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - CKRDY - - - - CFD 0: The corresponding interrupt is cleared. 1: The corresponding interrupt is pending. This bit is cleared when the corresponding bit in ICR is written to one. This bit is set on a zero-to-one transition of the corresponding bit in the Status Register (SR). 175 32099I–01/2012 AT32UC3L016/32/64 12.7.13 Interrupt Clear Register Name: ICR Access Type: Write-only Offset: 0x0D0 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 AE - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - CKRDY - - - - CFD Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in ISR. 176 32099I–01/2012 AT32UC3L016/32/64 12.7.14 Status Register Name: SR Access Type: Read-only Offset: 0x0D4 Reset Value: 0x00000020 31 30 29 28 27 26 25 24 AE - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - CKRDY - - - - CFD • AE: Access Error 0: No access error has occurred. 1: A write to lock protected register without unlocking it has occurred. • CKRDY: Clock Ready 0: One of the CPUSEL/PBxSEL registers has been written, and the new clock setting is not yet effective. 1: The synchronous clocks have frequencies as indicated in the CPUSEL/PBxSEL registers. • CFD: Clock Failure Detected 0: Main clock is running correctly. 1: Failure on main clock detected. Main clock is now running on RCSYS. 177 32099I–01/2012 AT32UC3L016/32/64 12.7.15 Peripheral Power Control Register Name: PPCR Access Type: Read/Write Offset: 0x004 Reset Value: 0x000001FA 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 PPC[31:24] 23 22 21 20 PPC[23:16] 15 14 13 12 PPC[15:8] 7 6 5 4 PPC[7:0] Table 12-11. Peripheral Power Control Bit Name 0 RSTPUN 1 FRC32 2 RSTTM 3 CATRCMASK 4 ACIFBCRCMASK 5 ADCIFBRCMASK 6 ASTRCMASK 7 TWIS0RCMASK 8 TWIS1RCMASK 31:9 - • RSTTM: Reset test mode 0: External reset not in test mode 1: External reset in test mode • FRC32: Force RC32 out 0: RC32 signal is not forced as output 1: RC32 signal is forced as output • RSTPUN: Reset Pullup, active low 0: Pull-up for external reset on 1: Pull-up for external reset off 178 32099I–01/2012 AT32UC3L016/32/64 • CATRCMASK: CAT Request Clock Mask 0: CAT Request Clock is disabled 1: CAT Request Clock is enabled • ACIFBRCMASK: ACIFB Request Clock Mask 0: ACIFB Request Clock is disabled 1: ACIFB Request Clock is enabled • ADCIFBRCMASK: ADCIFB Request Clock Mask 0: ADCIFB Request Clock is disabled 1: ADCIFB Request Clock is enabled • ASTRCMASK: AST Request Clock Mask 0: AST Request Clock is disabled 1: AST Request Clock is enabled • TWIS0RCMASK: TWIS0 Request Clock Mask 0: TWIS0 Request Clock is disabled 1: TWIS0 Request Clock is enabled • TWIS1RCMASK: TWIS1 Request Clock Mask 0: TWIS1 Request Clock is disabled 1: TWIS1 Request Clock is enabled Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 179 32099I–01/2012 AT32UC3L016/32/64 12.7.16 Reset Cause Register Name: RCAUSE Access Type: Read-only Offset: 0x180 Reset Value: Latest Reset Source 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - AWIRE - - JTAG OCDRST 7 6 5 4 3 2 1 0 - SLEEP - - WDT EXT BOD POR • AWIRE: aWire Reset This bit is set when the last reset was caused by the aWire. • JTAG: JTAG Reset This bit is set when the last reset was caused by the JTAG. • OCDRST: OCD Reset This bit is set when the last reset was due to the RES bit in the OCD Development Control register having been written to one. • SLEEP: Sleep Reset This bit is set when the last reset was due to the device waking up from the Shutdown sleep mode. • WDT: Watchdog Reset This bit is set when the last reset was due to a watchdog time-out. • EXT: External Reset Pin This bit is set when the last reset was due to the RESET_N pin being pulled low. • BOD: Brown-out Reset This bit is set when the last reset was due to the core supply voltage being lower than the brown-out threshold level. • POR: Power-on Reset This bit is set when the last reset was due to the core supply voltage VDDCORE being lower than the power-on threshold level (the reset is generated by the POR18 detector), or the internal regulator supply voltage being lower than the regulator power-on threshold level (generated by the POR33 detector), or the internal regulator supply voltage being lower than the minimum required input voltage (generated by the 3.3V supply monitor SM33). 180 32099I–01/2012 AT32UC3L016/32/64 12.7.17 Wake Cause Register Name: WCAUSE Access Type: Read-only Offset: 0x184 Reset Value: Latest Wake Source 31 30 29 28 27 26 25 24 18 17 16 10 9 8 2 1 0 WCAUSE[31:24] 23 22 21 20 19 WCAUSE[23:16] 15 14 13 12 11 WCAUSE[15:8] 7 6 5 4 3 WCAUSE[7:0] A bit in this register is set on wake up caused by the peripheral referred to in Table 12-12 on page 181. Table 12-12. Wake Cause Bit Wake Cause 0 CAT 1 ACIFB 2 ADCIFB 3 TWI Slave 0 4 TWI Slave 1 5 WAKE_N 6 ADCIFB Pen Detect 15:7 - 16 EIC 17 AST 31:18 - 181 32099I–01/2012 AT32UC3L016/32/64 12.7.18 Asynchronous Wake Up Enable Register Name: AWEN Access Type: Read/Write Offset: 0x188 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 AWEN[31:24] 23 22 21 20 AWEN[23:16] 15 14 13 12 AWEN[15:8] 7 6 5 4 AWEN[7:0] Each bit in this register corresponds to an asynchronous wake-up source, according to Table 12-13 on page 182. 0: The corresponding wake-up source is disabled. 1: The corresponding wake-up source is enabled. Table 12-13. Asynchronous Wake-up Sources Bit Asynchronous Wake-up Source 0 CAT 1 ACIFB 2 ADCIFB 3 TWIS0 4 TWIS1 5 WAKEN 6 ADCIFBPD 31:7 - 182 32099I–01/2012 AT32UC3L016/32/64 12.7.19 Configuration Register Name: CONFIG Access Type: Read-Only Offset: 0x3F8 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 HSBPEVC - - - PBD PBC PBB PBA This register shows the configuration of the PM. • HSBPEVC:HSB PEVC Clock Implemented 0: HSBPEVC not implemented. 1: HSBPEVC implemented. • PBD: PBD Implemented 0: PBD not implemented. 1: PBD implemented. • PBC: PBC Implemented 0: PBC not implemented. 1: PBC implemented. • PBB: PBB Implemented 0: PBB not implemented. 1: PBB implemented. • PBA: PBA Implemented 0: PBA not implemented. 1: PBA implemented. 183 32099I–01/2012 AT32UC3L016/32/64 12.7.20 Version Register Name: VERSION Access Type: Read-Only Offset: 0x3FC Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant Number Reserved. No functionality associated. • VERSION: Version Number Version number of the module. No functionality associated. 184 32099I–01/2012 AT32UC3L016/32/64 12.8 Module Configuration The specific configuration for each PM instance is listed in the following tables. The module bus clocks listed here are connected to the system bus clocks. Please refer to the “Synchronous Clocks”, “Peripheral Clock Masking” and “Sleep Modes” sections for details. Table 12-14. Power Manager Clocks Clock Name Description CLK_PM Clock for the PM bus interface Table 12-15. Register Reset Values Register Reset Value VERSION 0x00000411 Table 12-16. Effect of the Different Reset Events Power-On Reset External Reset Watchdog Reset BOD Reset SM33 Reset CPU Error Reset OCD Reset JTAG Reset CPU/HSB/PBx (excluding Power Manager) Y Y Y Y Y Y Y Y 32KHz oscillator Y N N N N N N N RC Oscillator Calibration register Y N N N N N N N Other oscillator control registers Y Y Y Y Y Y Y Y AST register, except interrupt registers Y N N N N N N N Watchdog control register Y Y N Y Y Y Y Y Voltage Calibration register Y N N N N N N N SM33 control register Y Y Y Y Y Y Y Y BOD control register Y Y Y N Y Y Y Y Clock control registers Y Y Y Y Y Y Y Y OCD system and OCD registers Y Y N Y Y Y N Y 185 32099I–01/2012 AT32UC3L016/32/64 13. System Control Interface (SCIF) Rev: 1.0.2.2 13.1 Features • • • • • • • • • • • • • 13.2 Supports crystal oscillator 0.45-16MHz (OSC0) Supports Digital Frequency Locked Loop 40-150MHz (DFLL) Supports 32KHz ultra-low-power oscillator (OSC32K) Supports 32kHz RC oscillator (RC32K) Integrated low-power RC oscillator (RCSYS) Generic clocks (GCLK) with wide frequency range provided Controls Brown-out detectors (BOD) and supply monitors Controls Bandgap Controls Voltage Regulator (VREG) behavior and calibration Controls Temperature Sensor Controls Supply Monitor 33 (SM33) operating modes and calibration Controls 120MHz integrated RC Oscillator (RC120M) Four 32-bit general-purpose backup registers Overview The System Control Interface (SCIF) controls the oscillators, Generic Clocks, BODs, Bandgap, VREG, Temperature Sensor, and Backup Registers. 13.3 I/O Lines Description Table 13-1. 13.4 I/O Lines Description Pin Name Pin Description Type RC32OUT RC32 output at start-up Output XIN0 Crystal 0 input Analog/Digital XIN32 Crystal 32 input (primary location) Analog/Digital XIN32_2 Crystal 32 input (secondary location) Analog/Digital XOUT0 Crystal 0 output Analog XOUT32 Crystal 32 output (primary location) Analog XOUT32_2 Crystal 32 output (secondary location) Analog GCLK4-GCLK0 Generic clock output Output Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 13.4.1 I/O Lines The SCIF provides a number of generic clock outputs, which can be connected to output pins, multiplexed with GPIO lines. The programmer must first program the GPIO controller to assign these pins to their peripheral function. If the I/O pins of the SCIF are not used by the application, they can be used for other purposes by the GPIO controller. Oscillator pins are also multiplexed 186 32099I–01/2012 AT32UC3L016/32/64 with GPIO. When oscillators are used, the related pins are controlled directly by the SCIF, overriding the GPIO settings. RC32OUT will be output after reset, and the GPIO controller can assign this pin to other peripheral function after start-up. 13.4.2 Power Management The BOD and the oscillators, except the 32 KHz oscillator (OSC32K), are turned off in some sleep modes and turned automatically on when the device wakes up. The Voltage Regulator is set in low power mode in some sleep modes and automatically set back in normal mode when the device wakes up. Please refer to the Power Manager chapter for details. The BOD control registers will not be reset by the Power Manager on a BOD reset. 13.4.3 Clocks The SCIF controls all oscillators in the device. The oscillators can be used as source for the CPU and peripherals. Selection of source is done in the Power Manager. The oscillators can also be used as source for generic clocks. 13.4.4 Interrupts The SCIF interrupt request line is connected to the interrupt controller. Using the SCIF interrupt requires the interrupt controller to be programmed first. 13.4.5 13.5 13.5.1 Debug Operation The SCIF does not interact with debug operations. Functional Description Oscillator (OSC) Operation Rev: 1.0.0.0 The main oscillator (OSCn) is designed to be used with an external 0.450 to 16MHz crystal and two biasing capacitors, as shown in the Electrical Characteristics chapter, or with an external clock connected to the XIN. The oscillator can be used as source for the main clock in the device, as described in the Power Manager chapter. The oscillator can be used as source for the generic clocks, as described in the Generic Clocks section. The oscillator is disabled by default after reset. When the oscillator is disabled, the XIN and XOUT pins can be used as general purpose I/Os. When the oscillator is enabled, the XIN and XOUT pins are controlled directly by the SCIF, overriding GPIO settings. When the oscillator is configured to use an external clock, the clock must be applied to the XIN pin while the XOUT pin can be used as general purpose I/O. The oscillator is enabled by writing a one to the Oscillator Enable bit in the Oscillator Control register (OSCCTRLn.OSCEN). Operation mode (external clock or crystal) is selected by writing to the Oscillator Mode bit in OSCCTRLn (OSCCTRLn.MODE). The oscillator is automatically disabled in certain sleep modes to reduce power consumption, as described in the Power Manager chapter. After a hard reset, or when waking up from a sleep mode where the oscillators were disabled, the oscillator will need a certain amount of time to stabilize on the correct frequency. This startup time can be set in the OSCCTRLn register. 187 32099I–01/2012 AT32UC3L016/32/64 The SCIF masks the oscillator outputs during the start-up time, to ensure that no unstable clocks propagate to the digital logic. The OSCn Ready bit in the Power and Clock Status Register (PCLKSR.OSCnRDY) is set when the oscillator is stable and ready to be used as clock source. An interrupt can be generated on a zero-to-one transition on OSCnRDY if the OSCnRDY bit in the Interrupt Mask Register (IMR.OSCnRDY) is set. This bit is set by writing a one to the corresponding bit in the Interrupt Enable Register (IER.OSCnRDY). 13.5.2 32KHz Oscillator (OSC32K) Operation Rev: 1.0.1.1 The 32KHz oscillator operates as described for the oscillator above. The 32KHz oscillator can be used as source clock for the Asynchronous Timer (AST) and the Watchdog Timer (WDT). The 32KHz oscillator can also be used as source for the generic clocks. The oscillator is disabled by default after reset. When the oscillator is disabled, the XIN32 and XOUT32 pins can be used as general-purpose I/Os. When the oscillator is enabled, the XIN32 and XOUT32 pins are controlled directly by the SCIF, overriding GPIO settings. When the oscillator is configured to use an external clock, the clock must be applied to the XIN32 pin while the XOUT32 pin can be used as general-purpose I/O. The oscillator is enabled writing a one to the OSC32 Enable bit in the 32KHz Oscillator Control Register (OSCCTRL32OSC32EN). The oscillator is disabled by writing a zero to the OSC32EN bit, while keeping the other bits unchanged. Writing to OSC32EN while also writing to other bits may result in unpredictable behavior. Operation mode (external clock or crystal) is selected by writing to the Oscillator Mode bit in OSCCTRL32 (OSCCTRL32.MODE). The oscillator is an ultra-low-power design and remains enabled in all sleep modes. The start-up time of the 32KHz oscillator is selected by writing to the Oscillator Start-up Time field in the OSCCTRL32 register (OSCCTRL32.STARTUP). The SCIF masks the oscillator output during the start-up time, to ensure that no unstable clock cycles propagate to the digital logic. The OSC32 Ready bit in the Power and Clock Status Register (PCLKSR.OSC32RDY) is set when the oscillator is stable and ready to be used as clock source. An interrupt can be generated on a zero-to-one transition on PCLKSR.OSC32RDY if the OSC32RDY bit in the Interrupt Mask Register (IMR.OSC32RDY) is set. This bit is set by writing a one to the corresponding bit in the Interrupt Enable Register (IER.OSC32RDY). .As a crystal oscillator usually requires a very long start-up time (up to 1 second), the 32 KHz oscillator will keep running across resets, except a Power-on Reset (POR). The 32KHz oscillator also has a 1KHz output. This is enabled by writing a one to the Enable 1KHz output bit in OSCCTRL32 register (OSCCTRL32.EN1K). If the 32KHz output clock is not needed when 1K is enabled, this can be disabled by writing a zero to the Enable 32KHz output bit in the OSCCTRL32 register (OSCCTRL32.EN32K). OSCCTRL32.EN32K is set after a POR. The 32 KHz oscillator has two possible sets of pins. To select between them write to the Pin Select bit in the OSCCTRL32 register (OSCCTRL32.PINSEL). If the 32KHz oscillator is to be used in Shutdown mode, PINSEL must be written to one, and XIN32_2 and XOUT32_2 must be used. 188 32099I–01/2012 AT32UC3L016/32/64 13.5.3 Digital Frequency Locked Loop (DFLL) Operation Rev: 2.0.1.1 The DFLL is controlled by the Digital Frequency Locked Loop Interface (DFLLIF). The DFLL is disabled by default, but can be enabled to provide a high-frequency source clock for synchronous and generic clocks. Features: • Internal oscillator with no external components • 40-150MHz frequency in closed loop mode • Can operate standalone as a high-frequency programmable oscillator in open loop mode • Can operate as an accurate frequency multiplier against a known frequency in closed loop mode • Optional spread-spectrum clock generation • Very high-frequency multiplication supported - can generate all frequencies from a 32KHz clock The DFLL can operate in both open loop mode and closed loop mode. In closed loop mode a low frequency clock with high accuracy can be used as reference clock to get high accuracy on the output clock (CLK_DFLL). To prevent unexpected writes due to software bugs, write access to the configuration registers is protected by a locking mechanism. For details please refer to the UNLOCK register description. Figure 13-1. DFLLIF Block Diagram COARSE 8 DFLL CLK_DFLL FINE 9 8+9 32 CSTEP FSTEP IMUL FMUL FREQUENCY TUNER CLK_DFLLIF_DITHER CLK_DFLLIF_REF DFLLLOCKF DFLLLOCKLOSTF DFLLLOCKC DFLLLOCKA DFLLLOCKLOSTC DFLLLOCKLOSTA 13.5.3.1 Enabling the DFLL The DFLL is enabled by writing a one to the Enable bit (EN) in the DFLLn Configuration Register (DFLLnCONF). No other bits or fields in DFLLnCONF must be changed simultaneously, or before the DFLL is enabled. 189 32099I–01/2012 AT32UC3L016/32/64 13.5.3.2 Internal synchronization Due to multiple clock domains in the DFLLIF, values in the DFLLIF configuration registers need to be synchronized to other clock domains. The status of this synchronization can be read from the Power and Clocks Status Register (PCLKSR). Before writing to a DFLLIF configuration register, the user must check that the DFLLn Synchronization Ready bit (DFLLnRDY) in PCLKSR is set. When this bit is set, the DFLL can be configured, and CLK_DFLL is ready to be used. Any write to a DFLLIF configuration register while DFLLnRDY is cleared will be ignored. Before reading the value in any of the DFLL configuration registers a one must be written to the Synchronization bit (SYNC) in the DFLLn Synchronization Register (DFLLnSYNC). The DFLL configuration registers are ready to be read when PCLKSR.DFLLnRDY is set. 13.5.3.3 Disabling the DFLL The DFLL is disabled by writing a zero to DFLLnCONF.EN. No other bits or fields in DFLLnCONF must be changed simultaneously. After disabling the DFLL, PCLKSR.DFLLnRDY will not be set. It is not required to wait for PCLKSR.DFLLnRDY to be set before re-enabling the DFLL. 13.5.3.4 Open loop operation After enabling the DFLL, open loop mode is selected by writing a zero to the Mode Selection bit (MODE) in DFLLnCONF. When operating in open loop mode the output frequency of the DFLL will be determined by the values written to the Coarse Calibration Value field (COARSE) and the Fine Calibration Value field (FINE) in the DFLLnCONF register. When writing to COARSE and FINE, be aware that the output frequency must not exceed the maximum frequency of the device after the division in the clock generator. It is possible to change the value of COARSE and FINE, and thereby the output frequency of the DFLL, while the DFLL is enabled and in use. The DFLL clock is ready to be used when PCLKSR.DFLLnRDY is cleared after enabling the DFLL. The frequency range in open loop mode is 40-150MHz, but maximum frequency can be higher, and the minimum frequency can be lower. The best way to start the DFLL at a specific frequency in open loop mode is to first configure it for closed loop mode, see Section 13.5.3.5. When a lock is achieved, read back the COARSE and FINE values and switch to open loop mode using these values. An alternative approach is to use the Frequency Meter (FREQM) to monitor the DFLL frequency and adjust the COARSE and FINE values based on measurement results form the FREQM. Please refer to the FREQM chapter for more information on how to use it. Note that the output frequency of the DFLL will drift when in open loop mode due to temperature and voltage changes. Please refer to the Electrical Characteristics chapter for details. 13.5.3.5 Closed loop operation The DFLL must be correctly configured before closed loop operation can be enabled. After enabling the DFLL, enable and select a reference clock (CLK_DFLLIF_REF). CLK_DFLLIF_REF is a generic clock, please refer to Generic Clocks section for details. Then set the maximum step size allowed in finding the COARSE and FINE values by setting the Coarse Maximum Step field (CSTEP) and Fine Maximum Step field (FSTEP) in the DFLLn Maximum Step Register (DFLLnSTEP). A small step size will ensure low overshoot on the output frequency, but can typically result in longer lock times. A high value might give a big overshoot, but can typically give faster locking. DFLLnSTEP.CSTEP and DFLLnSTEP.FSTEP must be lower than 50% of the maximum value of DFLLnCONF.COARSE and DFLLnCONF.FINE respectively. Then select the multiplication factor in the Integer Multiply Factor field (IMUL) and 190 32099I–01/2012 AT32UC3L016/32/64 the Fractional Multiply field (FMUL) in the DFLLn Multiplier Register (DFLLnMUL). Care must be taken when choosing IMUL and FMUL so the output frequency does not exceed the maximum frequency of the device. Start the closed loop mode by writing a one to DFLLnCONF.MODE bit. The frequency of CLK_DFLL (fDFLL) is given by: FMUL⎞ f DFLL = ⎛ IMUL + ---------------- f 16 ⎠ REF ⎝ 2 where fREF is the frequency of CLK_DFLLIF_REF. COARSE and FINE in DFLLnCONF are readonly in closed loop mode, and are controlled by the DFLLIF to meet user specified frequency. The values in COARSE when the closed loop mode is enabled is used by the frequency tuner as a starting point for COARSE. Setting COARSE to a value close to the final value will reduce the time needed to get a lock on COARSE. Frequency locking The locking of the frequency in closed loop mode is divided into three stages. In the COARSE stage the control logic quickly finds the correct value for DFLLnCONF.COARSE and thereby sets the output frequency to a value close to the correct frequency. The DFLLn Locked on Coarse Value bit (DFLLnLOCKC) in PCLKSR will be set when this is done. In the FINE stage the control logic tunes the value in DFLLnCONF.FINE so the output frequency will be very close to the desired frequency. DFLLn Locked on Fine Value bit (DFLLnLOCKF) in PCLKSR will be set when this is done. In the ACCURATE stage the DFLL frequency tuning mechanism uses dithering on the FINE bits to obtain an accurate average output frequency. DFLLn Locked on Accurate Value bit (DFLLnLOCKA) in PCLKSR will be set when this is done. The ACCURATE stage will only be executed if the Dithering Enable bit (DITHER) in DFLLnCONF has been written to a one. If DITHER is written to a zero DFLLnLOCKA will never occur. If dithering is enabled, the frequency of the dithering is decided by a generic clock (CLK_DFLLIF_DITHER). This clock has to be set up correctly before enabling dithering. Please refer to the Generic Clocks section for details. Figure 13-2. DFLL Closed loop State Diagram Measure fDFLLn DFLLnLOCKC 1 DFLLnLOCKF 1 DITHER 1 DFLLnLOCKA 0 0 0 0 Calculate new COARSE value Calculate new FINE value Compensate for drift Calculate new dithering dutycycle 1 Compensate for drift When dithering is enabled the accuracy of the average output frequency of the DFLL will be higher. However, the actual frequency will be alternating between two frequencies. If a fixed frequency is required, the dithering should not be enabled. 191 32099I–01/2012 AT32UC3L016/32/64 Figure 13-3. DFLL Locking in Closed loop Target frequency Initial frequency DFLLnLOCKC DFLLnLOCKF DFLLnLOCKA CLK_DFLL is ready to be used when the DFLLn Synchronization Ready bit (DFLLnRDY) in PCLKSR is set after enabling the DFLL. However, the accuracy of the output frequency depends on which locks are set. For lock times, please refer to the Electrical Characteristics chapter. Drift compensation The frequency tuner will automatically compensate for drift in the fDFLL without losing either of the locks. If the FINE value overflows or underflows, which should normally not happen, but could occur due to large drift in temperature and voltage, all locks will be lost, and the COARSE and FINE values will be recalibrated as described earlier. If any lock is lost the corresponding bit in PCLKSR will be set, DFLLn Lock Lost on Coarse Value bit (DFLLnLOCKLOSTC) for lock lost on COARSE value, DFLLn Lock Lost on Fine Value bit (DFLLnLOCKLOSTF) for lock lost on FINE value and DFLLn Lock Lost on Accurate Value bit (DFLLnLOCKLOSTA) for lock lost on ACCURATE value. The corresponding lock status bit will be cleared when the lock lost bit is set, and vice versa. Reference clock stop detection If CLK_DFLLIF_REF stops or is running at a very slow frequency, the DFLLn Reference Clock Stopped bit (DFLLnRCS) in PCLKSR will be set. Note that the detection of the clock stop will take a long time. The DFLLIF operate as if it was in open loop mode if it detects that the reference clock has stopped. This means that the COARSE and FINE values will be kept constant while PCLKSR.DFLLnRCS is set. Closed loop mode operation will automatically resume if the CLK_DFLLIF_REF is restarted, and compensate for any drift during the time CLK_DFLLIF_REF was stopped. No locks will be lost. Frequency error measurement The ratio between CLK_DFLLIF_REF and CLK_DFLL is measured automatically by the DFLLIF. The difference between this ratio and DFLLnMUL is stored in the Multiplication Ratio Difference field (RATIODIFF) in the DFLLn Ratio Register (DFLLnRATIO). The relative error on CLK_DFLL compared to the target frequency can be calculated as follows: 192 32099I–01/2012 AT32UC3L016/32/64 RATIODIFF ⋅ f REF error = -----------------------------------------------2 NUMREF ⋅ f DFLL where 2 NUMREF is the number of reference clock cycles the DFLLIF is using for calculating the ratio. 13.5.3.6 Spread Spectrum Generator (SSG) When the DFLL is used as the main clock source for the device, the EMI radiated from the device will be synchronous to fDFLL. To provide better Electromagnetic Compatibility (EMC) the DFLLIF can provide a clock with the energy spread in the frequency domain. This is done by adding or subtracting values from the FINE value. SSG is enabled by writing a one to the Enable bit (EN) in the DFLLn Spread Spectrum Generator Control Register (DFLLnSSG). A generic clock sets the rate at which the SSG changes the frequency of the DFLL clock to generate a spread spectrum (CLK_DFLLIF_DITHER). This is the same clock used by the dithering mechanism. The frequency of this clock should be higher than fREF to ensure that the DFLLIF can lock. Please refer to the Generic clocks section for details. Optionally, the clock ticks can be qualified by a Pseudo Random Binary Sequence (PRBS) if the PRBS bit in DFLLnSSG is one. This reduces the modulation effect of CLK_DFLLIF_DITHER frequency onto fDFLL. The amplitude of the frequency variation can be selected by setting the SSG Amplitude field (AMPLITUDE) in DFLLnSSG. If AMPLITUDE is zero the SSG will toggle on the LSB of the FINE value. If AMPLITUDE is one the SSG will add the sequence {1,-1, 0} to FINE. The step size of the SSG is selected by writing to the SSG Step Size field (STEPSIZE) in DFLLnSSG. STEPSIZE equal to zero or one will result in a step size equal to one. If the step size is set to n, the output value from the SSG will be incremented/decremented by n on every tick of the source clock. The Spread Spectrum Generator is available in both open and closed loop mode. When spread spectrum is enabled in closed loop mode, and the AMPLITUDE is high, an overflow/underflow in FINE is more likely to occur. Figure 13-4. Spread Spectrum Generator Block Diagram. FINE CLK_DFLLIF_DITHER Pseudorandom Binary Sequence To DFLL 1 Spread Spectrum Generator 9 0 PRBS 13.5.3.7 AMPLITUDE, STEPSIZE Wake from sleep modes When waking up from a sleep mode where the DFLL has been turned off, and CLK_DFLL was the main clock before going to sleep, the DFLL will be re-enabled and start running with the same configuration as before it was stopped even if the reference clock is not available. The 193 32099I–01/2012 AT32UC3L016/32/64 locks will not be lost. When the reference clock has restarted, the FINE tracking will quickly compensate for any frequency drift during sleep. 13.5.3.8 Accuracy There are mainly three factors that decide the accuracy of the fDFLL. These can be tuned to obtain maximum accuracy when fine lock is achieved. • FINE resolution: The frequency step between two FINE values. This is relatively smaller for high output frequencies. • Resolution of the measurement: If the resolution of the measured fDFLL is low, i.e. the ratio between CLK_DFLL frequency and CLK_DFLLIF_REF is small, then the DFLLIF might lock at a frequency that is lower than the targeted frequency. It is recommended to use a reference clock frequency of 32 KHz or lower to avoid this issue for low target frequencies. • The accuracy of the reference clock. 13.5.3.9 Interrupts A interrupt can be generated on a zero-to-one transaction on DFLLnLOCKC, DFLLnLOCKF, DFLLnLOCKA, DFLLnLOCKLOSTC, DFLLnLOCKLOSTF, DFLLnLOCKLOSTA, DFLLnRDY or DFLLnRCS. 13.5.4 Brown-Out Detection (BOD) Rev: 1.0.1.0 The Brown-Out Detector monitors the VDDCORE supply pin and compares the supply voltage to the brown-out detection level. The BOD is disabled by default, and is enabled by writing to the BOD Control field in the BOD Control Register (BOD.CTRL). This field can also be updated by flash fuses. The BOD is powered by VDDIO and will not be powered during Shutdown sleep mode. To prevent unexpected writes to the BOD register due to software bugs, write access to this register is protected by a locking mechanism. For details please refer to the UNLOCK register description. To prevent further modifications by software, the content of the BOD register can be set as readonly by writing a one to the Store Final Value bit (BOD.SFV). When this bit is one, software can not change the BOD register content. This bit is cleared after flash calibration and after a reset except after a BOD reset. The brown-out detection level is selected by writing to the BOD Level field in BOD (BOD.LEVEL). Please refer to the Electrical Characteristics chapter for parametric details. If the BOD is enabled (BOD.CTRL is one or two) and the supply voltage goes below the detection level, the Brown-Out Detection bit in the Power and Clocks Status Register (PCLKSR.BODDET) is set. This bit is cleared when the supply voltage goes above the detection level. An interrupt request will be generated on a zero-to-one transition on PCLKSR.BODDET if the Brown-Out Detection bit in the Interrupt Mask Register (IMR.BODDET) is set. This bit is set by writing a one to the corresponding bit in the Interrupt Enable Register (IER.BODDET). If BOD.CTRL is one, a BOD reset will be generated when the supply voltage goes below the detection level. If BOD.CTRL is two, the device will not be reset. Writing a one to the BOD Hysteresis bit in BOD (BOD.HYST) will add a hysteresis on the BOD detection level. 194 32099I–01/2012 AT32UC3L016/32/64 Note that the BOD must be disabled before changing BOD.LEVEL, to avoid spurious reset or interrupt. After enabling the BOD, the BOD output will be masked during one half of a RCSYS clock cycle and two main clocks cycles to avoid false results. When the JTAG or aWire is enabled, the BOD reset and interrupt are masked. The CTRL, HYST, and LEVEL fields in the BOD Control Register are loaded factory defined calibration values from flash fuses after a reset. If the Flash Calibration Done bit in the BOD Control Register (BOD.FCD) is zero, the flash calibration will be redone after any reset, and the BOD.FCD bit will be set before program execution starts in the CPU. If BOD.FCD is one, the flash calibration is redone after any reset except for a BOD reset. The BOD.FCD bit is cleared after a reset, except for a BOD reset. BOD.FCD is set when these fields have been updated after a flash calibration. It is possible to override the values in the BOD.CTRL, BOD.HYST, and BOD.LEVEL fields after reset by writing to the BOD Control Register. Please refer to the Fuse Settings chapter for more details about BOD fuses and how to program the fuses. Figure 13-5. BOD Block Diagram VDDCORE BOD Hyst BOD Level Interrupt Reset Reset SCIF POWER MANAGER(PM) 13.5.5 Enable BOD Bod Detected POR18 INTC Bandgap Rev: 1.0.1.0 The flash memory, the BOD, and the Temperature Sensor need a stable voltage reference to operate. This reference voltage is provided by an internal Bandgap voltage reference. This refer- 195 32099I–01/2012 AT32UC3L016/32/64 ence is automatically turned on at start-up and turned off during some sleep modes to save power. The Bandgap reference is powered by the internal regulator supply voltage and will not be powered during Shutdown sleep mode. Please refer to the Power Manager chapter for details. 13.5.6 System RC Oscillator (RCSYS) Rev: 1.0.1.0 The system RC oscillator has a startup time of three cycles, and is always available except in some sleep modes. Please refer to the Power Manager chapter for details. The system RC oscillator operates at a nominal frequency of 115kHz, and is calibrated using the Calibration Value field (CALIB) in the RC Oscillator Calibration Register (RCCR). After a Power-on Reset (POR), the RCCR.CALIB field is loaded with a factory defined value stored in the Flash fuses. Please refer to the Fuse setting chapter for more details about RCCR fuses and how to program the fuses. If the Flash Calibration Done (FCD) bit in the RCCR is zero at any reset, the flash calibration will be redone and the RCCR.FCD bit will be set before program execution starts in the CPU. If the RCCR.FCD is one, the flash calibration will only be redone after a Power-on Reset. To prevent unexpected writes to RCCR due to software bugs, write access to this register is protected by a locking mechanism. For details please refer to the UNLOCK register description. Although it is not recommended to override default factory settings, it is still possible to override the default values by writing to RCCR.CALIB. 13.5.7 Voltage Regulator (VREG) Rev: 1.0.1.0 The embedded voltage regulator can be used to provide the VDDCORE voltage from the internal regulator supply voltage. It is controlled by the Voltage Regulator Calibration Register (VREGCR). The voltage regulator is enabled by default at start-up but can be disabled by software if an external voltage is provided on the VDDCORE pin. The VREGCR also contains bits to control the POR33 detector. 13.5.7.1 Register protection To prevent unexpected writes to VREGCR due to software bugs, write access to this register is protected by a locking mechanism. For details please refer to the UNLOCK register description. To prevent further modifications by software, the content of the VREGCR register can be set as read-only by writing a one to the Store Final Value bit (VREGCR.SFV). Once this bit is set, software can not change the VREGCR content until a Power-on Reset (POR) is applied. 13.5.7.2 Controlling voltage regulator output The voltage regulator is always enabled at start-up, i.e. after a POR or when waking up from Shutdown mode. It can be disabled by software by writing a zero to the Enable bit (VREGCR.EN). This bit is set after a POR. Because of internal synchronization, the voltage regulator is not immediately enabled or disabled. The actual state of the voltage regulator can be read from the ON bit (VREGCR.ON). The voltage regulator output level is controlled by the Select VDD field (SELVDD) in VREGCR. The default value of this field corresponds to a regulator output voltage of 1.8V. Other values of this field are not defined, and it is not recommended to change the value of this field. 196 32099I–01/2012 AT32UC3L016/32/64 The Voltage Regulator OK bit (VREGCR.VREGOK) bit indicates when the voltage regulator output has reached the voltage threshold level. 13.5.7.3 Factory calibration After a Power-on Reset (POR) the VREGCR.CALIB field is loaded with a factory defined calibration value. This value is chosen so that the normal output voltage of the regulator after a powerup is 1.8V. Although it is not recommended to override default factory settings, it is still possible to override these default values by writing to VREGCR.CALIB. If the Flash Calibration Done bit in VREGCR (VREGCR.FCD) is zero, the flash calibration will be redone after any reset, and the VREGCR.FCD bit will be set before program execution starts in the CPU. If VREGCR.FCD is one, the flash calibration will only be redone after a POR. 13.5.7.4 POR33 control VREGCR includes control bits for the Power-on Reset 3.3V (POR33) detector that monitors the internal regulator supply voltage. The POR33 detector is enabled by default but can be disabled by software to reduce power consumption. The 3.3V Supply Monitor (SM33) can then be used to monitor the regulator power supply. The POR33 detector is disabled by writing a zero to the POR33 Enable bit (VREGCR.POR33EN). Because of internal synchronisation, the POR33 detector is not immediately enabled or disabled. The actual state of the POR33 detector can be read from the POR33 Status bit (VREGCR.POR33STATUS). The 32kHz RC oscillator (RC32K) must be enabled before disabling the POR33 detector. Once the POR33 detector has been disabled, the RC32K oscillator can be disabled again. To avoid spurious resets, it is mandatory to mask the Power-on Reset when enabling or disabling the POR33 detector. The Power-on Reset generated by the POR33 detector can be ignored by writing a one to the POR33 Mask bit (VREGCR.POR33MASK). Because of internal synchronization, the masking is not immediately effective, so software should wait for the VREGCR.POR33MASK to read as a one before assuming the masking is effective. 13.5.8 3.3 V Supply Monitor (SM33) Rev: 1.0.0.0 The 3.3V supply monitor is a specific voltage detector for the internal regulator supply voltage. It will indicate if the internal regulator supply voltage is above the minimum required input voltage threshold. The user can choose to generate either a Power-on Reset (POR) and an interrupt request, or only an interrupt request, when the internal regulator supply voltage drops below this threshold. Please refer to the Electrical Characteristics chapter for parametric details. 13.5.8.1 Register protection To prevent unexpected writes to SM33 register due to software bugs, write access to this register is protected by a locking mechanism. For details please refer to the UNLOCK register description. 197 32099I–01/2012 AT32UC3L016/32/64 To prevent further modifications by software, the content of the register can be set as read-only by writing a one to the Store Final Value bit (SM33.SFV). When this bit is one, software can not change the SM33 register content until the device is reset. 13.5.8.2 Operating modes The SM33 is disabled by default and is enabled by writing to the Supply Monitor Control field in the SM33 control register (SM33.CTRL). The current state of the SM33 can be read from the Supply Monitor On Indicator bit in SM33 (SM33.ONSM). Enabling the SM33 will disable the POR33 detector. The SM33 can operate in continuous mode or in sampling mode. In sampling mode, the SM33 is periodically enabled for a short period of time, just enough to make a a measurement, and then disabled for a longer time to reduce power consumption. By default, the SM33 operates in sampling mode during DeepStop and Static mode and in continuous mode for other sleep modes. Sampling mode can also be forced during sleep modes other than DeepStop and Static, and during normal operation, by writing a one to the Force Sampling Mode bit in the SM33 register (SM33.FS). The user can select the sampling frequency by writing to the Sampling Frequency field in SM33 (SM33.SAMPFREQ). The sampling mode uses the 32 kHz RC oscillator (RC32K) as clock source. The 32kHz RC oscillator is automatically enabled when the SM33 operates in sampling mode. 13.5.8.3 Interrupt and reset generation If the SM33 is enabled (SM33.CTRL is one or two) and the regulator supply voltage drops below the SM33 threshold, the SM33DET bit in the Power and Clocks Status Register (PCLKSR.SM33DET) is set. This bit is cleared when the supply voltage goes above the threshold. An interrupt request is generated on a zer-to-one transition of PCLKSR.SM33DET if the Supply Monitor 3.3V Detection bit in the Interrupt Mask Register (IMR.SM33DET) is set. This bit is set by writing a one to the corresponding bit in the Interrupt Enable Register (IER.SM33DET). If SM33.CTRL is one, a POR will be generated when the voltage drops below the threshold. If SM33.CTRL is two, the device will not be reset. 13.5.8.4 Factory calibration After a reset the SM33.CALIB field is loaded with a factory defined value. This value is chosen so that the nominal threshold value is 1.75V. The flash calibration is redone after any reset, and the Flash Calibration Done bit in SM33 (SM33.FCD) is set before program execution starts in the CPU. Although it is not recommended to override default factory settings, it is still possible to override the default value by writing to SM33.CALIB 13.5.9 Temperature Sensor Rev: 1.0.0.0 The Temperature Sensor is connected to an ADC channel, please refer to the ADC chapter for details. It is enabled by writing one to the Enable bit (EN) in the Temperature Sensor Configuration Register (TSENS). The Temperature Sensor can not be calibrated. Please refer to the Electrical Characteristics chapter for more details. 198 32099I–01/2012 AT32UC3L016/32/64 13.5.10 120MHz RC Oscillator (RC120M) Rev: 1.0.1.0 The 120MHz RC Oscillator can be used as source for the main clock in the device, as described in the Power Manager chapter. The oscillator can also be used as source for the generic clocks, as described in Generic Clock section. The RC120M must be enabled before it is used as a source clock. To enable the clock, the user must write a one to the Enable bit in the 120MHz RC Oscillator Control Register (RC120MCR.EN), and read back the RC120MCR register until the EN bit reads one. The clock is disabled by writing a zero to RC120MCR.EN. The EN bit must be read back as zero before the RC120M is re-enabled. If not, undefined behavior may occur. The oscillator is automatically disabled in certain sleep modes to reduce power consumption, as described in the Power Manager chapter. 13.5.11 Backup Registers (BR) Rev: 1.0.0.1 Four 32-bit backup registers are available to store values when the device is in Shutdown mode. These registers will keep their content even when the VDDCORE supply and the internal regulator supply voltage supplies are removed. The backup registers can be accessed by reading from and writing to the BR0, BR1, BR2, and BR3 registers. After writing to one of the backup registers the user must wait until the Backup Register Interface Ready bit in tne Power and Clocks Status Register (PCLKSR.BRIFARDY) is set before writing to another backup register. Writes to the backup register while PCLKSR.BRIFARDY is zero will be discarded. An interrupt can be generated on a zero-to-one transition on PCLKSR.BRIFARDY if the BRIFARDY bit in the Interrupt Mask Register (IMR.BRIFARDY) is set. This bit is set by writing a one to the corresponding bit in the Interrupt Enable Register (IER.BRIFARDY). After powering up the device the Backup Register Interface Valid bit in PCLKSR (PCLKSR.BRIFAVALID) is cleared, indicating that the content of the backup registers has not been written and contains the reset value. After writing to one of the backup registers the PCLKSR.BRIFAVALID bit is set. During writes to the backup registers (when BRIFARDY is zero) BRIFAVALID will be zero. If a reset occurs when BRIFARDY is zero, BRIFAVALID will be cleared after the reset, indicating that the content of the backup registers is not valid. If BRIFARDY is one when a reset occurs, BRIFAVALID will be one and the content is the same as before the reset. The user must ensure that BRIFAVALID and BRIFARDY are both set before reading the backup register values. 13.5.12 32kHz RC Oscillator (RC32K) Rev: 1.0.0.0 The RC32K can be used as source for the generic clocks, as described in The Generic Clocks section. The 32 kHz RC oscillator (RC32K) is forced on after reset, and output on PA20. The clock is available on the pad until the PPCR.FRC32 bit in the Power Manager has been cleared or a different peripheral function has been chosen on PA20 (PA20 will start with peripheral function F by default). Note that the forcing will only enable the clock output. To be able to use the RC32K normally the oscillator must be enabled as described below. 199 32099I–01/2012 AT32UC3L016/32/64 The oscillator is enabled by writing a one to the Enable bit in the 32kHz RC Oscillator Configuration Register (RC32KCR.EN) and disabled by writing a zero to RC32KCR.EN. The oscillator is also automatically enabled when the sampling mode is requested for the SM33. In this case, writing a zero to RC32KCR.EN will not disable the RC32K until the sampling mode is no longer requested. 13.5.13 Generic Clocks Rev: 1.0.0.0 Timers, communication modules, and other modules connected to external circuitry may require specific clock frequencies to operate correctly. The SCIF defines a number of generic clocks that can provide a wide range of accurate clock frequencies. Each generic clock runs from either clock source listed in the “Generic Clock Sources” table in the SCIF Module Configuration section. The selected source can optionally be divided by any even integer up to 512. Each clock can be independently enabled and disabled, and is also automatically disabled along with peripheral clocks by the Sleep Controller in the Power Manager. Figure 13-6. Generic Clock Generation S leep C ontroller G eneric C lock S ources M ask D ivider OSCSEL 13.5.13.1 f G C LK 0 fS R C D IV G eneric C lock 1 D IV E N CEN Enabling a generic clock A generic clock is enabled by writing a one to the Clock Enable bit (CEN) in the Generic Clock Control Register (GCCTRL). Each generic clock can individually select a clock source by writing to the Oscillator Select field (OSCSEL). The source clock can optionally be divided by writing a one to the Divide Enable bit (DIVEN) and the Division Factor field (DIV), resulting in the output frequency: f SRC f GCLK = ---------------------------2 ( DIV + 1 ) where fSRC is the frequency of the selected source clock, and fGCLK is the output frequency of the generic clock. 13.5.13.2 Disabling a generic clock A generic clock is disabled by writing a zero to CEN or entering a sleep mode that disables the PB clocks. In either case, the generic clock will be switched off on the first falling edge after the disabling event, to ensure that no glitches occur. After CEN has been written to zero, the bit will still read as one until the next falling edge occurs, and the clock is actually switched off. When 200 32099I–01/2012 AT32UC3L016/32/64 writing a zero to CEN the other bits in GCCTRL should not be changed until CEN reads as zero, to avoid glitches on the generic clock. The generic clocks will be automatically re-enabled when waking from sleep. 13.5.13.3 Changing clock frequency When changing the generic clock frequency by changing OSCSEL or DIV, the clock should be disabled before being re-enabled with the new clock source or division setting. This prevents glitches during the transition. 13.5.13.4 Generic clock allocation The generic clocks are allocated to different functions as shown in the “Generic Clock Allocation” table in the SCIF Module Configuration section. 13.5.14 Interrupts The SCIF has the following interrupt sources: • AE - Access Error: – A protected SCIF register was accessed without first being correctly unlocked. • BRIFARDY - Backup Register Interface Ready. – A 0 to 1 transition on the PCLKSR.BRIFARDY bit is detected. • DFLL0RCS - DFLL Reference Clock Stopped: – A 0 to 1 transition on the PCLKSR.DFLLRCS bit is detected. • DFLL0RDY - DFLL Ready: – A 0 to 1 transition on the PCLKSR.DFLLRDY bit is detected. • DFLL0LOCKLOSTA - DFLL lock lost on Accurate value: – A 0 to 1 transition on the PCLKSR.DFLLLOCKLOSTA bit is detected. • DFLL0LOCKLOSTF - DFLL lock lost on Fine value: – A 0 to 1 transition on the PCLKSR.DFLLLOCKLOSTF bit is detected. • DFLL0LOCKLOSTC - DFLL lock lost on Coarse value: – A 0 to 1 transition on the PCLKSR.DFLLLOCKLOSTC bit is detected. • DFLL0LOCKA - DFLL Locked on Accurate value: – A 0 to 1 transition on the PCLKSR.DFLLLOCKA bit is detected. • DFLL0LOCKF - DFLL Locked on Fine value: – A 0 to 1 transition on the PCLKSR.DFLLLOCKF bit is detected. • DFLL0LOCKC - DFLL Locked on Coarse value: – A 0 to 1 transition on the PCLKSR.DFLLLOCKC bit is detected. • BODDET - Brown out detection: – A 0 to 1 transition on the PCLKSR.BODDET bit is detected. • SM33DET - Supply Monitor 3.3V Detector: – A 0 to 1 transition on the PCLKSR.SM33DET bit is detected. • VREGOK - Voltage Regulator OK: – A 0 to 1 transition on the PCLKSR.VREGOK bit is detected. • OSC0RDY - Oscillator Ready: – A 0 to 1 transition on the PCLKSR.OSC0RDY bit is detected. 201 32099I–01/2012 AT32UC3L016/32/64 • OSC32RDY - 32KHz Oscillator Ready: – A 0 to 1 transition on the PCLKSR.OSC32RDY bit is detected. The interrupt sources will generate an interrupt request if the corresponding bit in the Interrupt Mask Register is set. The interrupt sources are ORed together to form one interrupt request. The SCIF will generate an interrupt request if at least one of the bits in the Interrupt Mask Register (IMR) is set. Bits in IMR are set by writing a one to the corresponding bit in the Interrupt Enable Register (IER), and cleared by writing a one to the corresponding bit in the Interrupt Disable Register (IDR). The interrupt request remains active until the corresponding bit in the Interrupt Status Register (ISR) is cleared by writing a one to the corresponding bit in the Interrupt Clear Register (ICR). Because all the interrupt sources are ORed together, the interrupt request from the SCIF will remain active until all the bits in ISR are cleared. 202 32099I–01/2012 AT32UC3L016/32/64 13.6 User Interface Table 13-2. SCIF Register Memory Map Offset Register Register Name Access Reset 0x0000 Interrupt Enable Register IER Write-only 0x00000000 0x0004 Interrupt Disable Register IDR Write-only 0x00000000 0x0008 Interrupt Mask Register IMR Read-only 0x00000000 0x000C Interrupt Status Register ISR Read-only 0x00000000 0x0010 Interrupt Clear Register ICR Write-only 0x00000000 0x0014 Power and Clocks Status Register PCLKSR Read-only 0x00000000 0x0018 Unlock Register UNLOCK Write-only 0x00000000 0x001C Oscillator 0 Control Register OSCCTRL0 Read/Write 0x00000000 0x0020 Oscillator 32 Control Register OSCCTRL32 Read/Write 0x00000004 0x0024 DFLL Config Register DFLL0CONF Read/Write 0x00000000 0x0028 DFLL Multiplier Register DFLL0MUL Write-only 0x00000000 0x002C DFLL Step Register DFLL0STEP Read/Write 0x00000000 0x0030 DFLL Spread Spectrum Generator Control Register DFLL0SSG Write-only 0x00000000 0x0034 DFLL Ratio Register DFLL0RATIO Read-only 0x00000000 0x0038 DFLL Synchronization Register DFLL0SYNC Write-only 0x00000000 0x003C BOD Control Register BOD Read/Write -(2) 0x0044 Voltage Regulator Calibration Register VREGCR Read/Write -(2) 0x0048 System RC Oscillator Calibration Register RCCR Read/Write -(2) 0x004C Supply Monitor 33 Calibration Register SM33 Read/Write -(2) 0x0050 Temperature Sensor Configuration Register TSENS Read/Write 0x00000000 0x0058 120MHz RC Oscillator Control Register RC120MCR Read/Write 0x00000000 0x005C-0x0068 Backup Registers BR Read/Write 0x00000000 0x006C 32kHz RC Oscillator Control Register RC32KCR Read/Write 0x00000000 0x0070 Generic Clock Control0 GCCTRL0 Read/Write 0x00000000 0x0074 Generic Clock Control1 GCCTRL1 Read/Write 0x00000000 0x0078 Generic Clock Control2 GCCTRL2 Read/Write 0x00000000 0x007C Generic Clock Control3 GCCTRL3 Read/Write 0x00000000 0x0080 Generic Clock Control4 GCCTR4 Read/Write 0x00000000 0x03C8 Oscillator 0 Version Register OSC0VERSION Read-only -(1) 0x03CC 32 KHz Oscillator Version Register OSC32VERSION Read-only -(1) 0x03D0 DFLL Version Register DFLLIFVERSION Read-only -(1) 0x03D4 BOD Version Register BODIFAVERSION Read-only -(1) 0x03D8 Voltage Regulator Version Register VREGIFBVERSION Read-only -(1) 203 32099I–01/2012 AT32UC3L016/32/64 Table 13-2. Note: SCIF Register Memory Map Offset Register Register Name Access Reset 0x03DC System RC Oscillator Version Register RCOSCIFAVERSION Read-only -(1) 0x03E0 3.3V Supply Monitor Version Register SM33IFAVERSION Read-only -(1) 0x03E4 Temperature Sensor Version Register TSENSIFAVERSION Read-only -(1) 0x03EC 120MHz RC Oscillator Version Register RC120MIFAVERSION Read-only -(1) 0x03F0 Backup Register Interface Version Register BRIFAVERSION Read-only -(1) 0x03F4 32kHz RC Oscillator Version Register RC32KIFAVERSION Read-only -(1) 0x03F8 Generic Clock Version Register GCLKVERSION Read-only -(1) 0x03FC SCIF Version Register VERSION Read-only -(1) 1. The reset value is device specific. Please refer to the Module Configuration section at the end of this chapter. 2. The reset value of this register depends on factory calibration. 204 32099I–01/2012 AT32UC3L016/32/64 13.6.1 Interrupt Enable Register Name: IER Access Type: Write-only Offset: 0x0000 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 AE - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - BRIFARDY 15 14 13 12 11 10 9 8 DFLL0RCS DFLL0RDY DFLL0LOCK LOSTA DFLL0LOCK LOSTF DFLL0LOCK LOSTC DFLL0LOCK A DFLL0LOCK F DFLL0LOCK C 7 6 5 4 3 2 1 0 BODDET SM33DET VREGOK - - - OSC0RDY OSC32RDY Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 205 32099I–01/2012 AT32UC3L016/32/64 13.6.2 Interrupt Disable Register Name: IDR Access Type: Write-only Offset: 0x0004 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 AE - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - BRIFARDY 15 14 13 12 11 10 9 8 DFLL0RCS DFLL0RDY DFLL0LOCK LOSTA DFLL0LOCK LOSTF DFLL0LOCK LOSTC DFLL0LOCK A DFLL0LOCK F DFLL0LOCK C 7 6 5 4 3 2 1 0 BODDET SM33DET VREGOK - - - OSC0RDY OSC32RDY Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 206 32099I–01/2012 AT32UC3L016/32/64 13.6.3 Interrupt Mask Register Name: IMR Access Type: Read-only Offset: 0x0008 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 AE - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - BRIFARDY 15 14 13 12 11 10 9 8 DFLL0RCS DFLL0RDY DFLL0LOCK LOSTA DFLL0LOCK LOSTF DFLL0LOCK LOSTC DFLL0LOCK A DFLL0LOCK F DFLL0LOCK C 7 6 5 4 3 2 1 0 BODDET SM33DET VREGOK - - - OSC0RDY OSC32RDY 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. A bit in this register is cleared when the corresponding bit in IDR is written to one. A bit in this register is set when the corresponding bit in IER is written to one. 207 32099I–01/2012 AT32UC3L016/32/64 13.6.4 Interrupt Status Register Name: ISR Access Type: Read-only Offset: 0x000C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 AE - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - BRIFARDY 15 14 13 12 11 10 9 8 DFLL0RCS DFLL0RDY DFLL0LOCK LOSTA DFLL0LOCK LOSTF DFLL0LOCK LOSTC DFLL0LOCK A DFLL0LOCK F DFLL0LOCK C 7 6 5 4 3 2 1 0 BODDET SM33DET VREGOK - - - OSC0RDY OSC32RDY 0: The corresponding interrupt is cleared. 1: The corresponding interrupt is pending. A bit in this register is cleared when the corresponding bit in ICR is written to one. A bit in this register is set when the corresponding interrupt occurs. 208 32099I–01/2012 AT32UC3L016/32/64 13.6.5 Interrupt Clear Register Name: ICR Access Type: Write-only Offset: 0x0010 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 AE - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - BRIFARDY 15 14 13 12 11 10 9 8 DFLL0RCS DFLL0RDY DFLL0LOCK LOSTA DFLL0LOCK LOSTF DFLL0LOCK LOSTC DFLL0LOCK A DFLL0LOCK F DFLL0LOCK C 7 6 5 4 3 2 1 0 BODDET SM33DET VREGOK - - - OSC0RDY OSC32RDY Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in ISR. 209 32099I–01/2012 AT32UC3L016/32/64 13.6.6 Power and Clocks Status Register Name: PCLKSR Access Type: Read-only Offset: 0x0014 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - BRIFAVALID - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - BRIFARDY 15 14 13 12 11 10 9 8 DFLL0RCS DFLL0RDY DFLL0LOCK LOSTA DFLL0LOCK LOSTF DFLL0LOCK LOSTC DFLL0LOCK A DFLL0LOCK F DFLL0LOCK C 7 6 5 4 3 2 1 0 BODDET SM33DET VREGOK - - - OSC0RDY OSC32RDY • BRIFAVALID: Backup Register Interface Valid 0: The values in the backup registers are not valid. 1: The values in the backup registers are valid. • BRIFARDY: Backup Register Interface Ready 0: The backup register interface is busy updating the backup registers. Writes to BRn will be discarded. 1: The backup register interface is ready to accept new writes to the backup registers. • DFLL0RCS: DFLL0 Reference Clock Stopped 0: The DFLL reference clock is running, or has never been enabled. 1: The DFLL reference clock has stopped or is too slow. • DFLL0RDY: DFLL0 Synchronization Ready 0: Read or write to DFLL registers is invalid 1: Read or write to DFLL registers is valid • DFLL0LOCKLOSTA: DFLL0 Lock Lost on Accurate Value 0: DFLL has not lost its Accurate lock or has never been enabled. 1: DFLL has lost its Accurate lock, either by disabling the DFLL or due to faulty operation. • DFLL0LOCKLOSTF: DFLL0 Lock Lost on Fine Value 0: DFLL has not lost its Fine lock or has never been enabled. 1: DFLL has lost its Fine lock, either by disabling the DFLL or due to faulty operation. • DFLL0LOCKLOSTC: DFLL0 Lock Lost on Coarse Value 0: DFLL has not lost its Coarse lock or has never been enabled. 1: DFLL has lost its Coarse lock, either by disabling the DFLL or due to faulty operation. • DFLL0LOCKA: DFLL0 Locked on Accurate Value 0: DFLL is unlocked on Accurate value. 1: DFLL is locked on Accurate value, and is ready to be selected as clock source with an accurate output clock. • DFLL0LOCKF: DFLL0 Locked on Fine Value 210 32099I–01/2012 AT32UC3L016/32/64 • • • • • • 0: DFLL is unlocked on Fine value. 1: DFLL is locked on Fine value, and is ready to be selected as clock source with a high accuracy on the output clock. DFLL0LOCKC: DFLL0 Locked on Coarse Value 0: DFLL is unlocked on Coarse value. 1: DFLL is locked on Coarse value, and is ready to be selected as clock source with medium accuracy on the output clock. BODDET: Brown-Out Detection 0: No BOD Event. 1: BOD has detected that the supply voltage is below the BOD reference value. SM33DET: Supply Monitor 3.3V Detector 0: SM33 not enabled or the supply voltage is above the SM33 threshold. 1: SM33 enabled and the supply voltage is below the SM33 threshold. VREGOK: Voltage Regulator OK 0: Voltage regulator not enabled or not ready. 1: Voltage regulator has reached its output threshold value after being enabled. OSC0RDY: OSC0 Ready 0: Oscillator not enabled or not ready. 1: Oscillator is stable and ready to be used as clock source. OSC32RDY: 32 KHz oscillator Ready 0: OSC32K not enabled or not ready. 1: OSC32K is stable and ready to be used as clock source. 211 32099I–01/2012 AT32UC3L016/32/64 13.6.7 Unlock Register Name: UNLOCK Access Type: Write-only Offset: 0x0018 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 KEY 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - 7 6 5 4 3 2 ADDR[9:8] 1 0 ADDR[7:0] To unlock a write protected register, first write to the UNLOCK register with the address of the register to unlock in the ADDR field and 0xAA in the KEY field. Then, in the next PB access write to the register specified in the ADDR field. • KEY: Unlock Key Write this bit field to 0xAA to enable unlock. • ADDR: Unlock Address Write the address offset of the register to unlock to this field. 212 32099I–01/2012 AT32UC3L016/32/64 13.6.8 Oscillator Control Register Name: OSCCTRLn Access Type: Read/Write Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - OSCEN 15 14 13 12 11 10 9 8 - - - - 7 6 5 4 3 1 0 - - - - AGC STARTUP[3:0] 2 GAIN[1:0] MODE • OSCEN: Oscillator Enable 0: The oscillator is disabled. 1: The oscillator is enabled. • STARTUP: Oscillator Start-up Time Select start-up time for the oscillator. Please refer to the “Oscillator Startup Time” table in the SCIF Module Configuration section for details. • AGC: Automatic Gain Control For test purposes. • GAIN: Gain Selects the gain for the oscillator. Please refer to the “Oscillator Gain Settings” table in the SCIF Module Configuration section for details. • MODE: Oscillator Mode 0: External clock connected on XIN. XOUT can be used as general-purpose I/O (no crystal). 1: Crystal is connected to XIN/XOUT. Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 213 32099I–01/2012 AT32UC3L016/32/64 13.6.9 32KHz Oscillator Control Register Name: OSCCTRL32 Access Type: Read/Write Reset Value: 0x00000004 31 30 29 28 27 26 25 24 RESERVED - - - - - - - 23 22 21 20 19 18 17 16 - - - - - 15 14 13 12 11 - - - - - 7 6 5 4 3 2 1 0 - - - - EN1K EN32K PINSEL OSC32EN STARTUP[2:0] 10 9 8 MODE[2:0] Note: This register is only reset by Power-On Reset • RESERVED This bit must always be written to zero. • STARTUP: Oscillator Start-up Time Select start-up time for 32 KHz oscillator Table 13-3. Start-up Time for 32 KHz Oscillator Number of RCSYS Clock Cycle Approximative Equivalent Time (RCOSC = 115 kHz) 0 0 0 1 128 1.1 ms 2 8192 72.3 ms 3 16384 143 ms 4 65536 570 ms 5 131072 1.1 s 6 262144 2.3 s 7 524288 4.6 s STARTUP 214 32099I–01/2012 AT32UC3L016/32/64 • MODE: Oscillator Mode Table 13-4. MODE Operation Mode for 32 KHz Oscillator Description 0 External clock connected to XIN32, XOUT32 can be used as general-purpose I/O (no crystal) 1 Crystal mode. Crystal is connected to XIN32/XOUT32. 2 Reserved 3 Reserved 4 Crystal and high current mode. Crystal is connected to XIN32/XOUT32. 5 Reserved 6 Reserved 7 Reserved • EN1K: 1 KHz output Enable 0: The 1 KHz output is disabled. 1: The 1 KHz output is enabled. • EN32K: 32 KHz output Enable 0: The 32 KHz output is disabled. 1: The 32 KHz output is enabled. • PINSEL: Pins Select 0: Default pins used. 1: Alternate pins: XIN32_2 pin is used instead of XIN32 pin, XOUT32_2 pin is used instead of XOUT32. • OSC32EN: 32 KHz Oscillator Enable 0: The 32 KHz Oscillator is disabled 1: The 32 KHz Oscillator is enabled 215 32099I–01/2012 AT32UC3L016/32/64 13.6.10 DFLLn Configuration Register Name: DFLLnCONF Access Type: Read/Write Reset Value: 0x00000000 31 30 29 28 27 26 25 24 COARSE[7:0] 23 22 21 20 19 18 17 16 - - - - - - - FINE[8] 15 14 13 12 11 10 9 8 FINE[7:0] 7 6 5 4 3 2 1 0 - - - - - DITHER MODE EN • COARSE: Coarse Calibration Value Set the value of the coarse calibration register. If in closed loop mode, this field is Read-only. • FINE: FINE Calibration Value Set the value of the fine calibration register. If in closed loop mode, this field is Read-only. • DITHER: Enable Dithering 0: The fine LSB input to the VCO is constant. 1: The fine LSB input to the VCO is dithered to achieve sub-LSB approximation to the correct multiplication ratio. • MODE: Mode Selection 0: The DFLL is in open loop operation. 1: The DFLL is in closed loop operation. • EN: Enable 0: The DFLL is disabled. 1: The DFLL is enabled. Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 216 32099I–01/2012 AT32UC3L016/32/64 13.6.11 DFLLn Multiplier Register Name: DFLLnMUL Access Type: Read/Write Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 IMUL[15:8] 23 22 21 20 IMUL[7:0] 15 14 13 12 FMUL[15:8] 7 6 5 4 FMUL[7:0] • IMUL: Integer Multiply Factor This field, together with FMUL, determines the ratio between fDFLL and fREFthe DFLL. IMUL is the integer part, while the FMUL is the fractional part. In open loop mode, writing to this register has no effect. • FMUL: Fractional Multiply Factor This field, together with IMUL, determines the ratio between fDFLL and fREFthe DFLL. IMUL is the integer part, while the FMUL is the fractional part. In open loop mode, writing to this register has no effect. Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 217 32099I–01/2012 AT32UC3L016/32/64 13.6.12 DFLLn Maximum Step Register Name: DFLLnSTEP Access Type: Read/Write Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - FSTEP[8] 23 22 21 20 19 18 17 16 FSTEP[7:0] 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 CSTEP[7:0] • FSTEP: Fine Maximum Step This indicates the maximum step size during fine adjustment in closed-loop mode. When adjusting to a new frequency, the expected overshoot of that frequency depends on this step size. • CSTEP: Coarse Maximum Step This indicates the maximum step size during coarse adjustment in closed-loop mode. When adjusting to a new frequency, the expected overshoot of that frequency depends on this step size. Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 218 32099I–01/2012 AT32UC3L016/32/64 13.6.13 DFLLn Spread Spectrum Generator Control Register Name: DFLLnSSG Access Type: Read/Write Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - 15 14 13 9 8 - - - 7 6 5 4 3 2 1 0 - - - - - - PRBS EN STEPSIZE[4:0] 12 11 10 AMPLITUDE[4:0] • STEPSIZE: SSG Step Size Sets the step size of the spread spectrum. • AMPLITUDE: SSG Amplitude Sets the amplitude of the spread spectrum. • PRBS: Pseudo Random Bit Sequence 0: Each spread spectrum frequency is applied at constant intervals 1: Each spread spectrum frequency is applied at pseudo-random intervals • EN: Enable 0: SSG is disabled. 1: SSG is enabled. Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 219 32099I–01/2012 AT32UC3L016/32/64 13.6.14 DFLLn Ratio Register Name: DFLLnRATIO Access Type: Read-only Reset Value: 0x00000000 31 30 29 28 27 26 25 24 18 17 16 RATIODIFF[15:8] 23 22 21 20 19 RATIODIFF[7:0] 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - NUMREF[4:0] • RATIODIFF: Multiplication Ratio Difference In closed-loop mode, this field indicates the error in the ratio between the VCO frequency and the target frequency. • NUMREF: Numerical Reference The number of reference clock cycles used to measure the VCO frequency equals 2^NUMREF. 220 32099I–01/2012 AT32UC3L016/32/64 13.6.15 DFLLn Synchronization Register Name: DFLLnSYNC Access Type: Write-only Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - SYNC • SYNC: Synchronization To be able to read the current value of DFLLnCONF or DFLLnRATIO in closed-loop mode, this bit should be written to one. The updated value is available in DFLLnCONF and DFLLnRATIO when PCLKSR.DFLLnRDY is set. 221 32099I–01/2012 AT32UC3L016/32/64 13.6.16 BOD Control Register Name: BOD Access Type: Read/Write Reset Value: - 31 30 29 28 27 26 25 24 SFV - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - FCD 15 14 13 12 11 10 9 8 - - - - - - 7 6 5 4 3 2 - HYST CTRL 1 0 LEVEL • SFV: Store Final Value 0: The register is read/write 1: The register is read-only, to protect against further accidental writes. This bit is cleared after any reset except for a BOD reset, and during flash calibration. • FCD: Fuse Calibration Done 0: The flash calibration will be redone after any reset. 1: The flash calibration will be redone after any reset except for a BOD reset. This bit is cleared after any reset, except for a BOD reset. This bit is set when the CTRL, HYST and LEVEL fields have been updated by the flash fuses after a reset. • CTRL: BOD Control Table 13-5. CTRL Operation Mode for BOD Description 0 BOD is disabled. 1 BOD is enabled and can reset the device. An interrupt request will be generated, if enabled in the IMR register. 2 BOD is enabled but cannot reset the device. An interrupt request will be generated, if enabled in the IMR register. 3 Reserved. • HYST: BOD Hysteresis 0: No hysteresis. 1: Hysteresis on. • LEVEL: BOD Level This field sets the triggering threshold of the BOD. See Electrical Characteristics for actual voltage levels. Note that any change to the LEVEL field of the BOD register should be done with the BOD deactivated to avoid spurious reset or interrupt. 222 32099I–01/2012 AT32UC3L016/32/64 Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 223 32099I–01/2012 AT32UC3L016/32/64 13.6.17 Voltage Regulator Calibration Register Name: VREGCR Access Type: Read/Write Reset Value: - 31 30 29 28 27 26 25 24 SFV INTPD - - - DBG- - - 23 22 21 20 19 18 17 16 - - - POR33MASK POR33STAT US POR33EN DEEPDIS FCD 15 14 13 12 11 10 9 8 - - - - 7 6 5 4 3 1 0 ON VREGOK EN - - CALIB 2 SELVDD • SFV: Store Final Value 0: The register is read/write. 1: The register is read-only, to protect against further accidental writes. This bit is cleared by a Power-on Reset. • INTPD: Internal Pull-down This bit is used for test purposes only. 0: The voltage regulator output is not pulled to ground. 1: The voltage regulator output has a pull-down to ground. • POR33MASK: Power-on Reset 3.3V Output Mask 0: Power-on Reset 3.3V is not masked. 1: Power-on Reset 3.3V is masked. • POR33STATUS: Power-on Reset 3.3V Status 0: Power-on Reset is disabled. 1: Power-on Reset is enabled. This bit is read-only. Writing to this bit has no effect. • POR33EN: Power-on Reset 3.3V Enable 0: Writing a zero to this bit disables the POR33 detector. 1: Writing a one to this bit enables the POR33 detector. • DEEPDIS: Disable Regulator Deep Mode 0: Regulator will enter deep mode in low-power sleep modes for lower power consumption. 1: Regulator will stay in full-power mode in all sleep modes for shorter start-up time. • FCD: Flash Calibration Done 0: The flash calibration will be redone after any reset. 1: The flash calibration will only be redone after a Power-on Reset. This bit is cleared after a Power-on Reset. This bit is set when the CALIB field has been updated by flash calibration after a reset. • CALIB: Calibration Value Calibration value for Voltage Regulator. This is calibrated during production and should not be changed. 224 32099I–01/2012 AT32UC3L016/32/64 • ON: Voltage Regulator On Status 0: The voltage regulator is currently disabled. 1: The voltage regulator is currently enabled. This bit is read-only. Writing to this bit has no effect. • VREGOK: Voltage Regulator OK Status 0: The voltage regulator is disabled or has not yet reached a stable output voltage. 1: The voltage regulator has reached the output voltage threshold level after being enabled. This bit is read-only. Writing to this bit has no effect. • EN: Enable 0: The voltage regulator is disabled. 1: The voltage regulator is enabled. Note: This bit is set after a Power-on Reset (POR). • SELVDD: Select VDD Output voltage of the Voltage Regulator. The default value of this bit corresponds to an output voltage of 1.8V. Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 225 32099I–01/2012 AT32UC3L016/32/64 13.6.18 System RC Oscillator Calibration Register Name: RCCR Access Type: Read/Write Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - FCD 15 14 13 12 11 10 9 8 - - - - - - 7 6 5 4 3 2 CALIB[9:8] 1 0 CALIB[7:0] • FCD: Flash Calibration Done 0: The flash calibration will be redone after any reset. 1: The flash calibration will only be redone after a Power-on Reset. This bit is cleared after a POR. This bit is set when the CALIB field has been updated by the flash fuses after a reset. • CALIB: Calibration Value Calibration Value for the System RC oscillator (RCSYS). Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 226 32099I–01/2012 AT32UC3L016/32/64 13.6.19 Supply Monitor 33 Calibration Register Name: SM33 Access Type: Read/Write Reset Value: - 31 30 29 28 27 - - - - 23 22 21 20 19 - - - - 15 14 13 12 - - - - 7 6 5 4 FS - - - 26 25 24 18 17 16 - ONSM SFV FCD 11 10 9 8 1 0 SAMPFREQ CALIB 3 2 CTRL • SAMPFREQ: Sampling Frequency Selects the sampling mode frequency of the 3.3V supply monitor. In sampling mode, the SM33 performs a measurement every 2(SAMPFREQ+5) cycles of the internal 32kHz RC oscillator. • ONSM: Supply Monitor On Indicator 0: The supply monitor is disabled. 1: The supply monitor is enabled. This bit is read-only. Writing to this bit has no effect. • SFV: Store Final Value 0: The register is read/write 1: The register is read-only, to protect against further accidental writes. This bit is cleared after a reset. • FCD: Flash Calibration Done This bit is cleared after a reset. This bit is set when CALIB field has been updated after a reset. • CALIB: Calibration Value Calibration Value for the SM33. • FS: Force Sampling Mode 0: Sampling mode is enabled in DeepStop and Static mode only. 1: Sampling mode is always enabled. • CTRL: Supply Monitor Control 227 32099I–01/2012 AT32UC3L016/32/64 Selects the operating mode for the SM33. Table 13-6. CTRL Operation Mode for SM33 Description 0 SM33 is disabled. 1 SM33 is enabled and can reset the device. An interrupt request will be generated if the corresponding interrupt is enabled in the IMR register. 2 SM33 is enabled and cannot reset the device. An interrupt request will be generated if the corresponding interrupt is enabled in the IMR register. 3 SM33 is disabled 4-7 Reserved Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 228 32099I–01/2012 AT32UC3L016/32/64 13.6.20 Temperature Sensor Configuration Register Name: TSENS Access Type: Read/Write Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - EN • EN: Temperature Sensor Enable 0: The Temperature Sensor is disabled. 1: The Temperature Sensor is enabled. Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 229 32099I–01/2012 AT32UC3L016/32/64 13.6.21 120MHz RC Oscillator Configuration Register Name: RC120MCR Access Type: Read/Write Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - EN • EN: RC120M Enable 0: The 120 MHz RC oscillator is disabled. 1: The 120 MHz RC oscillator is enabled. Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 230 32099I–01/2012 AT32UC3L016/32/64 13.6.22 Backup Register n Name: BRn Access Type: Read/Write Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 DATA[31:24] 23 22 21 20 DATA[23:16] 15 14 13 12 DATA[15:8] 7 6 5 4 DATA[7:0] This is a set of general-purpose read/write registers. Data stored in these registers is retained when the device is in Shutdown. Before writing to these registers the user must ensure that PCLKSR.BRIFARDY is not set. Note that this registers are protected by a lock. To write to these registers the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 231 32099I–01/2012 AT32UC3L016/32/64 13.6.23 32kHz RC Oscillator Configuration Register Name: RC32KCR Access Type: Read/Write Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - EN • EN: RC32K Enable 0: The 32 kHz RC oscillator is disabled. 1: The 32 kHz RC oscillator is enabled. Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please refer to the UNLOCK register description for details. 232 32099I–01/2012 AT32UC3L016/32/64 13.6.24 Generic Clock Control Name: GCCTRL Access Type: Read/Write Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 DIV[15:8] 23 22 21 20 DIV[7:0] 15 14 13 12 - - - 7 6 5 4 3 2 1 0 - - - - - - DIVEN CEN OSCSEL[4:0] There is one GCCTRL register per generic clock in the design. • DIV: Division Factor The number of DIV bits for each generic clock is as shown in the “Generic Clock number of DIV bits” table in the SCIF Module Configuration section. • OSCSEL: Oscillator Select Selects the source clock for the generic clock. Please refer to the “Generic Clock Sources” table in the SCIF Module Configuration section. • DIVEN: Divide Enable 0: The generic clock equals the undivided source clock. 1: The generic clock equals the source clock divided by 2*(DIV+1). • CEN: Clock Enable 0: The generic clock is disabled. 1: The generic clock is enabled. 233 32099I–01/2012 AT32UC3L016/32/64 13.6.25 Oscillator 0 Version Register Name: OSC0VERSION Access Type: Read-only Offset: 0x03C8 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 234 32099I–01/2012 AT32UC3L016/32/64 13.6.26 32KHz Oscillator Version Register Name: OSC32VERSION Access Type: Read-only Offset: 0x03CC Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 235 32099I–01/2012 AT32UC3L016/32/64 13.6.27 Digital Frequency Locked Loop Version Register Name: DFLLIF VERSION Access Type: Read-only Offset: 0x03D0 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 236 32099I–01/2012 AT32UC3L016/32/64 13.6.28 Brown-Out Detector Version Register Name: BODIFAVERSION Access Type: Read-only Offset: 0x03D4 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 237 32099I–01/2012 AT32UC3L016/32/64 13.6.29 Voltage Regulator Version Register Name: VREGIFBVERSION Access Type: Read-only Offset: 0x03D8 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 238 32099I–01/2012 AT32UC3L016/32/64 13.6.30 RC Oscillator Version Register Name: RCOSCIFAVERSION Access Type: Read-only Offset: 0x03DC Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 239 32099I–01/2012 AT32UC3L016/32/64 13.6.31 3.3V Supply Monitor Version Register Name: SM33IFAVERSION Access Type: Read-only Offset: 0x03E0 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 240 32099I–01/2012 AT32UC3L016/32/64 13.6.32 Temperature Sensor Version Register Name: TSENSIFAVERSION Access Type: Read-only Offset: 0x03E4 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 241 32099I–01/2012 AT32UC3L016/32/64 13.6.33 120MHz RC Oscillator Version Register Name: RC120MIFAVERSION Access Type: Read-only Offset: 0x03EC Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 242 32099I–01/2012 AT32UC3L016/32/64 13.6.34 Backup Register Interface Version Register Name: BRIFAVERSION Access Type: Read-only Offset: 0x03F0 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 243 32099I–01/2012 AT32UC3L016/32/64 13.6.35 32kHz RC Oscillator Version Register Name: RC32KIFAVERSION Access Type: Read-only Offset: 0x03F4 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 244 32099I–01/2012 AT32UC3L016/32/64 13.6.36 Generic Clock Version Register Name: GCLKVERSION Access Type: Read-only Offset: 0x03F8 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 245 32099I–01/2012 AT32UC3L016/32/64 13.6.37 SCIF Version Register Name: VERSION Access Type: Read-only Offset: 0x03FC Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:0] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 246 32099I–01/2012 AT32UC3L016/32/64 13.7 Module Configuration The specific configuration for each SCIF instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 13-7. SCIF Clocks Module Name Clock Name Description SCIF CLK_SCIF Clock for the SCIF bus interface Table 13-8. Oscillator Startup Time Number of System RC Oscillator Clock Cycles Approximative Equivalent time (RCSYS = 115kHz) 0 0 0 1 64 560µs 2 128 1.1ms 3 2048 18ms 4 4096 36ms 5 8192 71ms 6 16384 142ms 7 Reserved Reserved STARTUP Table 13-9. Oscillator Gain Settings GAIN[1:0] Gain Setting 0 Oscillator is used with gain G0 (XIN from 0.45 MHz to 12.0 MHz) 1 Oscillator is used with gain G1 (XIN from 12.0 MHz to 16.0 MHz) 2 Oscillator is used with gain G2 (XIN equals 16.0 MHz. Used for e.g. increasing S/N ratio, better drive strength for high ESR crystals) 3 Oscillator is used with gain G3 (XIN equals 16.0 MHz. Used for e.g. increasing S/N ratio, better drive strength for high ESR crystals) In AT32UC3L016/32/64, there are six generic clocks. These are allocated to different functions as shown in Table 13-10. Note that only GCLK4-0 are routed out. Table 13-10. Generic Clock Allocation Clock number Function 0 DFLLIF main reference and GCLK0 pin (CLK_DFLLIF_REF) 1 DFLLIF dithering and ssg reference and GCLK1 pin (CLK_DFLLIF_DITHER) 2 AST and GCLK2 pin 247 32099I–01/2012 AT32UC3L016/32/64 Table 13-10. Generic Clock Allocation Clock number 3 PWMA and GCLK3 pin 4 CAT, ACIFB and GCLK4 pin (1) GLOC 5 Note: Function 1. GCLK5 is not routed out. Table 13-11. Generic Clock Sources OSCSEL Clock/Oscillator Description 0 RCSYS System RC oscillator clock 1 OSC32K Output clock from OSC32K 2 DFLL0 Output clock from DFLL0 3 OSC0 Output clock from Oscillator0 4 RC120M Output from 120MHz RCOSC 5 CLK_CPU The clock the CPU runs on 6 CLK_HSB High Speed Bus clock 7 CLK_PBA Peripheral Bus A clock 8 CLK_PBB Peripheral Bus B clock 9 RC32K Output from 32KHz RCOSC 10 Reserved 11 CLK_1K 12-31 Reserved 1KHz output clock from OSC32K . Table 13-12. Generic Clock number of DIV bits Generic Clock Number of DIV bits 0 8 1 8 2 8 3 8 4 8 5 8 248 32099I–01/2012 AT32UC3L016/32/64 Table 13-13. Register Reset Values Register Reset Value OSC0VERSION 0x00000100 OSC32VERSION 0x00000101 DFLLIFVERSION 0x00000201 BODIFAVERSION 0x00000101 VREGIFBVERSION 0x00000101 RCOSCIFAVERSION 0x00000101 SM33IFAVERSION 0x00000100 TSENSEIFAVERSION 0x00000100 RC120MIFAVERSION 0x00000101 BRIFAVERSION 0x00000100 RC32KIFAVERSION 0x00000100 GCLKVERSION 0x00000100 VERSION 0x00000102 249 32099I–01/2012 AT32UC3L016/32/64 14. Asynchronous Timer (AST) Rev: 3.0.0.1 14.1 Features • 32-bit counter with 32-bit prescaler • Clocked Source • • • • • 14.2 – System RC oscillator (RCSYS) – 32KHz crystal oscillator (OSC32K) – PB clock – Generic clock (GCLK) – 1KHz clock from 32KHz oscillator Operation and wakeup during shutdown Optional calendar mode supported Digital prescaler tuning for increased accuracy Periodic interrupt(s) and peripheral event(s) supported Alarm interrupt(s) and peripheral event(s) supported – Optional clear on alarm Overview The Asynchronous Timer (AST) enables periodic interrupts and periodic peripheral events, as well as interrupts and peripheral events at a specified time in the future. The AST consists of a 32-bit prescaler which feeds a 32-bit up-counter. The prescaler can be clocked from five different clock sources, including the low-power 32KHz oscillator, which allows the AST to be used as a real-time timer with a maximum timeout of more than 100 years. Also, the PB clock or a generic clock can be used for high-speed operation, allowing the AST to be used as a general timer. The AST can generate periodic interrupts and peripheral events from output from the prescaler, as well as alarm interrupts and peripheral events, which can trigger at any counter value. Additionally, the timer can trigger an overflow interrupt and peripheral event, and be reset on the occurrence of any alarm. This allows periodic interrupts and peripheral events at very long and accurate intervals. To keep track of time during shutdown the AST can run while the rest of the core is powered off. This will reduce the power consumption when the system is idle. The AST can also wake up the system from shutdown using either the alarm wakeup, periodic wakeup. or overflow wakeup mechanisms. The AST has been designed to meet the system tick and Real Time Clock requirements of most embedded operating systems. 250 32099I–01/2012 AT32UC3L016/32/64 14.3 Block Diagram Figure 14-1. Asynchronous Timer Block Diagram CLK_AST CONTROL REGISTER CLK_AST CSSEL RCSYS 32-bit Prescaler CLK_AST_PRSC GCLK CLK_AST_CNT 32-bit Counter others DIGITAL TUNER REGISTER 14.4 Wake Control COUNTER VALUE Wake EN PSEL OSC32 PB clock CLK_AST WAKE ENABLE REGISTER Periodic Interrupts Alarm Interrupts PERIODIC INTERVAL REGISTER ALARM REGISTER OVF Interrupt Status and Control IRQs Events Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 14.4.1 Power Management When the AST is enabled, it will remain clocked as long as its selected clock source is running. It can also wake the CPU from the currently active sleep mode. Refer to the Power Manager chapter for details on the different sleep modes. 14.4.2 Clocks The clock for the AST bus interface (CLK_AST) is generated by the Power Manager. This clock is turned on by default, and can be enabled and disabled in the Power Manager. A number of clocks can be selected as source for the internal prescaler clock CLK_AST_PRSC. The prescaler, counter, and interrupt will function as long as this selected clock source is active. The selected clock must be enabled in the System Control Interface (SCIF). The following clock sources are available: • System RC oscillator (RCSYS). This oscillator is always enabled, except in some sleep modes. Please refer to the Electrical Characteristics chapter for the characteristic frequency of this oscillator. • 32KHz crystal oscillator (OSC32K). This oscillator must be enabled before use. • Peripheral Bus clock (PB clock). This is the clock of the peripheral bus the AST is connected to. 251 32099I–01/2012 AT32UC3L016/32/64 • Generic clock (GCLK). One of the generic clocks is connected to the AST. This clock must be enabled before use, and remains enabled in sleep modes when the PB clock is active. • 1KHz clock from the 32KHz oscillator (CLK_1K). This clock is only available in crystal mode, and must be enabled before use. In Shutdown mode only the 32 KHz oscillator and the 1KHz clock are available, using certain pins. Please refer to the Power Manager chapter for details. 14.4.3 Interrupts The AST interrupt request lines are connected to the interrupt controller. Using the AST interrupts requires the interrupt controller to be programmed first. 14.4.4 Peripheral Events The AST peripheral events are connected via the Peripheral Event System. Refer to the Peripheral Event System chapter for details. 14.4.5 Debug Operation The AST prescaler and counter is frozen during debug operation, unless the Run In Debug bit in the Development Control Register is set and the bit corresponding to the AST is set in the Peripheral Debug Register (PDBG). Please refer to the On-Chip Debug chapter in the AVR32UC Technical Reference Manual, and the OCD Module Configuration section, for details. If the AST is configured in a way that requires it to be periodically serviced by the CPU through interrupts or similar, improper operation or data loss may result during debugging. 14.5 Functional Description 14.5.1 Initialization Before enabling the AST, the internal AST clock CLK_AST_PRSC must be enabled, following the procedure specified in Section 14.5.1.1. The Clock Source Select field in the Clock register (CLOCK.CSSEL) selects the source for this clock. The Clock Enable bit in the Clock register (CLOCK.CEN) enables the CLK_AST_PRSC. When CLK_AST_PRSC is enabled, the AST can be enabled by writing a one to the Enable bit in the Control Register (CR.EN). 14.5.1.1 Enabling and disabling the AST clock The Clock Source Selection field (CLOCK.CSSEL) and the Clock Enable bit (CLOCK.CEN) cannot be changed simultaneously. Special procedures must be followed for enabling and disabling the CLK_AST_PRSC and for changing the source for this clock. To enable CLK_AST_PRSC: • Write the selected value to CLOCK.CSSEL • Wait until SR.CLKBUSY reads as zero • Write a one to CLOCK.CEN, without changing CLOCK.CSSEL • Wait until SR.CLKBUSY reads as zero To disable the clock: • Write a zero to CLOCK.CEN to disable the clock, without changing CLOCK.CSSEL • Wait until SR.CLKBUSY reads as zero 252 32099I–01/2012 AT32UC3L016/32/64 14.5.1.2 Changing the source clock The CLK_AST_PRSC must be disabled before switching to another source clock. The Clock Busy bit in the Status Register (SR.CLKBUSY) indicates whether the clock is busy or not. This bit is set when the CEN bit in the CLOCK register is changed, and cleared when the CLOCK register can be changed. To change the clock: • Write a zero to CLOCK.CEN to disable the clock, without changing CLOCK.CSSEL • Wait until SR.CLKBUSY reads as zero • Write the selected value to CLOCK.CSSEL • Wait until SR.CLKBUSY reads as zero • Write a one to CLOCK.CEN to enable the clock, without changing the CLOCK.CSSEL • Wait until SR.CLKBUSY reads as zero 14.5.2 Basic Operation 14.5.2.1 Prescaler When the AST is enabled, the 32-bit prescaler will increment on the rising edge of CLK_AST_PRSC. The prescaler value cannot be read or written, but it can be reset by writing a one to the Prescaler Clear bit in the Control Register (CR.PCLR). The Prescaler Select field in the Control Register (CR.PSEL) selects the prescaler bit PSEL as source clock for the counter (CLK_AST_CNT). This results in a counter frequency of: f PRSC f CNT = ----------------------PSEL + 1 2 where fPRSC is the frequency of the internal prescaler clock CLK_AST_PRSC. 14.5.2.2 Counter operation When enabled, the AST will increment on every 0-to-1 transition of the selected prescaler tapping. When the Calender bit in the Control Register (CR.CAL) is zero, the counter operates in counter mode. It will increment until it reaches the top value of 0xFFFFFFFF, and then wrap to 0x00000000. This sets the status bit Overflow in the Status Register (SR.OVF). Optionally, the counter can also be reset when an alarm occurs (see Section 14.5.3.2 on page 255. This will also set the OVF bit. The AST counter value can be read from or written to the Counter Value (CV) register. Note that due to synchronization, continuous reading of the CV register with the lowest prescaler setting will skip every third value. In addition, if CLK_AST_PRSC is as fast as, or faster than, the CLK_AST, the prescaler value must be 3 or higher to be able to read the CV without skipping values. 253 32099I–01/2012 AT32UC3L016/32/64 14.5.2.3 Calendar operation When the CAL bit in the Control Register is one, the counter operates in calendar mode. Before this mode is enabled, the prescaler should be set up to give a pulse every second. The date and time can then be read from or written to the Calendar Value (CALV) register. Time is reported as seconds, minutes, and hours according to the 24-hour clock format. Date is the numeral date of month (starting on 1). Month is the numeral month of the year (1 = January, 2 = February, etc.). Year is a 6-bit field counting the offset from a software-defined leap year (e.g. 2000). The date is automatically compensated for leap years, assuming every year divisible by 4 is a leap year. All peripheral events and interrupts work the same way in calendar mode as in counter mode. However, the Alarm Register (ARn) must be written in time/date format for the alarm to trigger correctly. 14.5.3 Interrupts The AST can generate five separate interrupt requests: • OVF: OVF • PER: PER0, PER1 • ALARM: ALARM0, ALARM1 • CLKREADY • READY This allows the user to allocate separate handlers and priorities to the different interrupt types. The generation of the PER interrupt is described in Section 14.5.3.1., and the generation of the ALARM interrupt is described in Section 14.5.3.2. The OVF interrupt is generated when the counter overflows, or when the alarm value is reached, if the Clear on Alarm bit in the Control Register is one. The CLKREADY interrupt is generated when SR.CLKBUSY has a 1-to-0 transition, and indicates that the clock synchronization is completed. The READY interrupt is generated when SR.BUSY has a 1-to-0 transition, and indicates that the synchronization described in Section 14.5.8 is completed. An interrupt request will be generated if the corresponding bit in the Interrupt Mask Register (IMR) is set. Bits in IMR are set by writing a one to the corresponding bit in the Interrupt Enable Register (IER), and cleared by writing a one to the corresponding bit in the Interrupt Disable Register (IDR). The interrupt request remains active until the corresponding bit in SR is cleared by writing a one to the corresponding bit in the Status Clear Register (SCR). The AST interrupts can wake the CPU from any sleep mode where the source clock and the interrupt controller is active. 14.5.3.1 Periodic interrupt The AST can generate periodic interrupts. If the PERn bit in the Interrupt Mask Register (IMR) is one, the AST will generate an interrupt request on the 0-to-1 transition of the selected bit in the 254 32099I–01/2012 AT32UC3L016/32/64 prescaler when the AST is enabled. The bit is selected by the Interval Select field in the corresponding Periodic Interval Register (PIRn.INSEL), resulting in a periodic interrupt frequency of f CS f PA = ------------------------INSEL + 1 2 where fCS is the frequency of the selected clock source. The corresponding PERn bit in the Status Register (SR) will be set when the selected bit in the prescaler has a 0-to-1 transition. Because of synchronization, the transfer of the INSEL value will not happen immediately. When changing/setting the INSEL value, the user must make sure that the prescaler bit number INSEL will not have a 0-to-1 transition before the INSEL value is transferred to the register. In that case, the first periodic interrupt after the change will not be triggered. 14.5.3.2 Alarm interrupt The AST can also generate alarm interrupts. If the ALARMn bit in IMR is one, the AST will generate an interrupt request when the counter value matches the selected alarm value, when the AST is enabled. The alarm value is selected by writing the value to the VALUE field in the corresponding Alarm Register (ARn.VALUE). The corresponding ALARMn bit in SR will be set when the counter reaches the selected alarm value. Because of synchronization, the transfer of the alarm value will not happen immediately. When changing/setting the alarm value, the user must make sure that the counter will not count the selected alarm value before the value is transferred to the register. In that case, the first alarm interrupt after the change will not be triggered. If the Clear on Alarm bit in the Control Register (CR.CAn) is one, the corresponding alarm interrupt will clear the counter and set the OVF bit in the Status Register. This will generate an overflow interrupt if the OVF bit in IMR is set. 14.5.4 Peripheral events The AST can generate a number of peripheral events: • OVF • PER0 • PER1 • ALARM0 • ALARM1 The PERn peripheral event(s) is generated the same way as the PER interrupt, as described in Section 14.5.3.1. The ALARMn peripheral event(s) is generated the same way as the ALARM interrupt, as described in Section 14.5.3.2. The OVF peripheral event is generated the same way as the OVF interrupt, as described in Section 14.5.3- 255 32099I–01/2012 AT32UC3L016/32/64 The peripheral event will be generated if the corresponding bit in the Event Mask (EVM) register is set. Bits in EVM register are set by writing a one to the corresponding bit in the Event Enable (EVE) register, and cleared by writing a one to the corresponding bit in the Event Disable (EVD) register. 14.5.5 AST wakeup The AST can wake up the CPU directly, without the need to trigger an interrupt. A wakeup can be generated when the counter overflows, when the counter reaches the selected alarm value, or when the selected prescaler bit has a 0-to-1 transition. In this case, the CPU will continue executing from the instruction following the sleep instruction. The AST wakeup is enabled by writing a one to the corresponding bit in the Wake Enable Register (WER). When the CPU wakes from sleep, the wake signal must be cleared by writing a one to the corresponding bit in SCR to clear the internal wake signal to the sleep controller. If the wake signal is not cleared after waking from sleep, the next sleep instruction will have no effect because the CPU will wake immediately after this sleep instruction. The AST wakeup can wake the CPU from any sleep mode where the source clock is active. The AST wakeup can be configured independently of the interrupt masking. 14.5.6 Shutdown Mode If the AST is configured to use a clock that is available in Shutdown mode, the AST can be used to wake up the system from shutdown. Both the alarm wakeup, periodic wakeup, and overflow wakeup mechanisms can be used in this mode. When waking up from Shutdown mode all control registers will have the same value as before the shutdown was entered, except the Interrupt Mask Register (IMR). IMR will be reset with all interrupts turned off. The software must first reconfigure the interrupt controller and then enable the interrupts in the AST to again receive interrupts from the AST. The CV register will be updated with the current counter value directly after wakeup from shutdown. The SR will show the status of the AST, including the status bits set during shutdown operation. When waking up the system from shutdown the CPU will start executing code from the reset start address. 14.5.7 Digital Tuner The digital tuner adds the possibility to compensate for a too-slow or a too-fast input clock. The ADD bit in the Digital Tuner Register (DTR.ADD) selects if the prescaler frequency should be reduced or increased. If ADD is ‘0’, the prescaler frequency is reduced: ⎛ ⎞ ⎜ ⎟ 1 f TUNED = f 0 ⎜ 1 – --------------------------------------------------------------------------------⎟ 256 ⎜ roundup ⎛ --------------------⎞ ⋅ ( 2 EXP ) + 1⎟⎠ ⎝ ⎝ VALUE⎠ where fTUNED is the tuned frequency, f0 is the original prescaler frequency, and VALUE and EXP are the corresponding fields to be programmed in DTR. Note that DTR.EXP must be greater than zero. Frequency tuning is disabled by programming DTR.VALUE as zero. 256 32099I–01/2012 AT32UC3L016/32/64 If ADD is ‘1’, the prescaler frequency is increased: ⎛ ⎞ ⎜ ⎟ 1 f TUNED = f 0 ⎜ 1 + --------------------------------------------------------------------------------⎟ 256 ⎜ roundup ⎛ --------------------⎞ ⋅ ( 2 EXP ) – 1⎟⎠ ⎝ ⎝ VALUE⎠ Note that for these formulas to be within an error of 0.01%, it is recommended that the prescaler bit that is used as the clock for the counter (selected by CR.PSEL) or to trigger the periodic interrupt (selected by PIRn.INSEL) be bit 6 or higher. 14.5.8 Synchronization As the prescaler and counter operate asynchronously from the user interface, the AST needs a few clock cycles to synchronize the values written to the CR, CV, SCR, WER, EVE, EVD, PIRn, ARn, and DTR registers. The Busy bit in the Status Register (SR.BUSY) indicates that the synchronization is ongoing. During this time, writes to these registers will be discarded and reading will return a zero value. Note that synchronization takes place also if the prescaler is clocked from CLK_AST. 257 32099I–01/2012 AT32UC3L016/32/64 14.6 User Interface Table 14-1. Offset Register Register Name Access Reset 0x00 Control Register CR Read/Write 0x00000000 0x04 Counter Value CV Read/Write 0x00000000 0x08 Status Register SR Read-only 0x00000000 0x0C Status Clear Register SCR Write-only 0x00000000 0x10 Interrupt Enable Register IER Write-only 0x00000000 0x14 Interrupt Disable Register IDR Write-only 0x00000000 0x18 Interrupt Mask Register IMR Read-only 0x00000000 0x1C Wake Enable Register WER Read/write 0x00000000 AR0 Read/Write 0x00000000 0x20 Alarm Register 0 (2) Alarm Register 1 (2) AR1 Read/Write 0x00000000 0x30 (2) Periodic Interval Register 0 PIR0 Read/Write 0x00000000 0x34 Periodic Interval Register 1(2) PIR1 Read/Write 0x00000000 0x40 Clock Control Register CLOCK Read/Write 0x00000000 0x44 Digital Tuner Register DTR Read/Write 0x00000000 0x48 Event Enable EVE Write-only 0x00000000 0x4C Event Disable EVD Write-only 0x00000000 0x50 Event Mask EVM Read-only 0x00000000 0x54 Calendar Value CALV Read/Write 0x00000000 0xF0 Parameter Register PARAMETER Read-only -(1) 0xFC Version Register VERSION Read-only -(1) 0x24 Note: AST Register Memory Map 1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter. 2. The number of Alarm and Periodic Interval registers are device specific. Please refer to the Module Configuration section at the end of this chapter. 258 32099I–01/2012 AT32UC3L016/32/64 14.6.1 Name: Control Register CR Access Type: Read/Write Offset: 0x00 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - 15 14 13 12 11 10 9 8 - - - - - - CA1 CA0 7 6 5 4 3 2 1 0 - - - - - CAL PCLR EN PSEL When the SR.BUSY bit is set, writes to this register will be discarded and this register will read as zero. • PSEL: Prescaler Select Selects prescaler bit PSEL as source clock for the counter. • CAn: Clear on Alarm n 0: The corresponding alarm will not clear the counter. 1: The corresponding alarm will clear the counter. • CAL: Calendar Mode 0: The AST operates in counter mode. 1: The AST operates in calendar mode. • PCLR: Prescaler Clear Writing a zero to this bit has no effect. Writing a one to this bit clears the prescaler. This bit always reads as zero. • EN: Enable 0: The AST is disabled. 1: The AST is enabled. 259 32099I–01/2012 AT32UC3L016/32/64 14.6.2 Name: Counter Value CV Access Type: Read/Write Offset: 0x04 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 18 17 16 11 10 9 8 3 2 1 0 VALUE[31:24] 23 22 21 20 19 VALUE[23:16] 15 14 13 12 VALUE[15:8] 7 6 5 4 VALUE[7:0] When the SR.BUSY bit is set, writes to this register will be discarded and this register will read as zero. • VALUE: AST Value The current value of the AST counter. 260 32099I–01/2012 AT32UC3L016/32/64 14.6.3 Name: Status Register SR Access Type: Read-only Offset: 0x08 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - CLKRDY CLKBUSY - - READY BUSY 23 22 21 20 19 18 17 16 - - - - - - PER1 PER0 15 14 13 12 11 10 9 8 - - - - - - ALARM1 ALARM0 7 6 5 4 3 2 1 0 - - - - - - - OVF • CLKRDY: Clock Ready This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when the SR.CLKBUSY bit has a 1-to-0 transition. • CLKBUSY: Clock Busy 0: The clock is ready and can be changed. 1: CLOCK.CEN has been written and the clock is busy. • READY: AST Ready This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when the SR.BUSY bit has a 1-to-0 transition. • BUSY: AST Busy 0: The AST accepts writes to CR, CV, SCR, WER, EVE, EVD, ARn, PIRn, and DTR. 1: The AST is busy and will discard writes to CR, CV, SCR, WER, EVE, EVD, ARn, PIRn, and DTR. • PERn: Periodic n This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when the selected bit in the prescaler has a 0-to-1 transition. • ALARMn: Alarm n This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when the counter reaches the selected alarm value. • OVF: Overflow This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when an overflow has occurred. 261 32099I–01/2012 AT32UC3L016/32/64 14.6.4 Name: Status Clear Register SCR Access Type: Write-only Offset: 0x0C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - CLKRDY - - - READY - 23 22 21 20 19 18 17 16 - - - - - - PER1 PER0 15 14 13 12 11 10 9 8 - - - - - - ALARM1 ALARM0 7 6 5 4 3 2 1 0 - - - - - - - OVF When the SR.BUSY bit is set, writes to this register will be discarded. Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in SR and the corresponding interrupt request. 262 32099I–01/2012 AT32UC3L016/32/64 14.6.5 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x10 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - CLKRDY - - - READY - 23 22 21 20 19 18 17 16 - - - - - - PER1 PER0 15 14 13 12 11 10 9 8 - - - - - - ALARM1 ALARM0 7 6 5 4 3 2 1 0 - - - - - - - OVF Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 263 32099I–01/2012 AT32UC3L016/32/64 14.6.6 Name: Interrupt Disable Register IDR Access Type: Write-only Offset: 0x14 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - CLKRDY - - - READY - 23 22 21 20 19 18 17 16 - - - - - - PER1 PER0 15 14 13 12 11 10 9 8 - - - - - - ALARM1 ALARM0 7 6 5 4 3 2 1 0 - - - - - - - OVF Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 264 32099I–01/2012 AT32UC3L016/32/64 14.6.7 Name: Interrupt Mask Register IMR Access Type: Read-only Offset: 0x18 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - CLKRDY - - - READY - 23 22 21 20 19 18 17 16 - - - - - - PER1 PER0 15 14 13 12 11 10 9 8 - - - - - - ALARM1 ALARM0 7 6 5 4 3 2 1 0 - - - - - - - OVF 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. A bit in this register is cleared when the corresponding bit in IDR is written to one. A bit in this register is set when the corresponding bit in IER is written to one. 265 32099I–01/2012 AT32UC3L016/32/64 14.6.8 Name: Wake Enable Register WER Access Type: Read/Write Offset: 0x1C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - PER1 PER0 15 14 13 12 11 10 9 8 - - - - - - ALARM1 ALARM0 7 6 5 4 3 2 1 0 - - - - - - - OVF When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero. This register enables the wakeup signal from the AST. • PERn: Periodic n 0: The CPU will not wake up from sleep mode when the selected bit in the prescaler has a 0-to-1 transition. 1: The CPU will wake up from sleep mode when the selected bit in the prescaler has a 0-to-1 transition. • ALARMn: Alarm n 0: The CPU will not wake up from sleep mode when the counter reaches the selected alarm value. 1: The CPU will wake up from sleep mode when the counter reaches the selected alarm value. • OVF: Overflow 0: A counter overflow will not wake up the CPU from sleep mode. 1: A counter overflow will wake up the CPU from sleep mode. 266 32099I–01/2012 AT32UC3L016/32/64 14.6.9 Name: Alarm Register 0 AR0 Access Type: Read/Write Offset: 0x20 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 18 17 16 11 10 9 8 3 2 1 0 VALUE[31:24] 23 22 21 20 19 VALUE[23:16] 15 14 13 12 VALUE[15:8] 7 6 5 4 VALUE[7:0] When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero. • VALUE: Alarm Value When the counter reaches this value, an alarm is generated. 267 32099I–01/2012 AT32UC3L016/32/64 14.6.10 Name: Alarm Register 1 AR1 Access Type: Read/Write Offset: 0x24 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 18 17 16 11 10 9 8 3 2 1 0 VALUE[31:24] 23 22 21 20 19 VALUE[23:16] 15 14 13 12 VALUE[15:8] 7 6 5 4 VALUE[7:0] When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero. • VALUE: Alarm Value When the counter reaches this value, an alarm is generated. 268 32099I–01/2012 AT32UC3L016/32/64 14.6.11 Name: Periodic Interval Register 0 PIR0 Access Type: Read/Write Offset: 0x30 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - INSEL When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero. • INSEL: Interval Select The PER0 bit in SR will be set when the INSEL bit in the prescaler has a 0-to-1 transition. 269 32099I–01/2012 AT32UC3L016/32/64 14.6.12 Name: Periodic Interval Register 1 PIR1 Access Type: Read/Write Offset: 0x34 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - INSEL When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero. • INSEL: Interval Select The PER1 bit in SR will be set when the INSEL bit in the prescaler has a 0-to-1 transition. 270 32099I–01/2012 AT32UC3L016/32/64 14.6.13 Name: Clock Control Register CLOCK Access Type: Read/Write Offset: 0x40 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - 7 6 5 4 3 2 1 0 - - - - - - - CEN CSSEL When writing to this register, follow the sequence in Section 14.5.1 on page 252. • CSSEL: Clock Source Selection This field defines the clock source CLK_AST_PRSC for the prescaler: Table 14-2. CSSEL Clock Source Selection Clock Source 0 System RC oscillator (RCSYS) 1 32KHz oscillator (OSC32K) 2 PB clock 3 Generic clock (GCLK) 4 1KHz clock from 32KHz oscillator (CLK_1K) • CEN: Clock Enable 0: CLK_AST_PRSC is disabled. 1: CLK_AST_PRSC is enabled. 271 32099I–01/2012 AT32UC3L016/32/64 14.6.14 Name: Digital Tuner Register DTR Access Type: Read/Write Offset: 0x44 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 VALUE 7 6 5 - - ADD 4 EXP When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero. • VALUE: 0: The frequency is unchanged. 1-255: The frequency will be adjusted according to the formula below. • ADD: ⎛ ⎞ ⎜ ⎟ 1 0: The resulting frequency is f0 ⎜ 1 – -------------------------------------------------------------------------------⎟ 256 ⎜ for VALUE > 0 . roundup ⎛ --------------------⎞ ⋅ 2 ( EXP ) + 1⎟⎠ ⎝ ⎝ VALUE⎠ ⎛ ⎞ ⎜ ⎟ 1 1: The resulting frequency is f0 ⎜ 1 + -------------------------------------------------------------------------------⎟ 256 ⎜ for VALUE > 0 . roundup ⎛ --------------------⎞ ⋅ 2 ( EXP ) – 1⎟⎠ ⎝ ⎝ VALUE⎠ • EXP: The frequency will be adjusted according to the formula above. 272 32099I–01/2012 AT32UC3L016/32/64 14.6.15 Name: Event Enable Register EVE Access Type: Write-only Offset: 0x48 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - PER1 PER0 15 14 13 12 11 10 9 8 - - - - - - ALARM1 ALARM0 7 6 5 4 3 2 1 0 - - - - - - - OVF When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero. Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in EVM. 273 32099I–01/2012 AT32UC3L016/32/64 14.6.16 Name: Event Disable Register EVD Access Type: Write-only Offset: 0x4C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - PER1 PER0 15 14 13 12 11 10 9 8 - - - - - - ALARM1 ALARM0 7 6 5 4 3 2 1 0 - - - - - - - OVF When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero. Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in EVM. 274 32099I–01/2012 AT32UC3L016/32/64 14.6.17 Name: Event Mask Register EVM Access Type: Read-only Offset: 0x50 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - PER1 PER0 15 14 13 12 11 10 9 8 - - - - - - ALARM1 ALARM0 7 6 5 4 3 2 1 0 - - - - - - - OVF 0: The corresponding peripheral event is disabled. 1: The corresponding peripheral event is enabled. This bit is cleared when the corresponding bit in EVD is written to one. This bit is set when the corresponding bit in EVE is written to one. 275 32099I–01/2012 AT32UC3L016/32/64 14.6.18 Name: Calendar Value CALV Access Type: Read/Write Offset: 0x54 Reset Value: 0x00000000 31 30 29 28 27 26 25 YEAR 23 22 21 MONTH[3:2] 20 MONTH[1:0] 15 19 18 17 DAY 14 13 12 6 16 HOUR[4] 11 10 HOUR[3:0] 7 24 9 8 1 0 MIN[5:2] 5 4 3 MIN[1:0] 2 SEC When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero. • YEAR: Year Current year. The year is considered a leap year if YEAR[1:0] = 0. • MONTH: Month 1 = January 2 = February ... 12 = December • DAY: Day Day of month, starting with 1. • HOUR: Hour Hour of day, in 24-hour clock format. Legal values are 0 through 23. • MIN: Minute Minutes, 0 through 59. • SEC: Second Seconds, 0 through 59. 276 32099I–01/2012 AT32UC3L016/32/64 14.6.19 Name: Parameter Register PARAMETER Access Type: Read-only Offset: 0xF0 Reset Value: - 31 30 29 - - - 23 22 21 - - - 15 14 13 12 11 10 PIR1WA PIR0WA - NUMPIR - - 7 6 5 4 3 2 - • • • • • • • 28 27 26 25 24 17 16 9 8 PER1VALUE 20 19 18 PER0VALUE DTEXPVALUE NUMAR 1 0 DTEXPWA DT This register gives the configuration used in the specific device. Also refer to the Module Configuration section. DT: Digital Tuner 0: Digital tuner not implemented. 1: Digital tuner implemented. DTREXPWA: Digital Tuner Exponent Writeable 0: Digital tuner exponent is a constant value. Writes to EXP field in DTR will be discarded. 1: Digital tuner exponent is chosen by writing to EXP field in DTR. DTREXPVALUE: Digital Tuner Exponent Value Digital tuner exponent value if DTEXPWA is zero. NUMAR: Number of Alarm Comparators 0: Zero alarm comparators. 1: One alarm comparator. 2: Two alarm comparators. NUMPIR: Number of Periodic Comparators 0: One periodic comparator. 1: Two periodic comparator. PIRnWA: Periodic Interval n Writeable 0: Periodic interval n prescaler tapping is a constant value. Writes to INSEL field in PIRn register will be discarded. 1: Periodic interval n prescaler tapping is chosen by writing to INSEL field in PIRn register. PERnVALUE: Periodic Interval n Value Periodic interval prescaler n tapping if PIRnWA is zero. 277 32099I–01/2012 AT32UC3L016/32/64 14.6.20 Name: Version Register VERSION Access Type: Read-only Offset: 0xFC Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant Number Reserved. No functionality associated. • VERSION: Version Number Version number of the module. No functionality associated. 278 32099I–01/2012 AT32UC3L016/32/64 14.7 Module Configuration The specific configuration for each AST instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 14-3. AST Configuration Feature AST Number of alarm comparators 1 Number of periodic comparators 1 Digital tuner On Table 14-4. AST Clocks Clock Name Description CLK_AST Clock for the AST bus interface GCLK The generic clock used for the AST is GCLK2 PB clock Peripheral Bus clock from the PBA clock domain Table 14-5. Register Reset Values Register Reset Value VERSION 0x00000300 PARAMETER 0x00004103 279 32099I–01/2012 AT32UC3L016/32/64 15. Watchdog Timer (WDT) Rev: 4.0.2.0 15.1 Features • • • • • 15.2 Watchdog Timer counter with 32-bit counter Timing window watchdog Clocked from system RC oscillator or the 32 KHz crystal oscillator Configuration lock WDT may be enabled at reset by a fuse Overview The Watchdog Timer (WDT) will reset the device unless it is periodically serviced by the software. This allows the device to recover from a condition that has caused the system to be unstable. The WDT has an internal counter clocked from the system RC oscillator or the 32 KHz crystal oscillator. The WDT counter must be periodically cleared by software to avoid a watchdog reset. If the WDT timer is not cleared correctly, the device will reset and start executing from the boot vector. 15.3 Block Diagram Figure 15-1. WDT Block Diagram PB PB Clock Domain CLR SR CTRL WDTCLR WINDOW, CLEARED EN, MODE, PSEL, TBAN SYNC RCSYS 0 OSC32K 1 CLK_CNT 32-bit Counter Watchdog Detector Watchdog Reset CEN CLK_CNT Domain CSSEL 15.4 Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 280 32099I–01/2012 AT32UC3L016/32/64 15.4.1 Power Management When the WDT is enabled, the WDT remains clocked in all sleep modes. It is not possible to enter sleep modes where the source clock of CLK_CNT is stopped. Attempting to do so will result in the device entering the lowest sleep mode where the source clock is running, leaving the WDT operational. Please refer to the Power Manager chapter for details about sleep modes. After a watchdog reset the WDT bit in the Reset Cause Register (RCAUSE) in the Power Manager will be set. 15.4.2 Clocks The clock for the WDT bus interface (CLK_WDT) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the WDT before disabling the clock, to avoid freezing the WDT in an undefined state. There are two possible clock sources for the Watchdog Timer (CLK_CNT): • System RC oscillator (RCSYS): This oscillator is always enabled when selected as clock source for the WDT. Please refer to the Power Manager chapter for details about the RCSYS and sleep modes. Please refer to the Electrical Characteristics chapter for the characteristic frequency of this oscillator. • 32 KHz crystal oscillator (OSC32K): This oscillator has to be enabled in the System Control Interface before using it as clock source for the WDT. The WDT will not be able to detect if this clock is stopped. 15.4.3 Debug Operation The WDT counter is frozen during debug operation, unless the Run In Debug bit in the Development Control Register is set and the bit corresponding to the WDT is set in the Peripheral Debug Register (PDBG). Please refer to the On-Chip Debug chapter in the AVR32UC Technical Reference Manual, and the OCD Module Configuration section, for details. If the WDT counter is not frozen during debug operation it will need periodically clearing to avoid a watchdog reset. 15.4.4 Fuses The WDT can be enabled at reset. This is controlled by the WDTAUTO fuse, see Section 15.5.4 for details. Please refer to the Fuse Settings section in the Flash Controller chapter for details about WDTAUTO and how to program the fuses. 15.5 Functional Description 15.5.1 Basic Mode 15.5.1.1 WDT Control Register Access To avoid accidental disabling of the watchdog, the Control Register (CTRL) must be written twice, first with the KEY field set to 0x55, then 0xAA without changing the other bits. Failure to do so will cause the write operation to be ignored, and the value in the CTRL Register will not be changed. 15.5.1.2 Changing CLK_CNT Clock Source After any reset, except for watchdog reset, CLK_CNT will be enabled with the RCSYS as source. 281 32099I–01/2012 AT32UC3L016/32/64 To change the clock for the WDT the following steps need to be taken. Note that the WDT should always be disabled before changing the CLK_CNT source: 1. Write a zero to the Clock Enable (CEN) bit in the CTRL Register, leaving the other bits as they are in the CTRL Register. This will stop CLK_CNT. 2. Read back the CTRL Register until the CEN bit reads zero. The clock has now been stopped. 3. Modify the Clock Source Select (CSSEL) bit in the CTRL Register with your new clock selection and write it to the CTRL Register. 4. Write a one to the CEN bit, leaving the other bits as they are in the CTRL Register. This will enable the clock. 5. Read back the CTRL Register until the CEN bit reads one. The clock has now been enabled. 15.5.1.3 Configuring the WDT If the MODE bit in the CTRL Register is zero, the WDT is in basic mode. The Time Out Prescale Select (PSEL) field in the CTRL Register selects the WDT timeout period: Ttimeout = Tpsel = 2(PSEL+1) / fclk_cnt 15.5.1.4 Enabling the WDT To enable the WDT write a one to the Enable (EN) bit in the CTRL Register. Due to internal synchronization, it will take some time for the CTRL.EN bit to read back as one. 15.5.1.5 Clearing the WDT Counter The WDT counter is cleared by writing a one to the Watchdog Clear (WDTCLR) bit in the Clear (CLR) Register, at any correct write to the CTRL Register, or when the counter reaches Ttimeout and the device is reset. In basic mode the CLR.WDTCLR can be written at any time when the WDT Counter Cleared (CLEARED) bit in the Status Register (SR) is one. Due to internal synchronization, clearing the WDT counter takes some time. The SR.CLEARED bit is cleared when writing to CLR.WDTCLR bit and set when the clearing is done. Any write to the CLR.WDTCLR bit while SR.CLEARED is zero will not clear the counter. Writing to the CLR.WDTCLR bit has to be done in a particular sequence to be valid. The CLR Register must be written twice, first with the KEY field set to 0x55 and WDTCLR set to one, then a second write with the KEY set to 0xAA without changing the WDTCLR bit. Writing to the CLR Register without the correct sequence has no effect. If the WDT counter is periodically cleared within Tpsel no watchdog reset will be issued, see Figure 15-2 on page 283. 282 32099I–01/2012 AT32UC3L016/32/64 Figure 15-2. Basic Mode WDT Timing Diagram, normal operation. t= t 0 T psel T im e o u t W rite o n e to C L R .W D T C L R W a tc h d o g re s e t If the WDT counter is not cleared within Tpsel a watchdog reset will be issued at the end of Tpsel, see Figure 15-3 on page 283. Figure 15-3. Basic Mode WDT Timing Diagram, no clear within Tpsel. t= t 0 T psel T im e o u t W rite o ne to C L R .W D T C L R W a tch d o g re se t 15.5.1.6 Watchdog Reset A watchdog reset will result in a reset and the code will start executing from the boot vector, please refer to the Power Manager chapter for details. If the Disable After Reset (DAR) bit in the CTRL Register is zero, the WDT counter will restart counting from zero when the watchdog reset is released. If the CTRL.DAR bit is one the WDT will be disabled after a watchdog reset. Only the CTRL.EN bit will be changed after the watchdog reset. However, if WDTAUTO fuse is configured to enable the WDT after a watchdog reset, and the CTRL.FCD bit is zero, writing a one to the CTRL.DAR bit will have no effect. 15.5.2 Window Mode The window mode can protect against tight loops of runaway code. This is obtained by adding a ban period to timeout period. During the ban period clearing the WDT counter is not allowed. If the WDT Mode (MODE) bit in the CTRL Register is one, the WDT is in window mode. Note that the CTRL.MODE bit can only be changed when the WDT is disabled (CTRL.EN=0). 283 32099I–01/2012 AT32UC3L016/32/64 The PSEL and Time Ban Prescale Select (TBAN) fields in the CTRL Register selects the WDT timeout period Ttimeout = Ttban + Tpsel = (2(TBAN+1) + 2(PSEL+1)) / fclk_cnt where Ttban sets the time period when clearing the WDT counter by writing to the CLR.WDTCLR bit is not allowed. Doing so will result in a watchdog reset, the device will receive a reset and the code will start executing form the boot vector, see Figure 15-5 on page 284. The WDT counter will be cleared. Writing a one to the CLR.WDTCLR bit within the Tpsel period will clear the WDT counter and the counter starts counting from zero (t=t0), entering Ttban, see Figure 15-4 on page 284. If the value in the CTRL Register is changed, the WDT counter will be cleared without a watchdog reset, regardless of if the value in the WDT counter and the TBAN value. If the WDT counter reaches Ttimeout, the counter will be cleared, the device will receive a reset and the code will start executing form the boot vector. Figure 15-4. Window Mode WDT Timing Diagram t= t 0 T tb a n T pse l T im e o u t W rite o n e to C L R .W D T C L R W a tch d o g re se t Figure 15-5. Window Mode WDT Timing Diagram, clearing within Ttban, resulting in watchdog reset. t= t 0 T tb a n T psel T im e o u t W rite o n e to C L R .W D T C L R W a tc h d o g re s e t 284 32099I–01/2012 AT32UC3L016/32/64 15.5.3 Disabling the WDT The WDT is disabled by writing a zero to the CTRL.EN bit. When disabling the WDT no other bits in the CTRL Register should be changed until the CTRL.EN bit reads back as zero. If the CTRL.CEN bit is written to zero, the CTRL.EN bit will never read back as zero if changing the value from one to zero. 15.5.4 Flash Calibration The WDT can be enabled at reset. This is controlled by the WDTAUTO fuse. The WDT will be set in basic mode, RCSYS is set as source for CLK_CNT, and PSEL will be set to a value giving Tpsel above 100 ms. Please refer to the Fuse Settings chapter for details about WDTAUTO and how to program the fuses. If the Flash Calibration Done (FCD) bit in the CTRL Register is zero at a watchdog reset the flash calibration will be redone, and the CTRL.FCD bit will be set when the calibration is done. If CTRL.FCD is one at a watchdog reset, the configuration of the WDT will not be changed during flash calibration. After any other reset the flash calibration will always be done, and the CTRL.FCD bit will be set when the calibration is done. 15.5.5 Special Considerations Care must be taken when selecting the PSEL/TBAN values so that the timeout period is greater than the startup time of the device. Otherwise a watchdog reset will reset the device before any code has been run. This can also be avoided by writing the CTRL.DAR bit to one when configuring the WDT. If the Store Final Value (SFV) bit in the CTRL Register is one, the CTRL Register is locked for further write accesses. All writes to the CTRL Register will be ignored. Once the CTRL Register is locked, it can only be unlocked by a reset (e.g. POR, OCD, and WDT). The CTRL.MODE bit can only be changed when the WDT is disabled (CTRL.EN=0). 285 32099I–01/2012 AT32UC3L016/32/64 15.6 User Interface Table 15-1. Note: WDT Register Memory Map Offset Register Register Name Access Reset 0x000 Control Register CTRL Read/Write 0x00010080 0x004 Clear Register CLR Write-only 0x00000000 0x008 Status Register SR Read-only 0x00000003 0x3FC Version Register VERSION Read-only -(1) 1. The reset value for this register is device specific. Please refer to the Module Configuration section at the end of this chapter. 286 32099I–01/2012 AT32UC3L016/32/64 15.6.1 Name: Control Register CTRL Access Type: Read/Write Offset: 0x000 Reset Value: 0x00010080 31 30 29 28 27 26 25 24 19 18 17 16 CSSEL CEN 9 8 KEY 23 22 21 - 20 TBAN 15 14 13 12 11 10 - - - 7 6 5 4 3 2 1 0 FCD - - - SFV MODE DAR EN PSEL • KEY • • • • • • • This field must be written twice, first with key value 0x55, then 0xAA, for a write operation to be effective. This field always reads as zero. TBAN: Time Ban Prescale Select Counter bit TBAN is used as watchdog “banned” time frame. In this time frame clearing the WDT timer is forbidden, otherwise a watchdog reset is generated and the WDT timer is cleared. CSSEL: Clock Source Select 0: Select the system RC oscillator (RCSYS) as clock source. 1: Select the 32KHz crystal oscillator (OSC32K) as clock source. CEN: Clock Enable 0: The WDT clock is disabled. 1: The WDT clock is enabled. PSEL: Time Out Prescale Select Counter bit PSEL is used as watchdog timeout period. FCD: Flash Calibration Done This bit is set after any reset. 0: The flash calibration will be redone after a watchdog reset. 1: The flash calibration will not be redone after a watchdog reset. SFV: WDT Control Register Store Final Value 0: WDT Control Register is not locked. 1: WDT Control Register is locked. Once locked, the Control Register can not be re-written, only a reset unlocks the SFV bit. MODE: WDT Mode 0: The WDT is in basic mode, only PSEL time is used. 1: The WDT is in window mode. Total timeout period is now TBAN+PSEL. Writing to this bit when the WDT is enabled has no effect. 287 32099I–01/2012 AT32UC3L016/32/64 • DAR: WDT Disable After Reset 0: After a watchdog reset, the WDT will still be enabled. 1: After a watchdog reset, the WDT will be disabled. • EN: WDT Enable 0: WDT is disabled. 1: WDT is enabled. After writing to this bit the read back value will not change until the WDT is enabled/disabled. This due to internal synchronization. 288 32099I–01/2012 AT32UC3L016/32/64 15.6.2 Name: Clear Register CLR Access Type: Write-only Offset: 0x004 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 KEY 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - WDTCLR When the Watchdog Timer is enabled, this Register must be periodically written within the window time frame or within the watchdog timeout period, to prevent a watchdog reset. • KEY This field must be written twice, first with key value 0x55, then 0xAA, for a write operation to be effective. • WDTCLR: Watchdog Clear Writing a zero to this bit has no effect. Writing a one to this bit clears the WDT counter. 289 32099I–01/2012 AT32UC3L016/32/64 15.6.3 Name: Status Register SR Access Type: Read-only Offset: 0x008 Reset Value: 0x00000003 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - CLEARED WINDOW • CLEARED: WDT Counter Cleared This bit is cleared when writing a one to the CLR.WDTCLR bit. This bit is set when clearing the WDT counter is done. • WINDOW: Within Window This bit is cleared when the WDT counter is inside the TBAN period. This bit is set when the WDT counter is inside the PSEL period. 290 32099I–01/2012 AT32UC3L016/32/64 15.6.4 Name: Version Register VERSION Access Type: Read-only Offset: 0x3FC Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 291 32099I–01/2012 AT32UC3L016/32/64 15.7 Module Configuration The specific configuration for each WDT instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 15-2. WDT Clocks Clock Name Description CLK_WDT Clock for the WDT bus interface Table 15-3. Register Reset Values Register Reset Value VERSION 0x00000402 292 32099I–01/2012 AT32UC3L016/32/64 16. External Interrupt Controller (EIC) Rev: 3.0.1.0 16.1 Features • • • • • • • 16.2 Dedicated interrupt request for each interrupt Individually maskable interrupts Interrupt on rising or falling edge Interrupt on high or low level Asynchronous interrupts for sleep modes without clock Filtering of interrupt lines Non-Maskable NMI interrupt Overview The External Interrupt Controller (EIC) allows pins to be configured as external interrupts. Each external interrupt has its own interrupt request and can be individually masked. Each external interrupt can generate an interrupt on rising or falling edge, or high or low level. Every interrupt input has a configurable filter to remove spikes from the interrupt source. Every interrupt pin can also be configured to be asynchronous in order to wake up the part from sleep modes where the CLK_SYNC clock has been disabled. A Non-Maskable Interrupt (NMI) is also supported. This has the same properties as the other external interrupts, but is connected to the NMI request of the CPU, enabling it to interrupt any other interrupt mode. The EIC can wake up the part from sleep modes without triggering an interrupt. In this mode, code execution starts from the instruction following the sleep instruction. 16.3 Block Diagram Figure 16-1. EIC Block Diagram EN D IS E n a b le LEVEL MODE EDGE ASYNC P o la rity co n tro l A syn ch ro n u s d e te cto r F IL T E R LEVEL MODE EDGE F ilte r E d g e/L e ve l D e te cto r E X T IN T n NMI CTRL CLK_SYNC IC R CTRL IE R ID R IN T n M a sk IS R IM R W a ke d e te ct IR Q n E IC _ W A K E 293 32099I–01/2012 AT32UC3L016/32/64 16.4 I/O Lines Description Table 16-1. 16.5 I/O Lines Description Pin Name Pin Description Type NMI Non-Maskable Interrupt Input EXTINTn External Interrupt Input Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 16.5.1 I/O Lines The external interrupt pins (EXTINTn and NMI) may be multiplexed with I/O Controller lines. The programmer must first program the I/O Controller to assign the desired EIC pins to their peripheral function. If I/O lines of the EIC are not used by the application, they can be used for other purposes by the I/O Controller. It is only required to enable the EIC inputs actually in use. If an application requires two external interrupts, then only two I/O lines will be assigned to EIC inputs. 16.5.2 Power Management All interrupts are available in all sleep modes as long as the EIC module is powered. However, in sleep modes where CLK_SYNC is stopped, the interrupt must be configured to asynchronous mode. 16.5.3 Clocks The clock for the EIC bus interface (CLK_EIC) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. The filter and synchronous edge/level detector runs on a clock which is stopped in any of the sleep modes where the system RC oscillator (RCSYS) is not running. This clock is referred to as CLK_SYNC. 16.5.4 Interrupts The external interrupt request lines are connected to the interrupt controller. Using the external interrupts requires the interrupt controller to be programmed first. Using the Non-Maskable Interrupt does not require the interrupt controller to be programmed. 16.5.5 Debug Operation When an external debugger forces the CPU into debug mode, the EIC continues normal operation. If the EIC is configured in a way that requires it to be periodically serviced by the CPU through interrupts or similar, improper operation or data loss may result during debugging. 16.6 16.6.1 Functional Description External Interrupts The external interrupts are not enabled by default, allowing the proper interrupt vectors to be set up by the CPU before the interrupts are enabled. 294 32099I–01/2012 AT32UC3L016/32/64 Each external interrupt INTn can be configured to produce an interrupt on rising or falling edge, or high or low level. External interrupts are configured by the MODE, EDGE, and LEVEL registers. Each interrupt has a bit INTn in each of these registers. Writing a zero to the INTn bit in the MODE register enables edge triggered interrupts, while writing a one to the bit enables level triggered interrupts. If INTn is configured as an edge triggered interrupt, writing a zero to the INTn bit in the EDGE register will cause the interrupt to be triggered on a falling edge on EXTINTn, while writing a one to the bit will cause the interrupt to be triggered on a rising edge on EXTINTn. If INTn is configured as a level triggered interrupt, writing a zero to the INTn bit in the LEVEL register will cause the interrupt to be triggered on a low level on EXTINTn, while writing a one to the bit will cause the interrupt to be triggered on a high level on EXTINTn. Each interrupt has a corresponding bit in each of the interrupt control and status registers. Writing a one to the INTn bit in the Interrupt Enable Register (IER) enables the external interrupt from pin EXTINTn to propagate from the EIC to the interrupt controller, while writing a one to INTn bit in the Interrupt Disable Register (IDR) disables this propagation. The Interrupt Mask Register (IMR) can be read to check which interrupts are enabled. When an interrupt triggers, the corresponding bit in the Interrupt Status Register (ISR) will be set. This bit remains set until a one is written to the corresponding bit in the Interrupt Clear Register (ICR) or the interrupt is disabled. Writing a one to the INTn bit in the Enable Register (EN) enables the external interrupt on pin EXTINTn, while writing a one to INTn bit in the Disable Register (DIS) disables the external interrupt. The Control Register (CTRL) can be read to check which interrupts are enabled. If a bit in the CTRL register is set, but the corresponding bit in IMR is not set, an interrupt will not propagate to the interrupt controller. However, the corresponding bit in ISR will be set, and EIC_WAKE will be set. Note that an external interrupt should not be enabled before it has been configured correctly. If the CTRL.INTn bit is zero, the corresponding bit in ISR will always be zero. Disabling an external interrupt by writing a one to the DIS.INTn bit will clear the corresponding bit in ISR. Please refer to the Module Configuration section for the number of external interrupts. 16.6.2 Synchronization and Filtering of External Interrupts In synchronous mode the pin value of the EXTINTn pin is synchronized to CLK_SYNC, so spikes shorter than one CLK_SYNC cycle are not guaranteed to produce an interrupt. The synchronization of the EXTINTn to CLK_SYNC will delay the propagation of the interrupt to the interrupt controller by two cycles of CLK_SYNC, see Figure 16-2 and Figure 16-3 for examples (FILTER off). It is also possible to apply a filter on EXTINTn by writing a one to the INTn bit in the FILTER register. This filter is a majority voter, if the condition for an interrupt is true for more than one of the latest three cycles of CLK_SYNC the interrupt will be set. This will additionally delay the propagation of the interrupt to the interrupt controller by one or two cycles of CLK_SYNC, see Figure 16-2 and Figure 16-3 for examples (FILTER on). 295 32099I–01/2012 AT32UC3L016/32/64 Figure 16-2. Timing Diagram, Synchronous Interrupts, High Level or Rising Edge CLK_SYNC EXTINTn/NMI ISR.INTn: FILTER off ISR.INTn: FILTER on Figure 16-3. Timing Diagram, Synchronous Interrupts, Low Level or Falling Edge CLK_SYNC EXTINTn/NMI ISR.INTn: FILTER off ISR.INTn: FILTER on 16.6.3 Non-Maskable Interrupt The NMI supports the same features as the external interrupts, and is accessed through the same registers. The description in Section 16.6.1 should be followed, accessing the NMI bit instead of the INTn bits. The NMI is non-maskable within the CPU in the sense that it can interrupt any other execution mode. Still, as for the other external interrupts, the actual NMI input can be enabled and disabled by accessing the registers in the EIC. 16.6.4 Asynchronous Interrupts Each external interrupt can be made asynchronous by writing a one to INTn in the ASYNC register. This will route the interrupt signal through the asynchronous path of the module. All edge interrupts will be interpreted as level interrupts and the filter is disabled. If an interrupt is configured as edge triggered interrupt in asynchronous mode, a zero in EDGE.INTn will be interpreted as low level, and a one in EDGE.INTn will be interpreted as high level. EIC_WAKE will be set immediately after the source triggers the interrupt, while the corresponding bit in ISR and the interrupt to the interrupt controller will be set on the next rising edge of CLK_SYNC. Please refer to Figure 16-4 on page 297 for details. 296 32099I–01/2012 AT32UC3L016/32/64 When CLK_SYNC is stopped only asynchronous interrupts remain active, and any short spike on this interrupt will wake up the device. EIC_WAKE will restart CLK_SYNC and ISR will be updated on the first rising edge of CLK_SYNC. Figure 16-4. Timing Diagram, Asynchronous Interrupts C LK _SYN C CLK_SYNC E X T IN T n /N M I 16.6.5 E X T IN T n /N M I IS R .IN T n : ris in g E D G E o r h ig h LEVEL IS R .IN T n : ris in g E D G E o r h ig h LEVEL E IC _ W A K E : ris in g E D G E o r h ig h LEVEL E IC _ W A K E : ris in g E D G E o r h ig h LEVEL Wakeup The external interrupts can be used to wake up the part from sleep modes. The wakeup can be interpreted in two ways. If the corresponding bit in IMR is one, then the execution starts at the interrupt handler for this interrupt. If the bit in IMR is zero, then the execution starts from the next instruction after the sleep instruction. 297 32099I–01/2012 AT32UC3L016/32/64 16.7 User Interface Table 16-2. Note: EIC Register Memory Map Offset Register Register Name Access Reset 0x000 Interrupt Enable Register IER Write-only 0x00000000 0x004 Interrupt Disable Register IDR Write-only 0x00000000 0x008 Interrupt Mask Register IMR Read-only 0x00000000 0x00C Interrupt Status Register ISR Read-only 0x00000000 0x010 Interrupt Clear Register ICR Write-only 0x00000000 0x014 Mode Register MODE Read/Write 0x00000000 0x018 Edge Register EDGE Read/Write 0x00000000 0x01C Level Register LEVEL Read/Write 0x00000000 0x020 Filter Register FILTER Read/Write 0x00000000 0x024 Test Register TEST Read/Write 0x00000000 0x028 Asynchronous Register ASYNC Read/Write 0x00000000 0x030 Enable Register EN Write-only 0x00000000 0x034 Disable Register DIS Write-only 0x00000000 0x038 Control Register CTRL Read-only 0x00000000 0x3FC Version Register VERSION Read-only - (1) 1. The reset value is device specific. Please refer to the Module Configuration section at the end of this chapter. 298 32099I–01/2012 AT32UC3L016/32/64 16.7.1 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x000 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n Writing a zero to this bit has no effect. Writing a one to this bit will set the corresponding bit in IMR. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt Writing a zero to this bit has no effect. Wrting a one to this bit will set the corresponding bit in IMR. 299 32099I–01/2012 AT32UC3L016/32/64 16.7.2 Name: Interrupt Disable Register IDR Access Type: Write-only Offset: 0x004 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n Writing a zero to this bit has no effect. Writing a one to this bit will clear the corresponding bit in IMR. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt Writing a zero to this bit has no effect. Writing a one to this bit will clear the corresponding bit in IMR. 300 32099I–01/2012 AT32UC3L016/32/64 16.7.3 Name: Interrupt Mask Register IMR Access Type: Read-only Offset: 0x008 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. This bit is cleared when the corresponding bit in IDR is written to one. This bit is set when the corresponding bit in IER is written to one. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt 0: The Non-Maskable Interrupt is disabled. 1: The Non-Maskable Interrupt is enabled. This bit is cleared when the corresponding bit in IDR is written to one. This bit is set when the corresponding bit in IER is written to one. 301 32099I–01/2012 AT32UC3L016/32/64 16.7.4 Name: Interrupt Status Register ISR Access Type: Read-only Offset: 0x00C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n 0: An interrupt event has not occurred. 1: An interrupt event has occurred. This bit is cleared by writing a one to the corresponding bit in ICR. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt 0: An interrupt event has not occurred. 1: An interrupt event has occurred. This bit is cleared by writing a one to the corresponding bit in ICR. 302 32099I–01/2012 AT32UC3L016/32/64 16.7.5 Name: Interrupt Clear Register ICR Access Type: Write-only Offset: 0x010 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n Writing a zero to this bit has no effect. Writing a one to this bit will clear the corresponding bit in ISR. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt Writing a zero to this bit has no effect. Writing a one to this bit will clear the corresponding bit in ISR. 303 32099I–01/2012 AT32UC3L016/32/64 16.7.6 Name: Mode Register MODE Access Type: Read/Write Offset: 0x014 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n 0: The external interrupt is edge triggered. 1: The external interrupt is level triggered. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt 0: The Non-Maskable Interrupt is edge triggered. 1: The Non-Maskable Interrupt is level triggered. 304 32099I–01/2012 AT32UC3L016/32/64 16.7.7 Name: Edge Register EDGE Access Type: Read/Write Offset: 0x018 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n 0: The external interrupt triggers on falling edge. 1: The external interrupt triggers on rising edge. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt 0: The Non-Maskable Interrupt triggers on falling edge. 1: The Non-Maskable Interrupt triggers on rising edge. 305 32099I–01/2012 AT32UC3L016/32/64 16.7.8 Name: Level Register LEVEL Access Type: Read/Write Offset: 0x01C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n 0: The external interrupt triggers on low level. 1: The external interrupt triggers on high level. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt 0: The Non-Maskable Interrupt triggers on low level. 1: The Non-Maskable Interrupt triggers on high level. 306 32099I–01/2012 AT32UC3L016/32/64 16.7.9 Filter Register Name: FILTER Access Type: Read/Write Offset: 0x020 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n 0: The external interrupt is not filtered. 1: The external interrupt is filtered. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt 0: The Non-Maskable Interrupt is not filtered. 1: The Non-Maskable Interrupt is filtered. 307 32099I–01/2012 AT32UC3L016/32/64 16.7.10 Test Register Name: TEST Access Type: Read/Write Offset: 0x024 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 TESTEN INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • TESTEN: Test Enable 0: This bit disables external interrupt test mode. 1: This bit enables external interrupt test mode. • INTn: External Interrupt n Writing a zero to this bit will set the input value to INTn to zero, if test mode is enabled. Writing a one to this bit will set the input value to INTn to one, if test mode is enabled. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt Writing a zero to this bit will set the input value to NMI to zero, if test mode is enabled. Writing a one to this bit will set the input value to NMI to one, if test mode is enabled. If TESTEN is 1, the value written to this bit will be the value to the interrupt detector and the value on the pad will be ignored. 308 32099I–01/2012 AT32UC3L016/32/64 16.7.11 Asynchronous Register Name: ASYNC Access Type: Read/Write Offset: 0x028 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n 0: The external interrupt is synchronized to CLK_SYNC. 1: The external interrupt is asynchronous. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt 0: The Non-Maskable Interrupt is synchronized to CLK_SYNC. 1: The Non-Maskable Interrupt is asynchronous. 309 32099I–01/2012 AT32UC3L016/32/64 16.7.12 Enable Register Name: EN Access Type: Write-only Offset: 0x030 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n Writing a zero to this bit has no effect. Writing a one to this bit will enable the corresponding external interrupt. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt Writing a zero to this bit has no effect. Writing a one to this bit will enable the Non-Maskable Interrupt. 310 32099I–01/2012 AT32UC3L016/32/64 16.7.13 Disable Register Name: DIS Access Type: Write-only Offset: 0x034 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n Writing a zero to this bit has no effect. Writing a one to this bit will disable the corresponding external interrupt. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt Writing a zero to this bit has no effect. Writing a one to this bit will disable the Non-Maskable Interrupt. 311 32099I–01/2012 AT32UC3L016/32/64 16.7.14 Control Register Name: CTRL Access Type: Read-only Offset: 0x038 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - INT30 INT29 INT28 INT27 INT26 INT25 INT24 23 22 21 20 19 18 17 16 INT23 INT22 INT21 INT20 INT19 INT18 INT17 INT16 15 14 13 12 11 10 9 8 INT15 INT14 INT13 INT12 INT11 INT10 INT9 INT8 7 6 5 4 3 2 1 0 INT7 INT6 INT5 INT4 INT3 INT2 INT1 NMI • INTn: External Interrupt n 0: The corresponding external interrupt is disabled. 1: The corresponding external interrupt is enabled. Please refer to the Module Configuration section for the number of external interrupts. • NMI: Non-Maskable Interrupt 0: The Non-Maskable Interrupt is disabled. 1: The Non-Maskable Interrupt is enabled. 312 32099I–01/2012 AT32UC3L016/32/64 16.7.15 Name: Version Register VERSION Access Type: Read-only Offset: 0x3FC Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - 7 6 5 4 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VERSION: Version number Version number of the module. No functionality associated. 313 32099I–01/2012 AT32UC3L016/32/64 16.8 Module Configuration The specific configuration for each EIC instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 16-3. Module Configuration Feature EIC Number of external interrupts, including NMI 6 Table 16-4. EIC Clocks Clock Name Description CLK_EIC Clock for the EIC bus interface Table 16-5. Register Reset Values Register Reset Value VERSION 0x00000301 314 32099I–01/2012 AT32UC3L016/32/64 17. Frequency Meter (FREQM) Rev: 3.0.1.1 17.1 Features • • • • 17.2 Accurately measures a clock frequency Selectable reference clock A selectable clock can be measured Ratio can be measured with 24-bit accuracy Overview The Frequency Meter (FREQM) can be used to accurately measure the frequency of a clock by comparing it to a known reference clock. 17.3 Block Diagram Figure 17-1. Frequency Meter Block Diagram CLKSEL START CLK_MSR Counter VALUE CLK_REF Timer Trigger REFSEL 17.4 ISR REFNUM, START Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 17.4.1 Power Management The device can enter a sleep mode while a measurement is ongoing. However, make sure that neither CLK_MSR nor CLK_REF is stopped in the actual sleep mode. FREQM interrupts can wake up the device from sleep modes when the measurement is done, but only from sleep modes where CLK_FREQM is running. Please refer to the Power Manager chapter for details. 315 32099I–01/2012 AT32UC3L016/32/64 17.4.2 Clocks The clock for the FREQM bus interface (CLK_FREQM) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the FREQM before disabling the clock, to avoid freezing the FREQM ia an undefined state. A set of clocks can be selected as reference (CLK_REF) and another set of clocks can be selected for measurement (CLK_MSR). Please refer to the CLKSEL and REFSEL tables in the Module Configuration section for details. 17.4.3 Debug Operation When an external debugger forces the CPU into debug mode, the FREQM continues normal operation. If the FREQM is configured in a way that requires it to be periodically serviced by the CPU through interrupts or similar, improper operation or data loss may result during debugging. 17.4.4 Interrupts The FREQM interrupt request line is connected to the internal source of the interrupt controller. Using the FREQM interrupt requires the interrupt controller to be programmed first. 17.5 Functional Description The FREQM accuratly measures the frequency of a clock by comparing the frequency to a known frequency: fCLK_MSR = (VALUE/REFNUM)*fCLK_REF 17.5.1 Reference Clock The Reference Clock Selection (REFSEL) field in the Mode Register (MODE) selects the clock source for CLK_REF. The reference clock is enabled by writing a one to the Reference Clock Enable (REFCEN) bit in the Mode Register. This clock should have a known frequency. CLK_REF needs to be disabled before switching to another clock. The RCLKBUSY bit in the Status Register (SR) indicates whether the clock is busy or not. This bit is set when the MODE.REFCEN bit is written. To change CLK_REF: • Write a zero to the MODE.REFCEN bit to disable the clock, without changing the other bits/fields in the Mode Register. • Wait until the SR.RCLKBUSY bit reads as zero. • Change the MODE.REFSEL field. • Write a one to the MODE.REFCEN bit to enable the clock, without changing the other bits/fields in the Mode Register. • Wait until the SR.RCLKBUSY bit reads as zero. To enable CLK_REF: • Write the correct value to the MODE.REFSEL field. • Write a one to the MODE.REFCEN to enable the clock, without changing the other bits/fields in the Mode Register. • Wait until the SR.RCLKBUSY bit reads as zero. To disable CLK_REF: 316 32099I–01/2012 AT32UC3L016/32/64 • Write a zero to the MODE.REFCEN to disable he clock, without changing the other bits/fields in the Mode register. • Wait until the SR.RCLKBUSY bit reads as zero. 17.5.1.1 17.5.2 Cautionary note Note that if clock selected as source for CLK_REF is stopped during a measurement, this will not be detected by the FREQM. The BUSY bit in the STATUS register will never be cleared, and the DONE interrupt will never be triggered. If the clock selected as soruce for CLK_REF is stopped, it will not be possible to change the source for the reference clock as long as the selected source is not running. Measurement In the Mode Register the Clock Source Selection (CLKSEL) field selects CLK_MSR and the Number of Reference Clock Cycles (REFNUM) field selects the duration of the measurement. The duration is given in number of CLK_REF periodes. Writing a one to the START bit in the Control Register (CTRL) starts the measurement. The BUSY bit in SR is cleared when the measurement is done. The result of the measurement can be read from the Value Register (VALUE). The frequency of the measured clock CLK_MSR is then: fCLK_MSR = (VALUE/REFNUM)*fCLK_REF 17.5.3 Interrupts The FREQM has two interrupt sources: • DONE: A frequency measurement is done • RCLKRDY: The reference clock is ready These will generate an interrupt request if the corresponding bit in the Interrupt Mask Register (IMR) is set. The interrupt sources are ORed together to form one interrupt request. The FREQM will generate an interrupt request if at least one of the bits in the Interrupt Mask Register (IMR) is set. Bits in IMR are set by writing a one to the corresponding bit in the Interrupt Enable Register (IER) and cleared by writing a one to this bit in the Interrupt Disable Register (IDR). The interrupt request remains active until the corresponding bit in the Interrupt Status Register (ISR) is cleared by writing a one to this bit in the Interrupt Clear Register (ICR). Because all the interrupt sources are ORed together, the interrupt request from the FREQM will remain active until all the bits in ISR are cleared. 317 32099I–01/2012 AT32UC3L016/32/64 17.6 User Interface Table 17-1. Note: FREQM Register Memory Map Offset Register Register Name Access Reset 0x000 Control Register CTRL Write-only 0x00000000 0x004 Mode Register MODE Read/Write 0x00000000 0x008 Status Register STATUS Read-only 0x00000000 0x00C Value Register VALUE Read-only 0x00000000 0x010 Interrupt Enable Register IER Write-only 0x00000000 0x014 Interrupt Disable Register IDR Write-only 0x00000000 0x018 Interrupt Mask Register IMR Read-only 0x00000000 0x01C Interrupt Status Register ISR Read-only 0x00000000 0x020 Interrupt Clear Register ICR Write-only 0x00000000 0x3FC Version Register VERSION Read-only -(1) 1. The reset value for this register is device specific. Please refer to the Module Configuration section at the end of this chapter. 318 32099I–01/2012 AT32UC3L016/32/64 17.6.1 Name: Control Register CTRL Access Type: Write-only Offset: 0x000 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - START • START Writing a zero to this bit has no effect. Writing a one to this bit will start a measurement. 319 32099I–01/2012 AT32UC3L016/32/64 17.6.2 Name: Mode Register MODE Access Type: Read/Write Offset: 0x004 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 REFCEN - - - - - - - 23 22 21 20 19 18 17 16 - - - 15 14 13 CLKSEL 12 11 10 9 8 2 1 0 REFNUM 7 6 5 4 3 - - - - - REFSEL • REFCEN: Reference Clock Enable 0: The reference clock is disabled 1: The reference clock is enabled • CLKSEL: Clock Source Selection Selects the source for CLK_MSR. See table in Module Configuration chapter for details. • REFNUM: Number of Reference Clock Cycles Selects the duration of a measurement, given in number of CLK_REF cycles. • REFSEL: Reference Clock Selection Selects the source for CLK_REF. See table in Module Configuration chapter for details. 320 32099I–01/2012 AT32UC3L016/32/64 17.6.3 Status Register Name: STATUS Access Type: Read-only Offset: 0x008 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - RCLKBUSY BUSY • RCLKBUSY: FREQM Reference Clock Status 0: The FREQM ref clk is ready, so a measurement can start. 1: The FREQM ref clk is not ready, so a measurement should not be started. • BUSY: FREQM Status 0: The Frequency Meter is idle. 1: Frequency measurement is on-going. 321 32099I–01/2012 AT32UC3L016/32/64 17.6.4 Value Register Name: VALUE Access Type: Read-only Offset: 0x00C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 11 10 9 8 3 2 1 0 VALUE[23:16] 15 14 13 12 VALUE[15:8] 7 6 5 4 VALUE[7:0] • VALUE: Result from measurement. 322 32099I–01/2012 AT32UC3L016/32/64 17.6.5 Interrupt Enable Register Name: IER Access Type: Write-only Offset: 0x010 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - RCLKRDY DONE Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 323 32099I–01/2012 AT32UC3L016/32/64 17.6.6 Interrupt Disable Register Name: IDR Access Type: Write-only Offset: 0x014 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - RCLKRDY DONE Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 324 32099I–01/2012 AT32UC3L016/32/64 17.6.7 Interrupt Mask Register Name: IMR Access Type: Read-only Offset: 0x018 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - RCLKRDY DONE 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. A bit in this register is cleared when the corresponding bit in IDR is written to one. A bit in this register is set when the corresponding bit in IER is written to one. 325 32099I–01/2012 AT32UC3L016/32/64 17.6.8 Interrupt Status Register Name: ISR Access Type: Read-only Offset: 0x01C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - RCLKRDY DONE 0: The corresponding interrupt is cleared. 1: The corresponding interrupt is pending. A bit in this register is set when the corresponding bit in STATUS has a one to zero transition. A bit in this register is cleared when the corresponding bit in ICR is written to one. 326 32099I–01/2012 AT32UC3L016/32/64 17.6.9 Interrupt Clear Register Name: ICR Access Type: Write-only Offset: 0x020 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - RCLKRDY DONE Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in ISR and the corresponding interrupt request. 327 32099I–01/2012 AT32UC3L016/32/64 17.6.10 Name: Version Register VERSION Access Type: Read-only Offset: 0x3FC Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 328 32099I–01/2012 AT32UC3L016/32/64 17.7 Module Configuration The specific configuration for each FREQM instance is listed in the following tables. The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 17-2. FREQM Clock Name Module Name FREQM Table 17-3. Clock Name Description CLK_FREQM Bus interface clock CLK_MSR Measured clock CLK_REF Reference clock Register Reset Values Register Reset Value VERSION 0x00000301 Table 17-4. Clock Sources for CLK_MSR CLKSEL Clock/Oscillator Description 0 CLK_CPU The clock the CPU runs on 1 CLK_HSB High Speed Bus clock 2 CLK_PBA Peripheral Bus A clock 3 CLK_PBB Peripheral Bus B clock 4 OSC0 Output clock from Oscillator 0 5 OSC32K Output clock from OSC32K 6 RCSYS Output clock from RCSYS Oscillator 7 DFLL0 Output clock from DFLL0 8 Reserved 9-14 GCLK0-5 Generic clock 0 through 5 15 RC120M AW clock Output clock from RC120M to AW 16 RC120M Output clock from RC120M to main clock mux 17 RC32K Output clock from RC32K 18-31 Reserved Table 17-5. Clock Sources for CLK_REF REFSEL Clock/Oscillator Description 0 RCSYS System RC oscillator clock 1 OSC32K Output clock form OSC32K 2-7 Reserved 329 32099I–01/2012 AT32UC3L016/32/64 330 32099I–01/2012 AT32UC3L016/32/64 18. General-Purpose Input/Output Controller (GPIO) Rev: 2.1.1.5 18.1 Features • • • • • • • 18.2 Configurable pin-change, rising-edge, or falling-edge interrupt Configurable peripheral event generator Glitch filter providing rejection of pulses shorter than one clock cycle Input visibility and output control Multiplexing of peripheral functions on I/O pins Programmable internal pull-up resistor Optional locking of configuration to avoid accidental reconfiguration Overview The General Purpose Input/Output Controller (GPIO) controls the I/O pins of the microcontroller. Each GPIO pin may be used as a general-purpose I/O or be assigned to a function of an embedded peripheral. The GPIO is configured using the Peripheral Bus (PB). Some registers can also be configured using the low latency CPU Local Bus. See Section 18.6.2.7 for details. 18.3 Block Diagram Figure 18-1. GPIO Block Diagram Configuration Interface Interrupt Controller GPIO Interrupt Request PIN General Purpose Input/Output - GPIO Power Manager CLK_GPIO PIN PIN PIN MCU I/O Pins PIN Embedded Peripheral Pin Control Signals 331 32099I–01/2012 AT32UC3L016/32/64 18.4 I/O Lines Description Pin Name Description Type GPIOn GPIO pin n Digital 18.5 Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 18.5.1 Power Management If the CPU enters a sleep mode that disables clocks used by the GPIO, the GPIO will stop functioning and resume operation after the system wakes up from sleep mode. If a peripheral function is configured for a GPIO pin, the peripheral will be able to control the GPIO pin even if the GPIO clock is stopped. 18.5.2 Clocks The GPIO is connected to a Peripheral Bus clock (CLK_GPIO). This clock is generated by the Power Manager. CLK_GPIO is enabled at reset, and can be disabled by writing to the Power Manager. CLK_GPIO must be enabled in order to access the configuration registers of the GPIO or to use the GPIO interrupts. After configuring the GPIO, the CLK_GPIO can be disabled by writing to the Power Manager if interrupts are not used. If the CPU Local Bus is used to access the configuration interface of the GPIO, the CLK_GPIO must be equal to the CPU clock to avoid data loss. 18.5.3 Interrupts The GPIO interrupt request lines are connected to the interrupt controller. Using the GPIO interrupts requires the interrupt controller to be programmed first. 18.5.4 Peripheral Events The GPIO peripheral events are connected via the Peripheral Event System. Refer to the Peripheral Event System chapter for details. 18.5.5 Debug Operation When an external debugger forces the CPU into debug mode, the GPIO continues normal operation. If the GPIO is configured in a way that requires it to be periodically serviced by the CPU through interrupts or similar, improper operation or data loss may result during debugging. 332 32099I–01/2012 AT32UC3L016/32/64 18.6 Functional Description The GPIO controls the I/O pins of the microcontroller. The control logic associated with each pin is shown in the figure below. Figure 18-2. Overview of the GPIO PUER* ODER 1 0 Periph. Func. A Output Pullup 0 Periph.Func. B Periph. Func. C 1 .... GPER PMRn Output Enable 0 0 1 OVR PIN 1 Input PVR IER 0 Edge Detector 1 1 Glitch Filter IFR IMR1 GFER 0 Interrupt Request IMR0 *) Register value is overrided if a peripheral function that support this function is enabled 333 32099I–01/2012 AT32UC3L016/32/64 18.6.1 Basic Operation 18.6.1.1 Module Configuration The GPIO user interface registers are organized into ports and each port controls 32 different GPIO pins. Most of the registers supports bit wise access operations such as set, clear and toggle in addition to the standard word access. For details regarding interface registers, refer to Section 18.7. 18.6.1.2 Available Features The GPIO features implemented are device dependent, and not all functions are implemented on all pins. The user must refer to the Module Configuration section and the GPIO Function Multiplexing section in the Package and Pinout chapter for the device specific settings used in the AT32UC3L016/32/64. Device specific settings includes: • Number of GPIO pins • Functions implemented on each pin • Peripheral function(s) multiplexed on each GPIO pin • Reset state of registers 18.6.1.3 Inputs The level on each GPIO pin can be read through the Pin Value Register (PVR). This register indicates the level of the GPIO pins regardless of the pins being driven by the GPIO or by an external component. Note that due to power saving measures, the PVR register will only be updated when the corresponding bit in GPER is one or if an interrupt is enabled for the pin, i.e. IER is one for the corresponding pin. 18.6.1.4 Output Control When the GPIO pin is assigned to a peripheral function, i.e. the corresponding bit in GPER is zero, the peripheral determines whether the pin is driven or not. When the GPIO pin is controlled by the GPIO, the value of Output Driver Enable Register (ODER) determines whether the pin is driven or not. When a bit in this register is one, the corresponding GPIO pin is driven by the GPIO. When the bit is zero, the GPIO does not drive the pin. The level driven on a GPIO pin can be determined by writing the value to the corresponding bit in the Output Value Register (OVR). 18.6.1.5 Peripheral Muxing The GPIO allows a single GPIO pin to be shared by multiple peripheral pins and the GPIO itself. Peripheral pins sharing the same GPIO pin are arranged into peripheral functions that can be selected one at a time. Peripheral functions are configured by writing the selected function value to the Peripheral Mux Registers (PMRn). To allow a peripheral pin access to the shared GPIO pin, GPIO control must be disabled for that pin, i.e. the corresponding bit in GPER must read zero. A peripheral function value is set by writing bit zero to PMR0 and bit one to the same index position in PMR1 and so on. In a system with 4 peripheral functions A,B,C, and D, peripheral function C for GPIO pin four is selected by writing a zero to bit four in PMR0 and a one to the same bit index in PMR1. Refer to the GPIO Function Multiplexing chapter for details regarding pin function configuration for each GPIO pin. 334 32099I–01/2012 AT32UC3L016/32/64 18.6.2 Advanced Operation 18.6.2.1 Peripheral I/O Pin Control When a GPIO pin is assigned to a peripheral function, i.e. the corresponding bit in GPER is zero, output and output enable is controlled by the selected peripheral pin. In addition the peripheral may control some or all of the other GPIO pin functions listed in Table 18-1, if the peripheral supports those features. All pin features not controlled by the selected peripheral is controlled by the GPIO. Refer to the Module Configuration section for details regarding implemented GPIO pin functions and to the Peripheral chapter for details regarding I/O pin function control. Table 18-1. 18.6.2.2 I/O Pin function Control Function name GPIO mode Peripheral mode Output OVR Peripheral Output enable ODER Peripheral Pull-up PUER Peripheral if supported, else GPIO Pull-up Resistor Control Pull-up can be configured for each GPIO pin. Pull-up allows the pin and any connected net to be pulled up to VDD if the net is not driven. Pull-up is useful for detecting if a pin is unconnected or if a mechanical button is pressed, for various communication protocols and to keep unconnected pins from floating. Pull-up can be enabled and disabled by writing a one and a zero respectively to the corresponding bit in the Pull-up Enable Register (PUER). 18.6.2.3 Output Pin Timings Figure 18-3 shows the timing of the GPIO pin when writing to the Output Value Register (OVR). The same timing applies when performing a ‘set’ or ‘clear’ access, i.e. writing to OVRS or OVRC. The timing of PVR is also shown. Figure 18-3. Output Pin Timings CLK_GPIO Write OVR to 1 Write OVR to 0 PB Access PB Access OVR / I/O Line PVR 18.6.2.4 Interrupts The GPIO can be configured to generate an interrupt when it detects a change on a GPIO pin. Interrupts on a pin are enabled by writing a one to the corresponding bit in the Interrupt Enable 335 32099I–01/2012 AT32UC3L016/32/64 Register (IER). The module can be configured to generate an interrupt whenever a pin changes value, or only on rising or falling edges. This is controlled by the Interrupt Mode Registers (IMRn). Interrupts on a pin can be enabled regardless of the GPIO pin being controlled by the GPIO or assigned to a peripheral function. An interrupt can be generated on each GPIO pin. These interrupt generators are further grouped into groups of eight and connected to the interrupt controller. An interrupt request from any of the GPIO pin generators in the group will result in an interrupt request from that group to the interrupt controller if the corresponding bit for the GPIO pin in the IER is set. By grouping interrupt generators into groups of eight, four different interrupt handlers can be installed for each GPIO port. The Interrupt Flag Register (IFR) can be read by software to determine which pin(s) caused the interrupt. The interrupt flag must be manually cleared by writing a zero to the corresponding bit in IFR. GPIO interrupts will only be generated when CLK_GPIO is enabled. 18.6.2.5 Input Glitch Filter Input glitch filters can be enabled on each GPIO pin. When the glitch filter is enabled, a glitch with duration of less than 1 CLK_GPIO cycle is automatically rejected, while a pulse with duration of 2 CLK_GPIO cycles or more is accepted. For pulse durations between 1 and 2 CLK_GPIO cycles, the pulse may or may not be taken into account, depending on the precise timing of its occurrence. Thus for a pulse to be guaranteed visible it must exceed 2 CLK_GPIO cycles, whereas for a glitch to be reliably filtered out, its duration must not exceed 1 CLK_GPIO cycle. The filter introduces 2 clock cycles latency. The glitch filters are controlled by the Glitch Filter Enable Register (GFER). When a bit in GFER is one, the glitch filter on the corresponding pin is enabled. The glitch filter affects only interrupt inputs. Inputs to peripherals or the value read through PVR are not affected by the glitch filters. 18.6.2.6 Interrupt Timings Figure 18-4 shows the timing for rising edge (or pin-change) interrupts when the glitch filter is disabled. For the pulse to be registered, it must be sampled at the rising edge of the clock. In this example, this is not the case for the first pulse. The second pulse is sampled on a rising edge and will trigger an interrupt request. Figure 18-4. Interrupt Timing with Glitch Filter Disabled CLK_GPIO Pin Level IFR Figure 18-5 shows the timing for rising edge (or pin-change) interrupts when the glitch filter is enabled. For the pulse to be registered, it must be sampled on two subsequent rising edges. In the example, the first pulse is rejected while the second pulse is accepted and causes an interrupt request. 336 32099I–01/2012 AT32UC3L016/32/64 Figure 18-5. Interrupt Timing with Glitch Filter Enabled CLK_GPIO Pin Level IFR 18.6.2.7 CPU Local Bus The CPU Local Bus can be used for application where low latency read and write access to the Output Value Register (OVR) and Output Drive Enable Register (ODER) is required. The CPU Local Bus allows the CPU to configure the mentioned GPIO registers directly, bypassing the shared Peripheral Bus (PB). To avoid data loss when using the CPU Local Bus, the CLK_GPIO must run at the same frequency as the CLK_CPU. See Section 18.5.2 for details. The CPU Local Bus is mapped to a different base address than the GPIO but the OVER and ODER offsets are the same. See the CPU Local Bus Mapping section in the Memories chapter for details. 18.6.2.8 Peripheral Events Peripheral events allow direct peripheral to peripheral communication of specified events. See the Peripheral Event System chapter for more information. The GPIO can be programmed to output peripheral events whenever an interrupt condition is detected. The peripheral events configuration depends on the interrupt configuration. An event will be generated on the same condition as the interrupt (pin change, rising edge, or falling edge). The interrupt configuration is controlled by the IMR register. Peripheral event on a pin is enabled by writing a one to the corresponding bit in the Event Enable Register (EVER). The Peripheral Event trigger mode is shared with the interrupt trigger and is configured by writing to the IMR0 and IMR1 registers. Interrupt does not need to be enabled on a pin when peripheral events are enabled. Peripheral Events are also affected by the Input Glitch Filter settings. See Section 18.6.2.5 for more information. A peripheral event can be generated on each GPIO pin. Each port can then have up to 32 peripheral event generators. Groups of eight peripheral event generators in each port are ORed together to form a peripheral event line, so that each port has four peripheral event lines connected to the Peripheral Event System. 337 32099I–01/2012 AT32UC3L016/32/64 18.7 User Interface The GPIO controller manages all the GPIO pins on the 32-bit AVR microcontroller. The pins are managed as 32-bit ports that are configurable through a Peripheral Bus (PB) interface. Each port has a set of configuration registers. The overall memory map of the GPIO is shown below. The number of pins and hence the number of ports is product specific. Figure 18-6. Port Configuration Registers 0x0000 Port 0 Configuration Registers 0x0200 Port 1 Configuration Registers 0x0400 …. Port 2 Configuration Registers n*0x200 Port n Configuration Registers In the peripheral muxing table in the Package and Pinout chapter each GPIO pin has a unique number. Note that the PA, PB, PC, and PX ports do not necessarily directly correspond to the GPIO ports. To find the corresponding port and pin the following formulas can be used: GPIO port = floor((GPIO number) / 32), example: floor((36)/32) = 1 GPIO pin = GPIO number % 32, example: 36 % 32 = 4 Table 18-2 shows the configuration registers for one port. Addresses shown are relative to the port address offset. The specific address of a configuration register is found by adding the register offset and the port offset to the GPIO start address. One bit in each of the configuration registers corresponds to a GPIO pin. 18.7.1 Access Types Most configuration register can be accessed in four different ways. The first address location can be used to write the register directly. This address can also be used to read the register value. The following addresses facilitate three different types of write access to the register. Performing a “set” access, all bits written to one will be set. Bits written to zero will be unchanged by the operation. Performing a “clear” access, all bits written to one will be cleared. Bits written to zero will be unchanged by the operation. Finally, a toggle access will toggle the value of all bits writ- 338 32099I–01/2012 AT32UC3L016/32/64 ten to one. Again all bits written to zero remain unchanged. Note that for some registers (e.g. IFR), not all access methods are permitted. Note that for ports with less than 32 bits, the corresponding control registers will have unused bits. This is also the case for features that are not implemented for a specific pin. Writing to an unused bit will have no effect. Reading unused bits will always return 0. 18.7.2 Configuration Protection In order to protect the configuration of individual GPIO pins from software failure, configuration bits for individual GPIO pins may be locked by writing a one to the corresponding bit in the LOCK register. While this bit is one, any write to the same bit position in any lockable GPIO register using the Peripheral Bus (PB) will not have an effect. The CPU Local Bus is not checked and thus allowed to write to all bits in a CPU Local Bus mapped register no mather the LOCK value. The registers required to clear bits in the LOCK register are protected by the access protection mechanism described in Section 18.7.3, ensuring the LOCK mechanism itself is robust against software failure. 18.7.3 Access Protection In order to protect critical registers from software failure, some registers are protected by a key protection mechanism. These registers can only be changed by first writing the UNLOCK register, then the protected register. Protected registers are indicated in Table 18-2. The UNLOCK register contains a key field which must always be written to 0xAA, and an OFFSET field corresponding to the offset of the register to be modified. The next write operation resets the UNLOCK register, so if the register is to be modified again, the UNLOCK register must be written again. Attempting to write to a protected register without first writing the UNLOCK register results in the write operation being discarded, and the Access Error bit in the Access Status Register (ASR.AE) will be set. Table 18-2. GPIO Register Memory Map Offset Register Function Register Name Access Reset Config. Protection Access Protection 0x000 GPIO Enable Register Read/Write GPER Read/Write -(1) Y N 0x004 GPIO Enable Register Set GPERS Write-only Y N 0x008 GPIO Enable Register Clear GPERC Write-only Y N 0x00C GPIO Enable Register Toggle GPERT Write-only Y N Y N - (1) 0x010 Peripheral Mux Register 0 Read/Write PMR0 Read/Write 0x014 Peripheral Mux Register 0 Set PMR0S Write-only Y N 0x018 Peripheral Mux Register 0 Clear PMR0C Write-only Y N 0x01C Peripheral Mux Register 0 Toggle PMR0T Write-only Y N Y N (1) 0x020 Peripheral Mux Register 1 Read/Write PMR1 Read/Write - 0x024 Peripheral Mux Register 1 Set PMR1S Write-only Y N 0x028 Peripheral Mux Register 1 Clear PMR1C Write-only Y N 339 32099I–01/2012 AT32UC3L016/32/64 Table 18-2. GPIO Register Memory Map Offset Register Function Register Name Access 0x02C Peripheral Mux Register 1 Toggle PMR1T Write-only Reset - (1) Config. Protection Access Protection Y N Y N 0x030 Peripheral Mux Register 2 Read/Write PMR2 Read/Write 0x034 Peripheral Mux Register 2 Set PMR2S Write-only Y N 0x038 Peripheral Mux Register 2 Clear PMR2C Write-only Y N 0x03C Peripheral Mux Register 2 Toggle PMR2T Write-only Y N Y N (1) 0x040 Output Driver Enable Register Read/Write ODER Read/Write - 0x044 Output Driver Enable Register Set ODERS Write-only Y N 0x048 Output Driver Enable Register Clear ODERC Write-only Y N 0x04C Output Driver Enable Register Toggle ODERT Write-only Y N N N - (1) 0x050 Output Value Register Read/Write OVR Read/Write 0x054 Output Value Register Set OVRS Write-only N N 0x058 Output Value Register Clear OVRC Write-only N N 0x05c Output Value Register Toggle OVRT Write-only N N N N Depe nding on pin states 0x060 Pin Value Register Read PVR Read-only 0x064 Pin Value Register - - - N N 0x068 Pin Value Register - - - N N 0x06c Pin Value Register - - - N N 0x070 Pull-up Enable Register Read/Write PUER Read/Write Y N 0x074 Pull-up Enable Register Set PUERS Write-only Y N 0x078 Pull-up Enable Register Clear PUERC Write-only Y N 0x07C Pull-up Enable Register Toggle PUERT Write-only Y N 0x090 Interrupt Enable Register Read/Write IER Read/Write N N 0x094 Interrupt Enable Register Set IERS Write-only N N 0x098 Interrupt Enable Register Clear IERC Write-only N N 0x09C Interrupt Enable Register Toggle IERT Write-only N N 0x0A0 Interrupt Mode Register 0 Read/Write IMR0 Read/Write N N 0x0A4 Interrupt Mode Register 0 Set IMR0S Write-only N N 0x0A8 Interrupt Mode Register 0 Clear IMR0C Write-only N N 0x0AC Interrupt Mode Register 0 Toggle IMR0T Write-only N N 0x0B0 Interrupt Mode Register 1 Read/Write IMR1 Read/Write N N 0x0B4 Interrupt Mode Register 1 Set IMR1S Write-only N N 0x0B8 Interrupt Mode Register 1 Clear IMR1C Write-only N N 0x0BC Interrupt Mode Register 1 Toggle IMR1T Write-only N N -(1) -(1) -(1) -(1) 340 32099I–01/2012 AT32UC3L016/32/64 Table 18-2. GPIO Register Memory Map Offset Register Function Register Name Access Reset Config. Protection Access Protection 0x0C0 Glitch Filter Enable Register Read/Write GFER Read/Write -(1) N N 0x0C4 Glitch Filter Enable Register Set GFERS Write-only N N 0x0C8 Glitch Filter Enable Register Clear GFERC Write-only N N 0x0CC Glitch Filter Enable Register Toggle GFERT Write-only N N N N - (1) 0x0D0 Interrupt Flag Register Read IFR Read-only 0x0D4 Interrupt Flag Register - - - N N 0x0D8 Interrupt Flag Register Clear IFRC Write-only N N 0x0DC Interrupt Flag Register - - - N N N N - (1) 0x180 Event Enable Register Read EVER Read/Write 0x184 Event Enable Register Set EVERS Write-only N N 0x188 Event Enable Register Clear EVERC Write-only N N 0x18C Event Enable Register Toggle EVERT Write-only N N N Y 0x1A0 Lock Register Read/Write LOCK Read/Write 0x1A4 Lock Register Set LOCKS Write-only N N 0x1A8 Lock Register Clear LOCKC Write-only N Y 0x1AC Lock Register Toggle LOCKT Write-only N Y 0x1E0 Unlock Register Read/Write UNLOCK Write-only N N 0x1E4 Access Status Register Read/Write ASR Read/Write 0x1F8 Parameter Register Read PARAMETER Read-only -(1) N N Read-only (1) N N 0x1FC Note: Version Register Read VERSION - (1) N - 1. The reset values for these registers are device specific. Please refer to the Module Configuration section at the end of this chapter. 341 32099I–01/2012 AT32UC3L016/32/64 18.7.4 Name: GPIO Enable Register GPER Access: Read/Write, Set, Clear, Toggle Offset: 0x000, 0x004, 0x008, 0x00C Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-P31: GPIO Enable 0: A peripheral function controls the corresponding pin. 1: The GPIO controls the corresponding pin. 342 32099I–01/2012 AT32UC3L016/32/64 18.7.5 Name: Peripheral Mux Register 0 PMR0 Access: Read/Write, Set, Clear, Toggle Offset: 0x010, 0x014, 0x018, 0x01C Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Peripheral Multiplexer Select bit 0 343 32099I–01/2012 AT32UC3L016/32/64 18.7.6 Name: Peripheral Mux Register 1 PMR1 Access: Read/Write, Set, Clear, Toggle Offset: 0x020, 0x024, 0x028, 0x02C Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Peripheral Multiplexer Select bit 1 344 32099I–01/2012 AT32UC3L016/32/64 18.7.7 Name: Peripheral Mux Register 2 PMR2 Access: Read/Write, Set, Clear, Toggle Offset: 0x030, 0x034, 0x038, 0x03C Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Peripheral Multiplexer Select bit 2 {PMR2, PMR1, PMR0} 000 001 010 011 100 101 110 111 Selected Peripheral Function A B C D E F G H 345 32099I–01/2012 AT32UC3L016/32/64 18.7.8 Name: Output Driver Enable Register ODER Access: Read/Write, Set, Clear, Toggle Offset: 0x040, 0x044, 0x048, 0x04C Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Output Driver Enable 0: The output driver is disabled for the corresponding pin. 1: The output driver is enabled for the corresponding pin. 346 32099I–01/2012 AT32UC3L016/32/64 18.7.9 Name: Output Value Register OVR Access: Read/Write, Set, Clear, Toggle Offset: 0x050, 0x054, 0x058, 0x05C Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Output Value 0: The value to be driven on the GPIO pin is 0. 1: The value to be driven on the GPIO pin is 1. 347 32099I–01/2012 AT32UC3L016/32/64 18.7.10 Name: Pin Value Register PVR Access: Read-only Offset: 0x060, 0x064, 0x068, 0x06C Reset Value: Depending on pin states 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Pin Value 0: The GPIO pin is at level zero. 1: The GPIO pin is at level one. Note that the level of a pin can only be read when the corresponding pin in GPER is one or interrupt is enabled for the pin. 348 32099I–01/2012 AT32UC3L016/32/64 18.7.11 Name: Pull-up Enable Register PUER Access: Read/Write, Set, Clear, Toggle Offset: 0x070, 0x074, 0x078, 0x07C Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Pull-up Enable Writing a zero to a bit in this register will disable pull-up on the corresponding pin. Writing a one to a bit in this register will enable pull-up on the corresponding pin. 349 32099I–01/2012 AT32UC3L016/32/64 18.7.12 Name: Interrupt Enable Register IER Access: Read/Write, Set, Clear, Toggle Offset: 0x090, 0x094, 0x098, 0x09C Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Interrupt Enable 0: Interrupt is disabled for the corresponding pin. 1; Interrupt is enabled for the corresponding pin. 350 32099I–01/2012 AT32UC3L016/32/64 18.7.13 Name: Interrupt Mode Register 0 IMR0 Access: Read/Write, Set, Clear, Toggle Offset: 0x0A0, 0x0A4, 0x0A8, 0x0AC Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Interrupt Mode Bit 0 351 32099I–01/2012 AT32UC3L016/32/64 18.7.14 Name: Interrupt Mode Register 1 IMR1 Access: Read/Write, Set, Clear, Toggle Offset: 0x0B0, 0x0B4, 0x0B8, 0x0BC Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Interrupt Mode Bit 1 {IMR1, IMR0} 00 01 10 11 Interrupt Mode Pin Change Rising Edge Falling Edge Reserved 352 32099I–01/2012 AT32UC3L016/32/64 18.7.15 Name: Glitch Filter Enable Register GFER Access: Read/Write, Set, Clear, Toggle Offset: 0x0C0, 0x0C4, 0x0C8, 0x0CC Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Glitch Filter Enable 0: Glitch filter is disabled for the corresponding pin. 1: Glitch filter is enabled for the corresponding pin. NOTE! The value of this register should only be changed when the corresponding bit in IER is zero. Updating GFER while interrupt on the corresponding pin is enabled can cause an unintentional interrupt to be triggered. 353 32099I–01/2012 AT32UC3L016/32/64 18.7.16 Name: Interrupt Flag Register IFR Access: Read, Clear Offset: 0x0D0, 0x0D8 Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Interrupt Flag 0: No interrupt condition has been detected on the corresponding pin. 1: An interrupt condition has been detected on the corresponding pin. The number of interrupt request lines depends on the number of GPIO pins on the MCU. Refer to the product specific data for details. Note also that a bit in the Interrupt Flag register is only valid if the corresponding bit in IER is one. 354 32099I–01/2012 AT32UC3L016/32/64 18.7.17 Name: Event Enable Register EVER Access: Read/Write, Set, Clear, Toggle Offset: 0x180, 0x184, 0x188, 0x18C Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Event Enable 0: Peripheral Event is disabled for the corresponding pin. 1: Peripheral Event is enabled for the corresponding pin. 355 32099I–01/2012 AT32UC3L016/32/64 18.7.18 Name: Lock Register LOCK Access: Read/Write, Set, Clear, Toggle Offset: 0x1A0, 0x1A4, 0x1A8, 0x1AC Reset Value: - 31 30 29 28 27 26 25 24 P31 P30 P29 P28 P27 P26 P25 P24 23 22 21 20 19 18 17 16 P23 P22 P21 P20 P19 P18 P17 P16 15 14 13 12 11 10 9 8 P15 P14 P13 P12 P11 P10 P9 P8 7 6 5 4 3 2 1 0 P7 P6 P5 P4 P3 P2 P1 P0 • P0-31: Lock State 0: Pin is unlocked. The corresponding bit can be changed in any GPIO register for this port. 1: Pin is locked. The corresponding bit can not be changed in any GPIO register for this port. The value of LOCK determines which bits are locked in the lockable registers. The LOCK, LOCKC, and LOCKT registers are protected, which means they can only be written immediately after a write to the UNLOCK register with the proper KEY and OFFSET. LOCKS is not protected, and can be written at any time. 356 32099I–01/2012 AT32UC3L016/32/64 18.7.19 Name: Unlock Register UNLOCK Access: Write-only Offset: 0x1E0 Reset Value: - 31 30 29 28 27 26 25 24 KEY 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 - - - - - - 7 6 5 4 3 2 8 OFFSET 1 0 OFFSET • OFFSET: Register Offset This field must be written with the offset value of the LOCK, LOCKC or LOCKT register to unlock. This offset must also include the port offset for the register to unlock. LOCKS can not be locked so no unlock is required before writing to this register. • KEY: Unlocking Key This bitfield must be written to 0xAA for a write to this register to have an effect. This register always reads as zero. 357 32099I–01/2012 AT32UC3L016/32/64 18.7.20 Name: Access Status Register ASR Access: Read/Write Offset: 0x1E4 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - AE • AE: Access Error This bit is set when a write to a locked register occurs. This bit can be written to 0 by software. 358 32099I–01/2012 AT32UC3L016/32/64 18.7.21 Name: Parameter Register PARAMETER Access Type: Read-only Offset: 0x1F8 Reset Value: - 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 PARAMETER 23 22 21 20 PARAMETER 15 14 13 12 PARAMETER 7 6 5 4 PARAMETER • PARAMETER: 0: The corresponding pin is not implemented in this GPIO port. 1: The corresponding pin is implemented in this GPIO port. There is one PARAMETER register per GPIO port. Each bit in the Parameter Register indicates whether the corresponding GPER bit is implemented. 359 32099I–01/2012 AT32UC3L016/32/64 18.7.22 Name: Version Register VERSION Access Type: Read-only Offset: 0x1FC Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 1 0 VARIANT 11 10 VERSION[11:8] 3 2 VERSION[7:0] • VARIANT: Variant Number Reserved. No functionality associated. • VERSION: Version Number Version number of the module. No functionality associated. 360 32099I–01/2012 AT32UC3L016/32/64 18.8 Module Configuration The specific configuration for each GPIO instance is listed in the following tables. The module bus clocks listed here are connected to the system bus clocks. Refer to the Power Manager chapter for details. Table 18-3. GPIO Configuration Feature GPIO Number of GPIO ports 2 Number of peripheral functions 8 Table 18-4. Implemented Pin Functions Pin Function Implemented Notes Pull-up On all pins Controlled by PUER or peripheral Table 18-5. GPIO Clocks Module Name Clock Name Description GPIO CLK_GPIO Clock for the GPIO bus interface The reset values for all GPIO registers are zero, with the following exceptions: Table 18-6. Register Reset Values Port Register Reset Value 0 GPER 0x004DFF5F 0 PMR0 0x00320020 0 PMR1 0x00020080 0 PMR2 0x00100800 0 PUER 0x00000001 0 GFER 0x007FFFFF 0 PARAMETER 0x007FFFFF 0 VERSION 0x00000211 1 GPER 0x00001FCF 1 PMR0 0x00000030 1 PMR1 0x00000030 1 GFER 0x00001FFF 1 PARAMETER 0x00001FFF 1 VERSION 0x00000211 361 32099I–01/2012 AT32UC3L016/32/64 19. Universal Synchronous Asynchronous Receiver Transmitter (USART) Rev: 4.4.0.6 19.1 Features • Configurable baud rate generator • 5- to 9-bit full-duplex, synchronous and asynchronous, serial communication • • • • 19.2 – 1, 1.5, or 2 stop bits in asynchronous mode, and 1 or 2 in synchronous mode – Parity generation and error detection – Framing- and overrun error detection – MSB- or LSB-first – Optional break generation and detection – Receiver frequency over-sampling by 8 or 16 times – Optional RTS-CTS hardware handshaking – Receiver Time-out and transmitter Timeguard – Optional Multidrop mode with address generation and detection SPI Mode – Master or slave – Configurable serial clock phase and polarity – CLK SPI serial clock frequency up to a quarter of the CLK_USART internal clock frequency LIN Mode – Compliant with LIN 1.3 and LIN 2.0 specifications – Master or slave – Processing of Frames with up to 256 data bytes – Configurable response data length, optionally defined automatically by the Identifier – Self synchronization in slave node configuration – Automatic processing and verification of the “Break Field” and “Sync Field” – The “Break Field” is detected even if it is partially superimposed with a data byte – Optional, automatic identifier parity management – Optional, automatic checksum management – Supports both “Classic” and “Enhanced” checksum types – Full LIN error checking and reporting – Frame Slot Mode: the master allocates slots to scheduled frames automatically. – Wakeup signal generation Test Modes – Automatic echo, remote- and local loopback Supports two Peripheral DMA Controller channels – Buffer transfers without processor intervention Overview The Universal Synchronous Asynchronous Receiver Transceiver (USART) provides a full duplex, universal, synchronous/asynchronous serial link. Data frame format is widely configurable, including basic length, parity, and stop bit settings, maximizing standards support. The receiver implements parity-, framing-, and overrun error detection, and can handle un-fixed frame lengths with the time-out feature. The USART supports several operating modes, providing an interface to, LIN, and SPI buses and infrared transceivers. Communication with slow and remote devices is eased by the timeguard. Duplex multidrop communication is supported by address and data differentiation through the parity bit. The hardware handshaking feature enables an out-of-band flow control, automatically managing RTS and CTS pins. The Peripheral DMA Controller connection enables memory transactions, and the USART supports chained 362 32099I–01/2012 AT32UC3L016/32/64 buffer management without processor intervention. Automatic echo, remote-, and local loopback -test modes are also supported. 19.3 Block Diagram Figure 19-1. USART Block Diagram Peripheral DMA Controller Channel Channel USART I/O Controller RXD Receiver RTS Interrupt Controller USART Interrupt TXD Transmitter CTS CLK_USART Power Manager DIV BaudRate Generator CLK_USART/DIV CLK User Interface Peripheral bus Table 19-1. SPI Operating Mode PIN USART SPI Slave SPI Master RXD RXD MOSI MISO TXD TXD MISO MOSI RTS RTS – CS CTS CTS CS – 363 32099I–01/2012 AT32UC3L016/32/64 19.4 I/O Lines Description Table 19-2. I/O Lines Description Name Description Type CLK Serial Clock I/O TXD Transmit Serial Data or Master Out Slave In (MOSI) in SPI master mode or Master In Slave Out (MISO) in SPI slave mode Output RXD Receive Serial Data or Master In Slave Out (MISO) in SPI master mode or Master Out Slave In (MOSI) in SPI slave mode Input CTS Clear to Send or Slave Select (NSS) in SPI slave mode Input Low RTS Request to Send or Slave Select (NSS) in SPI master mode Output Low 19.5 19.5.1 Active Level Product Dependencies I/O Lines The USART pins may be multiplexed with the I/O Controller lines. The user must first configure the I/O Controller to assign these pins to their peripheral functions. Unused I/O lines may be used for other purposes. To prevent the TXD line from falling when the USART is disabled, the use of an internal pull up is required. If the hardware handshaking feature or modem mode is used, the internal pull up on TXD must also be enabled. 19.5.2 Clocks The clock for the USART bus interface (CLK_USART) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the USART before disabling the clock, to avoid freezing the USART in an undefined state. 19.5.3 Interrupts The USART interrupt request line is connected to the interrupt controller. Using the USART interrupt requires the interrupt controller to be programmed first. 364 32099I–01/2012 AT32UC3L016/32/64 19.6 Functional Description 19.6.1 Baud Rate Generator The baud rate generator provides the bit period clock named the Baud Rate Clock to both receiver and transmitter. It is based on a 16-bit divider, which is specified in the Clock Divider field in the Baud Rate Generator Register (BRGR.CD). A non-zero value enables the generator, and if CD is one, the divider is bypassed and inactive. The Clock Selection field in the Mode Register (MR.USCLKS) selects clock source between: • CLK_USART (internal clock) • CLK_USART/DIV (a divided CLK_USART, refer to Module Configuration section) • CLK (external clock, available on the CLK pin) If the external CLK clock is selected, the duration of the low and high levels of the signal provided on the CLK pin must be at least 4.5 times longer than those provided by CLK_USART. Figure 19-2. Baud Rate Generator USCLKS CLK_USART CLK_USART/DIV CLK Reserved CD CD 0 1 2 CLK 16-bit Counter FIDI >1 3 1 0 SYNC OVER 0 Sampling Divider 0 0 BaudRate Clock 1 1 SYNC USCLKS= 3 19.6.1.1 Sampling Clock Baud Rate in Asynchronous Mode If the USART is configured to operate in an asynchronous mode, the selected clock is divided by the CD value before it is provided to the receiver as a sampling clock. Depending on the Oversampling Mode bit (MR.OVER) value, the clock is then divided by either 8 (OVER=1), or 16 (OVER=0). The baud rate is calculated with the following formula: SelectedClock BaudRate = -----------------------------------------------( 8 ( 2 – OVER )CD ) This gives a maximum baud rate of CLK_USART divided by 8, assuming that CLK_USART is the fastest clock possible, and that OVER is one. 19.6.1.2 Baud Rate Calculation Example Table 19-3 shows calculations based on the CD field to obtain 38400 baud from different source clock frequencies. This table also shows the actual resulting baud rate and error. 365 32099I–01/2012 AT32UC3L016/32/64 Table 19-3. Baud Rate Example (OVER=0) Source Clock (Hz) Expected Baud Rate (bit/s) Calculation Result CD Actual Baud Rate (bit/s) Error 3 686 400 38 400 6.00 6 38 400.00 0.00% 4 915 200 38 400 8.00 8 38 400.00 0.00% 5 000 000 38 400 8.14 8 39 062.50 1.70% 7 372 800 38 400 12.00 12 38 400.00 0.00% 8 000 000 38 400 13.02 13 38 461.54 0.16% 12 000 000 38 400 19.53 20 37 500.00 2.40% 12 288 000 38 400 20.00 20 38 400.00 0.00% 14 318 180 38 400 23.30 23 38 908.10 1.31% 14 745 600 38 400 24.00 24 38 400.00 0.00% 18 432 000 38 400 30.00 30 38 400.00 0.00% 24 000 000 38 400 39.06 39 38 461.54 0.16% 24 576 000 38 400 40.00 40 38 400.00 0.00% 25 000 000 38 400 40.69 40 38 109.76 0.76% 32 000 000 38 400 52.08 52 38 461.54 0.16% 32 768 000 38 400 53.33 53 38 641.51 0.63% 33 000 000 38 400 53.71 54 38 194.44 0.54% 40 000 000 38 400 65.10 65 38 461.54 0.16% 50 000 000 38 400 81.38 81 38 580.25 0.47% 60 000 000 38 400 97.66 98 38 265.31 0.35% The baud rate is calculated with the following formula (OVER=0): BaudRate = ( CLKUSART ) ⁄ ( CD × 16 ) The baud rate error is calculated with the following formula. It is not recommended to work with an error higher than 5%. ExpectedBaudRate Error = 1 – ⎛ ---------------------------------------------------⎞ ⎝ ActualBaudRate ⎠ 19.6.1.3 Fractional Baud Rate in Asynchronous Mode The baud rate generator has a limitation: the source frequency is always a multiple of the baud rate. An approach to this problem is to integrate a high resolution fractional N clock generator, outputting fractional multiples of the reference source clock. This fractional part is selected with the Fractional Part field (BRGR.FP), and is activated by giving it a non-zero value. The resolution is one eighth of CD. The resulting baud rate is calculated using the following formula: SelectedClock BaudRate = -------------------------------------------------------------------⎛ 8 ( 2 – OVER ) ⎛ CD + FP -------⎞⎠ ⎞⎠ ⎝ ⎝ 8 The modified architecture is presented below: 366 32099I–01/2012 AT32UC3L016/32/64 Figure 19-3. Fractional Baud Rate Generator FP USCLKS CLK_USART CLK_USART/DIV CLK Reserved 0 1 2 3 CD Modulus Control FP CD 16-bit Counter glitch-free logic FIDI >1 1 0 CLK 0 OVER Sampling Divider 0 SYNC 0 BaudRate Clock 1 1 SYNC USCLKS = 3 19.6.1.4 Sampling Clock Baud Rate in Synchronous and SPI Mode If the USART is configured to operate in synchronous mode, the selected clock is divided by the BRGR.CD field. This does not apply when CLK is selected. SelectedClock BaudRate = -------------------------------------CD When CLK is selected the external frequency must be at least 4.5 times lower than the system clock, and when either CLK or CLK_USART/DIV are selected, CD must be even to ensure a 50/50 duty cycle. If CLK_USART is selected, the generator ensures this regardless of value. 19.6.2 Receiver and Transmitter Control After a reset, the transceiver is disabled. The receiver/transmitter is enabled by writing a one to either the Receiver Enable, or Transmitter Enable bit in the Control Register (CR.RXEN, or CR.TXEN). They may be enabled together and can be configured both before and after they have been enabled. The user can reset the USART receiver/transmitter at any time by writing a one to either the Reset Receiver (CR.RSTRX), or Reset Transmitter (CR.RSTTX) bit. This software reset clears status bits and resets internal state machines, immediately halting any communication. The user interface configuration registers will retain their values. The user can disable the receiver/transmitter by writing a one to either the Receiver Disable, or Transmitter Disable bit (CR.RXDIS, or CR.TXDIS). If the receiver is disabled during a character reception, the USART will wait for the current character to be received before disabling. If the transmitter is disabled during transmission, the USART will wait until both the current character and the character stored in the Transmitter Holding Register (THR) are transmitted before disabling. If a timeguard has been implemented it will remain functional during the transaction. 367 32099I–01/2012 AT32UC3L016/32/64 19.6.3 Synchronous and Asynchronous Modes 19.6.3.1 Transmitter Operations The transmitter performs equally in both synchronous and asynchronous operating modes (MR.SYNC). One start bit, up to 9 data bits, an optional parity bit, and up to two stop bits are successively shifted out on the TXD pin at each falling edge of the serial clock. The number of data bits is selected by the Character Length field (MR.CHRL) and the MR.MODE9 bit. Nine bits are selected by writing a one to MODE9, overriding any value in CHRL. The parity bit configuration is selected in the MR.PAR field. The Most Significant Bit First bit (MR.MSBF) selects which data bit to send first. The number of stop bits is selected by the MR.NBSTOP field. The 1.5 stop bit configuration is only supported in asynchronous mode. Figure 19-4. Character Transmit Example: 8-bit, Parity Enabled One Stop Baud Rate Clock TXD Start Bit D0 D1 D2 D3 D4 D5 D6 D7 Parity Bit Stop Bit The characters are sent by writing to the Character to be Transmitted field (THR.TXCHR). The transmitter reports status with the Transmitter Ready (TXRDY) and Transmitter Empty (TXEMPTY) bits in the Channel Status Register (CSR). TXRDY is set when THR is empty. TXEMPTY is set when both THR and the transmit shift register are empty (transmission complete). Both TXRDY and TXEMPTY are cleared when the transmitter is disabled. Writing a character to THR while TXRDY is zero has no effect and the written character will be lost. Figure 19-5. Transmitter Status Baud Rate Clock TXD Start D0 Bit D1 D2 D3 D4 D5 D6 D7 Parity Stop Start D0 Bit Bit Bit D1 D2 D3 D4 D5 D6 D7 Parity Stop Bit Bit Write THR TXRDY TXEMPTY 19.6.3.2 Asynchronous Receiver If the USART is configured in an asynchronous operating mode (MR.SYNC = 0), the receiver will oversample the RXD input line by either 8 or 16 times the baud rate clock, as selected by the Oversampling Mode bit (MR.OVER). If the line is zero for half a bit period (four or eight consecutive samples, respectively), a start bit will be assumed, and the following 8th or 16th sample will determine the logical value on the line, in effect resulting in bit values being determined at the middle of the bit period. 368 32099I–01/2012 AT32UC3L016/32/64 The number of data bits, endianess, parity mode, and stop bits are selected by the same bits and fields as for the transmitter (MR.CHRL, MODE9, MSBF, PAR, and NBSTOP). The synchronization mechanism will only consider one stop bit, regardless of the used protocol, and when the first stop bit has been sampled, the receiver will automatically begin looking for a new start bit, enabling resynchronization even if there is a protocol miss-match. Figure 19-6 and Figure 19-7 illustrate start bit detection and character reception in asynchronous mode. Figure 19-6. Asynchronous Start Bit Detection Baud Rate Clock Sampling Clock (x16) RXD Sampling 1 2 3 4 5 6 7 8 1 2 3 4 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 D0 Sampling Start Detection RXD Sampling 1 2 3 4 5 6 7 0 1 Start Rejection Figure 19-7. Asynchronous Character Reception Example: 8-bit, Parity Enabled Baud Rate Clock RXD Start Detection 16 16 16 16 16 16 16 16 16 16 samples samples samples samples samples samples samples samples samples samples D0 19.6.3.3 D1 D2 D3 D4 D5 D6 D7 Parity Bit Stop Bit Synchronous Receiver In synchronous mode (SYNC=1), the receiver samples the RXD signal on each rising edge of the Baud Rate Clock. If a low level is detected, it is considered as a start bit. Configuration bits and fields are the same as in asynchronous mode. 369 32099I–01/2012 AT32UC3L016/32/64 Figure 19-8. Synchronous Mode Character Reception Example: 8-bit, Parity Enabled 1 Stop Baud Rate Clock RXD Sampling Start D0 D1 D2 D3 D4 D5 D6 Stop Bit D7 Parity Bit 19.6.3.4 Receiver Operations When a character reception is completed, it is transferred to the Received Character field in the Receive Holding Register (RHR.RXCHR), and the Receiver Ready bit in the Channel Status Register (CSR.RXRDY) is set. If RXRDY is already set, RHR will be overwritten and the Overrun Error bit (CSR.OVRE) is set. Reading RHR will clear RXRDY, and writing a one to the Reset Status bit in the Control Register (CR.RSTSTA) will clear OVRE. Figure 19-9. Receiver Status Baud Rate Clock RXD Start D0 Bit D1 D2 D3 D4 D5 D6 D7 Parity Stop Start D0 Bit Bit Bit D1 D2 D3 D4 D5 D6 D7 Parity Stop Bit Bit RSTSTA = 1 Write CR Read RHR RXRDY OVRE 19.6.3.5 Parity The USART supports five parity modes selected by MR.PAR. The PAR field also enables the Multidrop mode, see ”Multidrop Mode” on page 371. If even parity is selected, the parity bit will be a zero if there is an even number of ones in the data character, and if there is an odd number it will be a one. For odd parity the reverse applies. If space or mark parity is chosen, the parity bit will always be a zero or one, respectively. See Table 19-4. Table 19-4. Parity Bit Examples Parity Mode Alphanum Character Hex Bin Odd Even Mark Space None A 0x41 0100 0001 1 0 1 0 - V 0x56 0101 0110 1 0 1 0 - R 0x52 0101 0010 0 1 1 0 - 370 32099I–01/2012 AT32UC3L016/32/64 The receiver will report parity errors in CSR.PARE, unless parity is disabled. Writing a one to CR.RSTSTA will clear PARE. See Figure 19-10 Figure 19-10. Parity Error Baud Rate Clock RXD Start D0 Bit D1 D2 D3 D4 D5 D6 D7 Bad Stop Parity Bit Bit RSTSTA = 1 Write CR PARE RXRDY 19.6.3.6 Multidrop Mode If PAR is either 0x6 or 0x7, the USART runs in Multidrop mode. This mode differentiates data and address characters. Data has the parity bit zero and addresses have a one. By writing a one to the Send Address bit (CR.SENDA) the user will cause the next character written to THR to be transmitted as an address. Receiving a character with a one as parity bit will set PARE. 19.6.3.7 Transmitter Timeguard The timeguard feature enables the USART to interface slow devices by inserting an idle state on the TXD line in between two characters. This idle state corresponds to a long stop bit, whose duration is selected by the Timeguard Value field in the Transmitter Timeguard Register (TTGR.TG). The transmitter will hold the TXD line high for TG bit periods, in addition to the number of stop bits. As illustrated in Figure 19-11, the behavior of TXRDY and TXEMPTY is modified when TG has a non-zero value. If a pending character has been written to THR, the TXRDY bit will not be set until this characters start bit has been sent. TXEMPTY will remain low until the timeguard transmission has completed. Figure 19-11. Timeguard Operation TG = 4 TG = 4 Baud Rate Clock TXD Start D0 Bit D1 D2 D3 D4 D5 D6 D7 Parity Stop Bit Bit Start D0 Bit D1 D2 D3 D4 D5 D6 D7 Parity Stop Bit Bit Write THR TXRDY TXEMPTY 371 32099I–01/2012 AT32UC3L016/32/64 Table 19-5. 19.6.3.8 Maximum Baud Rate Dependent Timeguard Durations Baud Rate (bit/sec) Bit time (µs) Timeguard (ms) 1 200 833 212.50 9 600 104 26.56 14400 69.4 17.71 19200 52.1 13.28 28800 34.7 8.85 33400 29.9 7.63 56000 17.9 4.55 57600 17.4 4.43 115200 8.7 2.21 Receiver Time-out The Time-out Value field in the Receiver Time-out Register (RTOR.TO) enables handling of variable-length frames by detection of selectable idle durations on the RXD line. The value written to TO is loaded to a decremental counter, and unless it is zero, a time-out will occur when the amount of inactive bit periods match the initial counter value. If a time-out has not occurred, the counter will reload and restart every time a new character arrives. A time-out sets the TIMEOUT bit in CSR. Clearing TIMEOUT can be done in two ways: • Writing a one to the Start Time-out bit (CR.STTTO). This also aborts count down until the next character has been received. • Writing a one to the Reload and Start Time-out bit (CR.RETTO). This also reloads the counter and restarts count down immediately. Figure 19-12. Receiver Time-out Block Diagram TO Baud Rate Clock 1 D Q Clock 16-bit Time-out Counter 16-bit Value = STTTO Character Received Load Clear TIMEOUT 0 RETTO Table 19-6. Maximum Time-out Period Baud Rate (bit/sec) Bit Time (µs) Time-out (ms) 600 1 667 109 225 1 200 833 54 613 2 400 417 27 306 4 800 208 13 653 372 32099I–01/2012 AT32UC3L016/32/64 19.6.3.9 Baud Rate (bit/sec) Bit Time (µs) Time-out (ms) 9 600 104 6 827 14400 69 4 551 19200 52 3 413 28800 35 2 276 33400 30 1 962 56000 18 1 170 57600 17 1 138 200000 5 328 Framing Error The receiver is capable of detecting framing errors. A framing error has occurred if a stop bit reads as zero. This can occur if the transmitter and receiver are not synchronized. A framing error is reported by CSR.FRAME as soon as the error is detected, at the middle of the stop bit. Figure 19-13. Framing Error Status Baud Rate Clock RXD Start D0 Bit D1 D2 D3 D4 D5 D6 D7 Parity Stop Bit Bit RSTSTA = 1 Write CR FRAME RXRDY 19.6.3.10 Transmit Break When TXRDY is set, the user can request the transmitter to generate a break condition on the TXD line by writing a one to The Start Break bit (CR.STTBRK). The break is treated as a normal 0x00 character transmission, clearing TXRDY and TXEMPTY, but with zeroes for preambles, start, parity, stop, and time guard bits. Writing a one to the Stop Break bit (CR.STBRK) will stop the generation of new break characters, and send ones for TG duration or at least 12 bit periods, ensuring that the receiver detects end of break, before resuming normal operation. Figure 19-14 illustrates STTBRK and STPBRK effect on the TXD line. Writing to STTBRK and STPBRK simultaneously can lead to unpredictable results. Writes to THR before a pending break has started will be ignored. 373 32099I–01/2012 AT32UC3L016/32/64 Figure 19-14. Break Transmission Baud Rate Clock TXD Start D0 Bit D1 D2 D3 D4 D5 D6 D7 Parity Stop Bit Bit Break Transmission STTBRK = 1 End of Break STPBRK = 1 Write CR TXRDY TXEMPTY 19.6.3.11 Receive Break A break condition is assumed when incoming data, parity, and stop bits are zero. This corresponds to a framing error, but FRAME will remain zero while the Break Received/End Of Break bit (CSR.RXBRK) is set. Writing a one to CR.RSTSTA will clear RXBRK. An end of break will also set RXBRK, and is assumed when TX is high for at least 2/16 of a bit period in asynchronous mode, or when a high level is sampled in synchronous mode. 19.6.3.12 Hardware Handshaking The USART features an out-of-band hardware handshaking flow control mechanism, implementable by connecting the RTS and CTS pins with the remote device, as shown in Figure 1915. Figure 19-15. Connection with a Remote Device for Hardware Handshaking USART Remote Device TXD RXD RXD TXD CTS RTS RTS CTS Writing 0x2 to the MR.MODE field configures the USART to operate in this mode. The receiver will drive its RTS pin high when disabled or when the Reception Buffer Full bit (CSR.RXBUFF) is set by the Buffer Full signal from the Peripheral DMA controller. If the receivers RTS pin is high, the transmitters CTS pin will also be high and only the active character transactions will be completed. Allocating a new buffer to the DMA controller by clearing RXBUFF, will drive the RTS pin low, allowing the transmitter to resume transmission. Detected level changes on the CTS pin can trigger interrupts, and are reported by the CTS Input Change bit in the Channel Status Register (CSR.CTSIC). Figure 19-16 illustrates receiver functionality, and Figure 19-17 illustrates transmitter functionality. 374 32099I–01/2012 AT32UC3L016/32/64 Figure 19-16. Receiver Behavior when Operating with Hardware Handshaking RXD RXEN = 1 RXDIS = 1 Write CR RTS RXBUFF Figure 19-17. Transmitter Behavior when Operating with Hardware Handshaking CTS TXD Figure 19-18. 19.6.4 SPI Mode The USART features a Serial Peripheral Interface (SPI) link compliant mode, supporting synchronous, full-duplex communication, in both master and slave mode. Writing 0xE (master) or 0xF (slave) to MR.MODE will enable this mode. A SPI in master mode controls the data flow to and from the other SPI devices, who are in slave mode. It is possible to let devices take turns being masters (aka multi-master protocol), and one master may shift data simultaneously into several slaves, but only one slave may respond at a time. A slave is selected when its slave select (NSS) signal has been raised by the master. The USART can only generate one NSS signal, and it is possible to use standard I/O lines to address more than one slave. 19.6.4.1 Modes of Operation The SPI system consists of two data lines and two control lines: • Master Out Slave In (MOSI): This line supplies the data shifted from master to slave. In master mode this is connected to TXD, and in slave mode to RXD. • Master In Slave Out (MISO): This line supplies the data shifted from slave to master. In master mode this is connected to RXD, and in slave mode to TXD. • Serial Clock (CLK): This is controlled by the master. One period per bit transmission. In both modes this is connected to CLK. • Slave Select (NSS): This control line allows the master to select or deselect a slave. In master mode this is connected to RTS, and in slave mode to CTS. Changing SPI mode after initial configuration has to be followed by a transceiver software reset in order to avoid unpredictable behavior. 19.6.4.2 Baud Rate The baud rate generator operates as described in ”Baud Rate in Synchronous and SPI Mode” on page 367, with the following requirements: In SPI Master Mode: 375 32099I–01/2012 AT32UC3L016/32/64 • The Clock Selection field (MR.USCLKS) must not equal 0x3 (external clock, CLK). • The Clock Output Select bit (MR.CLKO) must be one. • The BRGR.CD field must be at least 0x4. • If USCLKS is one (internal divided clock, CLK_USART/DIV), the value in CD has to be even, ensuring a 50:50 duty cycle. CD can be odd if USCLKS is zero (internal clock, CLK_USART). In SPI Slave Mode: • CLK frequency must be at least four times lower than the system clock. 19.6.4.3 Data Transfer • Up to nine data bits are successively shifted out on the TXD pin at each edge. There are no start, parity, or stop bits, and MSB is always sent first. The SPI Clock Polarity (MR.CPOL), and SPI Clock Phase (MR.CPHA) bits configure CLK by selecting the edges upon which bits are shifted and sampled, resulting in four non-interoperable protocol modes see Table 19-7. A master/slave pair must use the same configuration, and the master must be reconfigured if it is to communicate with slaves using different configurations. See Figures 19-19 and 19-20. Table 19-7. SPI Bus Protocol Modes SPI Bus Protocol Mode CPOL CPHA 0 0 1 1 0 0 2 1 1 3 1 0 Figure 19-19. SPI Transfer Format (CPHA=1, 8 bits per transfer) CLK cycle (for reference) 1 2 4 3 5 7 6 8 CLK (CPOL= 0) CLK (CPOL= 1) MOSI SPI Master ->TXD SPI Slave ->RXD MISO SPI Master ->RXD SPI Slave ->TXD MSB MSB 6 5 4 3 2 1 LSB 6 5 4 3 2 1 LSB NSS SPI Master ->RTS SPI Slave ->CTS 376 32099I–01/2012 AT32UC3L016/32/64 Figure 19-20. SPI Transfer Format (CPHA=0, 8 bits per transfer) CLK cycle (for reference) 1 2 3 4 5 6 7 8 CLK (CPOL= 0) CLK (CPOL= 1) MOSI SPI Master -> TXD SPI Slave -> RXD MSB 6 5 4 3 2 1 LSB MISO SPI Master -> RXD SPI Slave -> TXD MSB 6 5 4 3 2 1 LSB NSS SPI Master -> RTS SPI Slave -> CTS 19.6.4.4 Receiver and Transmitter Control See ”Transmitter Operations” on page 368, and ”Receiver Operations” on page 370. 19.6.4.5 Character Transmission and Reception In SPI master mode, the slave select line (NSS) is asserted low one bit period before the start of transmission, and released high one bit period after every character transmission. A delay for at least three bit periods is always inserted in between characters. In order to address slave devices supporting the Chip Select Active After Transfer (CSAAT) mode, NSS can be forced low by writing a one to the Force SPI Chip Select bit (CR.RTSEN/FCS). Releasing NSS when FCS is one, is only possible by writing a one to the Release SPI Chip Select bit (CR.RTSDIS/RCS). In SPI slave mode, a low level on NSS for at least one bit period will allow the slave to initiate a transmission or reception. The Underrun Error bit (CSR.UNRE) is set if a character must be sent while THR is empty, and TXD will be high during character transmission, as if 0xFF was being sent. If a new character is written to THR it will be sent correctly during the next transmission slot. Writing a one to CR.RSTSTA will clear UNRE. To ensure correct behavior of the receiver in SPI slave mode, the master device sending the frame must ensure a minimum delay of one bit period in between each character transmission. 19.6.4.6 19.6.5 Receiver Time-out Receiver Time-out’s are not possible in SPI mode as the baud rate clock is only active during data transfers. LIN Mode The USART features a LIN (Local Interconnect Network) 1.3 and 2.0 compliant mode, embedding full error checking and reporting, automatic frame processing with up to 256 data bytes, 377 32099I–01/2012 AT32UC3L016/32/64 customizable response data lengths, and requires minimal CPU resources. Writing 0xA (master) or 0xB (slave) to MR.MODE enables this mode. 19.6.5.1 Modes of operation Changing LIN mode after initial configuration has to be followed by a transceiver software reset in order to avoid unpredictable behavior. 19.6.5.2 Receiver and Transmitter Control See Section “19.6.2” on page 367. 19.6.5.3 Baud Rate Configuration The LIN nodes baud rate is configured in the Baud Rate Generator Register (BRGR), See Section “19.6.1.1” on page 365. 19.6.5.4 Character Transmission and Reception See ”Transmitter Operations” on page 368, and ”Receiver Operations” on page 370. 19.6.5.5 Header Transmission (Master Node Configuration) All LIN frames start with a header sent by the master. As soon as the identifier has been written to the Identifier Character field in the LIN Identifier Register (LINIR.IDCHR), TXRDY is cleared and the header is sent. The header consists of a Break, Sync, and Identifier field. TXRDY is set when the identifier has been transferred into the transmitters shift register. The Break field consists of 13 dominant bits, the break, and one recessive bit, the break delimiter. The Sync field is the character 0x55. The Identifier field contains the Identifier as written to IDCHR. The identifier parity bits can be generated automatically (see Section 19.6.5.8). Figure 19-21. Header Transmission Baud Rate Clock TXD Break Field 13 dominant bits (at 0) Write LINIR LINIR Break Delimiter 1 recessive bit (at 1) Start 1 Bit 0 1 0 1 0 Synch Byte = 0x55 1 0 Stop Start Stop ID0 ID1 ID2 ID3 ID4 ID5 ID6 ID7 Bit Bit Bit ID TXRDY 19.6.5.6 Header Reception (Slave Node Configuration) The USART stays idle until it detects a break field, consisting of at least 11 consecutive dominant bits (zeroes) on the bus. A received break will set the Lin Break bit (CSR.LINBK). The Sync field is used to synchronize the baud rate (see Section 19.6.5.7). IDCHR is updated and the LIN Identifier bit (CSR.LINID) is set when the Identifier has been received. The Identifier parity bits can be automatically checked (see Section 19.6.5.8). Writing a one to RSTSTA will clear LINBK and LINID. 378 32099I–01/2012 AT32UC3L016/32/64 Figure 19-22. Header Reception Baud Rate Clock RXD Break Field 13 dominant bits (at 0) LINID Break Delimiter 1 recessive bit (at 1) Start 1 Bit 0 1 0 1 0 1 0 Synch Byte = 0x55 Stop Stop Start ID0 ID1 ID2 ID3 ID4 ID5 ID6 ID7 Bit Bit Bit US_LINIR Write US_CR With RSTSTA=1 19.6.5.7 Slave Node Synchronization Synchronization is only done by the slave. If the Sync field is not 0x55, an Inconsistent Sync Field error (CSR.LINISFE) is generated. The time between falling edges is measured by a 19-bit counter, driven by the sampling clock (see Section 19.6.1). Figure 19-23. Sync Field Synch Field 8 Tbit 2 Tbit 2 Tbit 2 Tbit 2 Tbit Start bit Stop bit The counter starts when the Sync field start bit is detected, and continues for eight bit periods. The 16 most significant bits (counter value divided by 8) becomes the new clock divider (BRGR.CD), and the three least significant bits (the remainder) becomes the new fractional part (BRGR.FP). Figure 19-24. Slave Node Synchronization Baud Rate Clock RXD Break Field 13 dominant bits (at 0) Break Delimiter 1 recessive bit (at 1) Start 1 Bit 0 1 0 1 0 Synch Byte = 0x55 1 0 Stop Start Stop ID0 ID1 ID2 ID3 ID4 ID5 ID6 ID7 Bit Bit Bit LINIDRX Reset Synchro Counter 000_0011_0001_0110_1101 BRGR Clcok Divider (CD) Initial CD 0000_0110_0010_1101 BRGR Fractional Part (FP) Initial FP 101 The synchronization accuracy depends on: • The theoretical slave node clock frequency; nominal clock frequency (FNom) • The baud rate 379 32099I–01/2012 AT32UC3L016/32/64 • The oversampling mode (OVER=0 => 16x, or OVER=1 => 8x) The following formula is used to calculate synchronization deviation, where FSLAVE is the real slave node clock frequency, and FTOL_UNSYNC is the difference between FNom and FSLAVE According to the LIN specification, FTOL_UNSYNCH may not exceed ±15%, and the bit rates between two nodes must be within ±2% of each other, resulting in a maximal BaudRate_deviation of ±1%. [ α × 8 × ( 2 – OVER ) + β ] × BaudRate BaudRate_deviation = ⎛ 100 × ---------------------------------------------------------------------------------------------------⎞ % ⎝ ⎠ 8 × F SLAVE ⎛ ⎞ ⎜ [ α × 8 × ( 2 – OVER ) + β ] × BaudRate⎟ BaudRate_deviation = ⎜ 100 × ---------------------------------------------------------------------------------------------------⎟ % F TOL_UNSYNC ⎜ ⎟ 8 × ⎛ ------------------------------------⎞ xF Nom ⎝ ⎠ ⎝ ⎠ 100 – 0,5 ≤ α ≤ +0,5 -1 < β < +1 Minimum nominal clock frequency with a fractional part: ⎛ ⎞ ⎜ [ 0,5 × 8 × ( 2 – OVER ) + 1 ] × BaudRate⎟ F Nom ( min ) = ⎜ 100 × -------------------------------------------------------------------------------------------------------⎟ Hz – 15 ⎜ ⎟ 8 × ⎛ ---------- + 1⎞ × 1% ⎝ ⎠ ⎝ 100 ⎠ Examples: • Baud rate = 20 kbit/s, OVER=0 (Oversampling 16x) => FNom(min) = 2.64 MHz • Baud rate = 20 kbit/s, OVER=1 (Oversampling 8x) => FNom(min) = 1.47 MHz • Baud rate = 1 kbit/s, OVER=0 (Oversampling 16x) => FNom(min) = 132 kHz • Baud rate = 1 kbit/s, OVER=1 (Oversampling 8x) => FNom(min) = 74 kHz If the fractional part is not used, the synchronization accuracy is much lower. The 16 most significant bits, added with the first least significant bit, becomes the new clock divider (CD). The equation of the baud rate deviation is the same as above, but the constants are: – 4 ≤ α ≤ +4 -1 < β < +1 Minimum nominal clock frequency without a fractional part: ⎛ ⎞ ⎜ [ 4 × 8 × ( 2 – OVER ) + 1 ] × Baudrate⎟ F Nom ( min ) = ⎜ 100 × -----------------------------------------------------------------------------------------------⎟ Hz – 15 ⎜ ⎟ 8 × ⎛ ---------- + 1⎞ × 1% ⎝ ⎠ ⎝ 100 ⎠ Examples: • Baud rate = 20 kbit/s, OVER=0 (Oversampling 16x) => FNom(min) = 19.12 MHz • Baud rate = 20 kbit/s, OVER=1 (Oversampling 8x) => FNom(min) = 9.71 MHz • Baud rate = 1 kbit/s, OVER=0 (Oversampling 16x) => FNom(min) = 956 kHz • Baud rate = 1 kbit/s, OVER=1 (Oversampling 8x) => FNom(min) = 485 kHz 19.6.5.8 Identifier Parity An identifier field consists of two sub-fields; the identifier and its parity. Bits 0 to 5 are assigned to the identifier, while bits 6 and 7 are assigned to parity. Automatic parity management is disabled by writing a one to the Parity Disable bit in the LIN Mode register (LINMR.PARDIS). 380 32099I–01/2012 AT32UC3L016/32/64 • PARDIS=0: During header transmission, the parity bits are computed and in the shift register they replace bits six and seven from IDCHR. During header reception, the parity bits are checked and can generate a LIN Identifier Parity Error (see Section 19.6.6). Bits six and seven in IDCHR read as zero when receiving. • PARDIS=1: During header transmission, all the bits in IDCHR are sent on the bus. During header reception, all the bits in IDCHR are updated with the received Identifier. 19.6.5.9 Node Action After an identifier transaction, a LIN response mode has to be selected. This is done in the Node Action field (LINMR.NACT). Below are some response modes exemplified in a small LIN cluster: • Response, from master to slave1: Master: NACT=PUBLISH Slave1: NACT=SUBSCRIBE Slave2: NACT=IGNORE • Response, from slave1 to master: Master: NACT=SUBSCRIBE Slave1: NACT=PUBLISH Slave2: NACT=IGNORE • Response, from slave1 to slave2: Master: NACT=IGNORE Slave1: NACT=PUBLISH Slave2: NACT=SUBSCRIBE 19.6.5.10 LIN Response Data Length The response data length is the number of data fields (bytes), excluding the checksum. Figure 19-25. Response Data Length User configuration: 1 - 256 data fields (DLC+1) Identifier configuration: 2/4/8 data fields Sync Break Sync Field Identifier Field Data Field Data Field Data Field Data Field Checksum Field The response data length can be configured, either by the user, or automatically by bits 4 and 5 in the Identifier (IDCHR), in accordance to LIN 1.1. The user selects mode by writing to the Data Length Mode bit (LINMR.DML): • DLM=0: the response data length is configured by the user by writing to the 8-bit Data Length Control field (LINMR.DLC). The response data length equals DLC + 1 bytes. 381 32099I–01/2012 AT32UC3L016/32/64 • DLM=1: the response data length is defined by the Identifier bits according to the table below. Table 19-8. 19.6.5.11 Response Data Length if DLM = 1 IDCHR[5] IDCHR[4] Response Data Length [bytes] 0 0 2 0 1 2 1 0 4 1 1 8 Checksum The last frame field is the checksum. It is configured by the Checksum Type (LINMR.CHKTYP), and the Checksum Disable (LINMR.CHKDIS) bits. TXRDY will not be set after the last THR data write if enabled. Writing a one to CHKDIS will disable the automatic checksum generation/checking, and the user may send/check this last byte manually, disguised as a normal data. The checksum is an inverted 8-bit sum with carry, either: • over all data bytes, called a classic checksum. This is used for LIN 1.3 compliant slaves, and automatically managed when CHKDIS=0, and CHKTYP=1. • over all data bytes and the protected identifier, called an enhanced checksum. This is used for LIN 2.0 compliant slaves, and automatically managed when CHKDIS=0, and CHKTYP=0. 19.6.5.12 Frame Slot Mode A LIN master can be configured to use frame slots with a pre-defined minimum length. Writing a one to the Frame Slot Mode Disable bit (LINMR.FSDIS) disables this mode. This mode will not allow TXRDY to be set after a frame transfer until the entire frame slot duration has elapsed, in effect preventing the master from sending a new header. The LIN Transfer Complete bit (CSR.LINTC) will still be set after the checksum has been sent. Writing a one to CR.RSTST clears LINTC. Figure 19-26. Frame Slot Mode with Automatic Checksum Frame slot = TFrame_Maximum Frame Header Break Synch Data3 Interframe space Response space Protected Identifier Response Data 1 Data N-1 Data N Checksum TXRDY Frame Slot Mode Frame Slot Mode Disabled Enabled Write LINID Write THR Data 1 Data 2 Data 3 Data N LINTC The minimum frame slot size is determined by TFrame_Maximum, and calculated below (all values in bit periods): • THeader_Nominal = 34 382 32099I–01/2012 AT32UC3L016/32/64 • TFrame_Maximum = 1.4 x (THeader_Nominal + TResponse_Nominal + 1)(Note:) Note: The term “+1” leads to an integer result for TFrame_Max (LIN Specification 1.3) If the Checksum is sent (CHKDIS=0): • TResponse_Nominal = 10 x (NData + 1) • TFrame_Maximum = 1.4 x (34 + 10 x (DLC + 1 + 1) + 1) • TFrame_Maximum = 77 + 14 x DLC If the Checksum is not sent (CHKDIS=1): • TResponse_Nominal = 10 x NData • TFrame_Maximum = 1.4 x (34 + 10 x (DLC + 1) + 1) • TFrame_Maximum = 63 + 14 x DLC 19.6.6 LIN Errors These error bits are cleared by writing a one to CSR.RSTSTA. 19.6.6.1 Slave Not Responding Error (CSR.LINSNRE) This error is generated if no valid message appears within the TFrame_Maximum time frame slot, while the USART is expecting a response from another node (NACT=SUBSCRIBE). 19.6.6.2 Checksum Error (CSR.LINCE) This error is generated if the received checksum is wrong. This error can only be generated if the checksum feature is enabled (CHKDIS=0). 19.6.6.3 Identifier Parity Error (CSR.LINIPE) This error is generated if the identifier parity is wrong. This error can only be generated if parity is enabled (PARDIS=0). 19.6.6.4 Inconsistent Sync Field Error (CSR.LINISFE) This error is generated in slave mode if the Sync Field character received is not 0x55. Synchronization procedure is aborted. 19.6.6.5 Bit Error (CSR.LINBE) This error is generated if the value transmitted by the USART on Tx differs from the value sampled on Rx. If a bit error is detected, the transmission is aborted at the next byte border. 19.6.7 LIN Frame Handling 19.6.7.1 Master Node Configuration • Write a one to CR.TXEN and CR.RXEN to enable both transmitter and receiver • Select LIN mode and master node by writing to MR.MODE • Configure the baud rate by writing to CD and FP in BRGR • Configure the frame transfer by writing to NACT, PARDIS, CHKDIS, CHKTYPE, DLCM, FSDIS, and DLC in LINMR • Check that CSR.TXRDY is one • Send the header by writing to LINIR.IDCHR The following procedure depends on the NACT setting: 383 32099I–01/2012 AT32UC3L016/32/64 • Case 1: NACT=PUBLISH, the USART sends a response – Wait until TXRDY is a one – Send a byte by writing to THR.TXCHR – Repeat the two previous steps until there is no more data to send – Wait until CSR.LINTC is a one – Check for LIN errors • Case 2: NACT=SUBSCRIBE, the USART receives a response – Wait until RXRDY is a one – Read RHR.RXCHR – Repeat the two previous steps until there is no more data to read – Wait until LINTC is a one – Check for LIN errors • Case 3: NACT=IGNORE, the USART is not concerned by a response – Wait until LINTC is a one – Check for LIN errors Figure 19-27. Master Node Configuration, NACT=PUBLISH Frame slot = TFrame_Maximum Frame Header Break Synch Data3 Interframe space Response space Protected Identifier Response Data 1 Data N-1 Data N Checksum TXRDY FSDIS=1 FSDIS=0 RXRDY Write LINIR Write THR Data 1 Data 2 Data 3 Data N LINTC 384 32099I–01/2012 AT32UC3L016/32/64 Figure 19-28. Master Node Configuration, NACT=SUBSCRIBE Frame slot = TFrame_Maximum Frame Header Break Synch Data3 Interframe space Response space Protected Identifier Response Data 1 Data N-1 Data N Checksum TXRDY FSDIS=1 FSDIS=0 RXRDY Write LINIR Read RHR Data 1 Data N-2 Data N-1 Data N LINTC Figure 19-29. Master Node Configuration, NACT=IGNORE Frame slot = TFrame_Maximum Frame Break Response space Header Data3 Synch Protected Identifier Interframe space Response Data 1 Data N-1 Data N Checksum TXRDY FSDIS=1 FSDIS=0 RXRDY Write LINIR LINTC 19.6.7.2 Slave Node Configuration This is identical to the master node configuration above, except for: • LIN mode selected in MR.MODE is slave • When the baud rate is configured, wait until CSR.LINID is a one, then; • Check for LINISFE and LINPE errors, clear errors and LINIDby writing a one to RSTSTA • Read IDCHR • Configure the frame transfer by writing to NACT, PARDIS, CHKDIS, CHKTYPE, DLCM, and DLC in LINMR IMPORTANT: if NACT=PUBLISH, and this field is already correct, the LINMR register must still be written with this value in order to set TXRDY, and to request the corresponding Peripheral DMA Controller write transfer. The different NACT settings result in the same procedure as for the master node, see page 383. 385 32099I–01/2012 AT32UC3L016/32/64 Figure 19-30. Slave Node Configuration, NACT=PUBLISH Break Synch Protected Identifier Data 1 Data N-1 Data N Checksum Data N Checksum TXRDY RXRDY LINIDRX Read LINID Write THR Data 1 Data 2 Data 3 Data N LINTC Figure 19-31. Slave Node Configuration, NACT=SUBSCRIBE Break Synch Protected Identifier Data 1 Data N-1 TXRDY RXRDY LINIDRX Read LINID Read RHR Data 1 Data N-2 Data N-1 Data N LINTC Figure 19-32. Slave Node Configuration, NACT=IGNORE Break Synch Protected Identifier Data 1 Data N-1 Data N Checksum TXRDY RXRDY LINIDRX Read LINID Read RHR LINTC 19.6.8 LIN Frame Handling With The Peripheral DMA Controller The USART can be used together with the Peripheral DMA Controller in order to transfer data without processor intervention. The DMA Controller uses the TXRDY and RXRDY bits, to trigger one byte writes or reads. It always writes to THR, and it always reads RHR. 386 32099I–01/2012 AT32UC3L016/32/64 19.6.8.1 Master Node Configuration The Peripheral DMA Controller Mode bit (LINMR.PDCM) allows the user to select configuration: • PDCM=0: LIN configuration must be written to LINMR, it is not stored in the write buffer. • PDCM=1: LIN configuration is written by the DMA Controller to THR, and is stored in the write buffer. Since data transfer size is a byte, the transfer is split into two accesses. The first writes the NACT, PARDIS, CHKDIS, CHKTYP, DLM and FSDIS bits, while the second writes the DLC field. If NACT=PUBLISH, the write buffer will also contain the Identifier. When NACT=SUBSCRIBE, the read buffer contains the data. Figure 19-33. Master Node with Peripheral DMA Controller (PDCM=0) WRITE BUFFER DATA 0 Peripheral bus DATA 1 NODE ACTION = SUBSCRIBE Peripheral bus READ BUFFER Peripheral DMA Controller | | | | NODE ACTION = PUBLISH USART LIN CONTROLLER DATA 0 RXRDY Peripheral DMA Controller RXRDY USART LIN CONTROLLER TXRDY | | | | DATA N DATA N Figure 19-34. Master Node with Peripheral DMA Controller (PDCM=1) WRITE BUFFER WRITE BUFFER NACT PARDIS CHKDIS CHKTYP DLM FSDIS NACT PARDIS CHKDIS CHKTYP DLM FSDIS DLC DLC Peripheral bus NODE ACTION = PUBLISH Peripheral bus IDENTIFIER IDENTIFIER USART LIN CONTROLLER Peripheral DMA Controller DATA 0 | | | | DATA N READ BUFFER Peripheral DMA Controller RXRDY NODE ACTION = SUBSCRIBE USART LIN CONTROLLER RXRDY DATA 0 TXRDY | | | | DATA N 387 32099I–01/2012 AT32UC3L016/32/64 19.6.8.2 Slave Node Configuration In this mode, the Peripheral DMA Controller transfers only data. The user reads the Identifier from LINIR, and selects LIN mode by writing to LINMR. When NACT=PUBLISH the data is in the write buffer, while the read buffer contains the data when NACT=SUBSCRIBE. IMPORTANT: if in slave mode, NACT is already configured correctly as PUBLISH, the LINMR register must still be written with this value in order to set TXRDY, and to request the corresponding Peripheral DMA Controller write transfer. Figure 19-35. Slave Node with Peripheral DMA Controller WRITE BUFFER READ BUFFER DATA 0 DATA 0 Peripheral bus | | | | | | | | USART LIN CONTROLLER Peripheral DMA Controller TXRDY DATA N 19.6.9 NACT = SUBSCRIBE Peripheral Bus Peripheral DMA Controller USART LIN CONTROLLER RXRDY DATA N Wake-up Request Any node in a sleeping LIN cluster may request a wake-up. By writing to the Wakeup Signal Type bit (LINMR.WKUPTYP), the user can choose to send either a LIN 1.3 (WKUPTYP=1), or a LIN 2.0 (WKUPTYP=0) compliant wakeup request. Writing a one to the Send LIN Wakeup Signal bit (CR.LINWKUP), transmits a wakeup, and when completed sets LINTC. According to LIN 1.3, the wakeup request should be generated with the character 0x80 in order to impose eight successive dominant bits. According to LIN 2.0, the wakeup request is issued by forcing the bus into the dominant state for 250µs to 5ms. Sending the character 0xF0 does this, regardless of baud rate. • Baud rate max = 20 kbit/s -> one bit period = 50µs -> five bit periods = 250µs • Baud rate min = 1 kbit/s -> one bit period = 1ms -> five bit periods = 5ms 19.6.10 Bus Idle Time-out LIN bus inactivity should eventually cause slaves to time-out and enter sleep mode. LIN 1.3 specifies this to 25000 bit periods, whilst LIN 2.0 specifies 4seconds. For the time-out counter operation see Section 19.6.3.8 ”Receiver Time-out” on page 372. Table 19-9. Receiver Time-out Values (RTOR.TO) LIN Specification 2.0 1.3 Baud Rate Time-out period TO 1 000 bit/s 4 000 2 400 bit/s 9 600 9 600 bit/s 4s 38 400 19 200 bit/s 76 800 20 000 bit/s 80 000 - 25 000 bit periods 25 000 388 32099I–01/2012 AT32UC3L016/32/64 19.6.11 Test Modes The internal loopback feature enables on-board diagnostics, and allows the USART to operate in three different test modes, with reconfigured pin functionality, as shown below. 19.6.11.1 Normal Mode During normal operation, a receivers RXD pin is connected to a transmitters TXD pin. Figure 19-36. Normal Mode Configuration RXD Receiver TXD Transmitter 19.6.11.2 Automatic Echo Mode Automatic echo mode allows bit-by-bit retransmission. When a bit is received on the RXD pin, it is also sent to the TXD pin, as shown in Figure 19-37. Transmitter configuration has no effect. Figure 19-37. Automatic Echo Mode Configuration RXD Receiver TXD Transmitter 19.6.11.3 Local Loopback Mode Local loopback mode connects the output of the transmitter directly to the input of the receiver, as shown in Figure 19-38. The TXD and RXD pins are not used. The RXD pin has no effect on the receiver and the TXD pin is continuously driven high, as in idle state. Figure 19-38. Local Loopback Mode Configuration RXD Receiver Transmitter 19.6.11.4 1 TXD Remote Loopback Mode Remote loopback mode connects the RXD pin to the TXD pin, as shown in Figure 19-39. The transmitter and the receiver are disabled and have no effect. This mode allows bit-by-bit retransmission. 389 32099I–01/2012 AT32UC3L016/32/64 Figure 19-39. Remote Loopback Mode Configuration Receiver 1 RXD TXD Transmitter 19.6.12 Write Protection Registers To prevent single software errors from corrupting USART behavior, certain address spaces can be write-protected by writing the correct Write Protect KEY and a one to the Write Protect Enable bit in the Write Protect Mode Register (WPMR.WPKEY, and WPMR.WPEN). Disabling the write protection is done by writing the correct key, and a zero to WPEN. Write attempts to a write protected register are detected and the Write Protect Violation Status bit in the Write Protect Status Register (WPSR.WPVS) is set, while the Write Protect Violation Source field (WPSR.WPVSRC) indicates the targeted register. Writing the correct key to the Write Protect KEY bit (WPMR.WPKEY) clears WPVSRC and WPVS. The protected registers are: • ”Mode Register” on page 394 • ”Baud Rate Generator Register” on page 404 • ”Receiver Time-out Register” on page 405 • ”Transmitter Timeguard Register” on page 406 390 32099I–01/2012 AT32UC3L016/32/64 19.7 User Interface Table 19-10. USART Register Memory Map Offset Register Name Access Reset 0x0000 Control Register CR Write-only 0x00000000 0x0004 Mode Register MR Read-write 0x00000000 0x0008 Interrupt Enable Register IER Write-only 0x00000000 0x000C Interrupt Disable Register IDR Write-only 0x00000000 0x0010 Interrupt Mask Register IMR Read-only 0x00000000 0x0014 Channel Status Register CSR Read-only 0x00000000 0x0018 Receiver Holding Register RHR Read-only 0x00000000 0x001C Transmitter Holding Register THR Write-only 0x00000000 0x0020 Baud Rate Generator Register BRGR Read-write 0x00000000 0x0024 Receiver Time-out Register RTOR Read-write 0x00000000 0x0028 Transmitter Timeguard Register TTGR Read-write 0x00000000 0x0054 LIN Mode Register LINMR Read-write 0x00000000 0x0058 LIN Identifier Register LINIR Read-write 0x00000000 0x00E4 Write Protect Mode Register WPMR Read-write 0x00000000 0x00E8 Write Protect Status Register WPSR Read-only 0x00000000 0x00FC Version Register VERSION Read-only 0x–(1) Note: 1. Values in the Version Register vary with the version of the IP block implementation. 391 32099I–01/2012 AT32UC3L016/32/64 19.7.1 Name: Control Register CR Access Type: Write-only Offset: 0x0 Reset Value: 0x00000000 31 – 30 – 29 – 28 – 27 – 26 – 25 – 24 – 23 – 22 – 21 LINWKUP 20 LINABT 19 RTSDIS/RCS 18 RTSEN/FCS 17 – 16 – 15 RETTO 14 RSTNACK 13 – 12 SENDA 11 STTTO 10 STPBRK 9 STTBRK 8 RSTSTA 7 TXDIS 6 TXEN 5 RXDIS 4 RXEN 3 RSTTX 2 RSTRX 1 – 0 – • LINWKUP: Send LIN Wakeup Signal Writing a zero to this bit has no effect. Writing a one to this bit will sends a wakeup signal on the LIN bus. • LINABT: Abort LIN Transmission Writing a zero to this bit has no effect. Writing a one to this bit will abort the current LIN transmission. • RTSDIS/RCS: Request to Send Disable/Release SPI Chip Select Writing a zero to this bit has no effect. Writing a one to this bit when USART is not in SPI master mode drives RTS pin high. Writing a one to this bit when USART is in SPI master mode releases NSS (RTS pin). • RTSEN/FCS: Request to Send Enable/Force SPI Chip Select Writing a zero to this bit has no effect. Writing a one to this bit when USART is not in SPI master mode drives RTS low. Writing a one to this bit when USART is in SPI master mode when; FCS=0: has no effect. FCS=1: forces NSS (RTS pin) low, even if USART is not transmitting, in order to address SPI slave devices supporting the CSAAT Mode (Chip Select Active After Transfer). • RETTO: Rearm Time-out Writing a zero to this bit has no effect. Writing a one to this bit reloads the time-out counter and clears CSR.TIMEOUT. • RSTNACK: Reset Non Acknowledge Writing a zero to this bit has no effect. Writing a one to this bit clears CSR.NACK. • SENDA: Send Address Writing a zero to this bit has no effect. Writing a one to this bit will in multidrop mode send the next character written to THR as an address. • STTTO: Start Time-out Writing a zero to this bit has no effect. Writing a one to this bit will abort any current time-out count down, and trigger a new count down when the next character has been received. CSR.TIMEOUT is also cleared. 392 32099I–01/2012 AT32UC3L016/32/64 • STPBRK: Stop Break Writing a zero to this bit has no effect. Writing a one to this bit will stop the generation of break signal characters, and then send ones for TTGR.TG duration, or at least 12 bit periods. No effect if no break is being transmitted. • STTBRK: Start Break Writing a zero to this bit has no effect. Writing a one to this bit will start transmission of break characters when current characters present in THR and the transmit shift register have been sent. No effect if a break signal is already being generated. • RSTSTA: Reset Status Bits Writing a zero to this bit has no effect. Writing a one to this bit will clear the following bits in CSR: PARE, FRAME, OVRE, LINBE, LINSFE, LINIPE, LINCE, LINSNRE, and RXBRK. • TXDIS: Transmitter Disable Writing a zero to this bit has no effect. Writing a one to this bit disables the transmitter. • TXEN: Transmitter Enable Writing a zero to this bit has no effect. Writing a one to this bit enables the transmitter if TXDIS is zero. • RXDIS: Receiver Disable Writing a zero to this bit has no effect. Writing a one to this bit disables the receiver. • RXEN: Receiver Enable Writing a zero to this bit has no effect. Writing a one to this bit enables the receiver if RXDIS is zero. • RSTTX: Reset Transmitter Writing a zero to this bit has no effect. Writing a one to this bit will reset the transmitter. • RSTRX: Reset Receiver Writing a zero to this bit has no effect. Writing a one to this bit will reset the receiver. 393 32099I–01/2012 AT32UC3L016/32/64 19.7.2 Name: Mode Register MR Access Type: Read-write Offset: 0x4 Reset Value: 0x00000000 31 – 30 – 29 – 28 – 27 – 26 25 – 24 23 – 22 – 21 – 20 INACK 19 OVER 18 CLKO 17 MODE9 16 MSBF/CPOL 14 13 12 11 10 PAR 9 8 SYNC/CPHA 4 3 2 1 0 15 CHMODE 7 NBSTOP 6 CHRL 5 USCLKS MODE This register can only be written if the WPEN bit is cleared in the Write Protect Mode Register. • INACK: Inhibit Non Acknowledge 0: The NACK is generated. 1: The NACK is not generated. • OVER: Oversampling Mode 0: Oversampling at 16 times the baud rate. 1: Oversampling at 8 times the baud rate. • CLKO: Clock Output Select 0: The USART does not drive the CLK pin. 1: The USART drives the CLK pin unless USCLKS selects the external clock. • MODE9: 9-bit Character Length 0: CHRL defines character length. 1: 9-bit character length. • MSBF/CPOL: Bit Order or SPI Clock Polarity If USART does not operate in SPI Mode: MSBF=0: Least Significant Bit is sent/received first. MSBF=1: Most Significant Bit is sent/received first. If USART operates in SPI Mode, CPOL is used with CPHA to produce the required clock/data relationship between devices. CPOL=0: The inactive state value of CLK is logic level zero. CPOL=1: The inactive state value of CLK is logic level one. 394 32099I–01/2012 AT32UC3L016/32/64 • CHMODE: Channel Mode Table 19-11. CHMODE Mode Description 0 0 Normal Mode 0 1 Automatic Echo. Receiver input is connected to the TXD pin. 1 0 Local Loopback. Transmitter output is connected to the Receiver input. 1 1 Remote Loopback. RXD pin is internally connected to the TXD pin. • NBSTOP: Number of Stop Bits Table 19-12. NBSTOP Asynchronous (SYNC=0) Synchronous (SYNC=1) 0 0 1 stop bit 1 stop bit 0 1 1.5 stop bits Reserved 1 0 2 stop bits 2 stop bits 1 1 Reserved Reserved • PAR: Parity Type Table 19-13. PAR Parity Type 0 0 0 Even parity 0 0 1 Odd parity 0 1 0 Parity forced to 0 (Space) 0 1 1 Parity forced to 1 (Mark) 1 0 x No parity 1 1 x Multidrop mode • SYNC/CPHA: Synchronous Mode Select or SPI Clock Phase If USART does not operate in SPI Mode (MODE is … 0xE and 0xF): SYNC = 0: USART operates in Asynchronous Mode. SYNC = 1: USART operates in Synchronous Mode. If USART operates in SPI Mode, CPHA determines which edge of CLK causes data to change and which edge causes data to be captured. CPHA is used with CPOL to produce the required clock/data relationship between master and slave devices. CPHA = 0: Data is changed on the leading edge of CLK and captured on the following edge of CLK. CPHA = 1: Data is captured on the leading edge of CLK and changed on the following edge of CLK. 395 32099I–01/2012 AT32UC3L016/32/64 • CHRL: Character Length. Table 19-14. CHRL Character Length 0 0 5 bits 0 1 6 bits 1 0 7 bits 1 1 8 bits • USCLKS: Clock Selection Table 19-15. USCLKS Note: Selected Clock 0 0 CLK_USART 0 1 CLK_USART/DIV(1) 1 0 Reserved 1 1 CLK 1. The value of DIV is device dependent. Please refer to the Module Configuration section at the end of this chapter. • MODE Table 19-16. MODE Mode of the USART 0 0 0 0 Normal 0 0 1 0 Hardware Handshaking 1 0 1 0 LIN Master 1 0 1 1 LIN Slave 1 1 1 0 SPI Master 1 1 1 1 SPI Slave Others Reserved 396 32099I–01/2012 AT32UC3L016/32/64 19.7.3 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x8 Reset Value: 0x00000000 31 – 30 – 29 LINSNRE 28 LINCE 27 LINIPE 26 LINISFE 25 LINBE 24 – 23 – 22 – 21 – 20 – 19 CTSIC 18 – 17 – 16 – 15 LINTC 14 LINID 13 NACK/LINBK 12 RXBUFF 11 – 10 ITER/UNRE 9 TXEMPTY 8 TIMEOUT 7 PARE 6 FRAME 5 OVRE 4 – 3 – 2 RXBRK 1 TXRDY 0 RXRDY Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 397 32099I–01/2012 AT32UC3L016/32/64 19.7.4 Name: Interrupt Disable Register IDR Access Type: Write-only Offset: 0xC Reset Value: 0x00000000 31 – 30 – 29 LINSNRE 28 LINCE 27 LINIPE 26 LINISFE 25 LINBE 24 – 23 – 22 – 21 – 20 – 19 CTSIC 18 – 17 – 16 – 15 LINTC 14 LINID 13 NACK/LINBK 12 RXBUFF 11 – 10 ITER/UNRE 9 TXEMPTY 8 TIMEOUT 7 PARE 6 FRAME 5 OVRE 4 – 3 – 2 RXBRK 1 TXRDY 0 RXRDY Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 398 32099I–01/2012 AT32UC3L016/32/64 19.7.5 Name: Interrupt Mask Register IMR Access Type: Read-only Offset: 0x10 Reset Value: 0x00000000 31 – 30 – 29 LINSNRE 28 LINCE 27 LINIPE 26 LINISFE 25 LINBE 24 – 23 – 22 – 21 – 20 – 19 CTSIC 18 – 17 – 16 – 15 LINTC 14 LINID 13 NACK/LINBK 12 RXBUFF 11 – 10 ITER/UNRE 9 TXEMPTY 8 TIMEOUT 7 PARE 6 FRAME 5 OVRE 4 – 3 – 2 RXBRK 1 TXRDY 0 RXRDY 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. A bit in this register is cleared when the corresponding bit in IDR is written to one. A bit in this register is set when the corresponding bit in IER is written to one. 399 32099I–01/2012 AT32UC3L016/32/64 19.7.6 Name: Channel Status Register CSR Access Type: Read-only Offset: 0x14 Reset Value: 0x00000000 31 – 30 – 29 LINSNRE 28 LINCE 27 LINIPE 26 LINISFE 25 LINBE 24 – 23 CTS 22 – 21 – 20 – 19 CTSIC 18 – 17 – 16 – 15 LINTC 14 LINID 13 NACK/LINBK 12 RXBUFF 11 – 10 ITER/UNRE 9 TXEMPTY 8 TIMEOUT 7 PARE 6 FRAME 5 OVRE 4 – 3 – 2 RXBRK 1 TXRDY 0 RXRDY • LINSNRE: LIN Slave Not Responding Error 0: No LIN Slave Not Responding Error has been detected since the last RSTSTA. 1: A LIN Slave Not Responding Error has been detected since the last RSTSTA. • LINCE: LIN Checksum Error 0: No LIN Checksum Error has been detected since the last RSTSTA. 1: A LIN Checksum Error has been detected since the last RSTSTA. • LINIPE: LIN Identifier Parity Error 0: No LIN Identifier Parity Error has been detected since the last RSTSTA. 1: A LIN Identifier Parity Error has been detected since the last RSTSTA. • LINISFE: LIN Inconsistent Sync Field Error 0: No LIN Inconsistent Sync Field Error has been detected since the last RSTSTA 1: The USART is configured as a Slave node and a LIN Inconsistent Sync Field Error has been detected since the last RSTSTA. • LINBE: LIN Bit Error 0: No Bit Error has been detected since the last RSTSTA. 1: A Bit Error has been detected since the last RSTSTA. • CTS: Image of CTS Input 0: CTS is low. 1: CTS is high. • CTSIC: Clear to Send Input Change Flag 0: No change has been detected on the CTS pin since the last CSR read. 1: At least one change has been detected on the CTS pin since the last CSR read. • LINTC: LIN Transfer Completed 0: The USART is either idle or a LIN transfer is ongoing. 1: A LIN transfer has been completed since the last RSTSTA. • LINID: LIN Identifier 0: No LIN Identifier has been sent or received. 1: A LIN Identifier has been sent (master) or received (slave), since the last RSTSTA. • NACK: Non Acknowledge 0: No Non Acknowledge has been detected since the last RSTNACK. 1: At least one Non Acknowledge has been detected since the last RSTNACK. • RXBUFF: Reception Buffer Full 0: The Buffer Full signal from the Peripheral DMA Controller channel is inactive. 400 32099I–01/2012 AT32UC3L016/32/64 1: The Buffer Full signal from the Peripheral DMA Controller channel is active. • ITER/UNRE: Max number of Repetitions Reached or SPI Underrun Error If USART does not operate in SPI Slave Mode: ITER=0: Maximum number of repetitions has not been reached since the last RSTSTA. ITER=1: Maximum number of repetitions has been reached since the last RSTSTA. If USART operates in SPI Slave Mode: UNRE=0: No SPI underrun error has occurred since the last RSTSTA. UNRE=1: At least one SPI underrun error has occurred since the last RSTSTA. • TXEMPTY: Transmitter Empty 0: The transmitter is either disabled or there are characters in THR, or in the transmit shift register. 1: There are no characters in neither THR, nor in the transmit shift register. • TIMEOUT: Receiver Time-out 0: There has not been a time-out since the last Start Time-out command (CR.STTTO), or RTOR.TO is zero. 1: There has been a time-out since the last Start Time-out command. • PARE: Parity Error 0: Either no parity error has been detected, or the parity bit is a zero in multidrop mode, since the last RSTSTA. 1: Either at least one parity error has been detected, or the parity bit is a one in multidrop mode, since the last RSTSTA. • FRAME: Framing Error 0: No stop bit has been found as low since the last RSTSTA. 1: At least one stop bit has been found as low since the last RSTSTA. • OVRE: Overrun Error 0: No overrun error has occurred since the last RSTSTA. 1: At least one overrun error has occurred since the last RSTSTA. • RXBRK: Break Received/End of Break 0: No Break received or End of Break detected since the last RSTSTA. 1: Break received or End of Break detected since the last RSTSTA. • TXRDY: Transmitter Ready 0: The transmitter is either disabled, or a character in THR is waiting to be transferred to the transmit shift register, or an STTBRK command has been requested. As soon as the transmitter is enabled, TXRDY becomes one. 1: There is no character in the THR. • RXRDY: Receiver Ready 0: The receiver is either disabled, or no complete character has been received since the last read of RHR. If characters were being received when the receiver was disabled, RXRDY changes to 1 when the receiver is enabled. 1: At least one complete character has been received and RHR has not yet been read. 401 32099I–01/2012 AT32UC3L016/32/64 19.7.7 Name: Receiver Holding Register RHR Access Type: Read-only Offset: 0x18 Reset Value: 0x00000000 31 – 30 – 29 – 28 – 27 – 26 – 25 – 24 – 23 – 22 – 21 – 20 – 19 – 18 – 17 – 16 – 15 – 14 – 13 – 12 – 11 – 10 – 9 – 8 RXCHR[8] 7 6 5 4 3 2 1 0 RXCHR[7:0] • RXCHR: Received Character Last received character. 402 32099I–01/2012 AT32UC3L016/32/64 19.7.8 Name: Transmitter Holding Register THR Access Type: Write-only Offset: 0x1C Reset Value: 0x00000000 31 – 30 – 29 – 28 – 27 – 26 – 25 – 24 – 23 – 22 – 21 – 20 – 19 – 18 – 17 – 16 – 15 – 14 – 13 – 12 – 11 – 10 – 9 – 8 TXCHR[8] 7 6 5 4 3 2 1 0 TXCHR[7:0] • TXCHR: Character to be Transmitted If TXRDY is zero this field contains the next character to be transmitted. 403 32099I–01/2012 AT32UC3L016/32/64 19.7.9 Name: Baud Rate Generator Register BRGR Access Type: Read-write Offset: 0x20 Reset Value: 0x00000000 31 – 30 – 29 – 28 – 27 – 26 – 25 – 24 – 23 – 22 – 21 – 20 – 19 – 18 17 FP 16 15 14 13 12 11 10 9 8 3 2 1 0 CD[15:8] 7 6 5 4 CD[7:0] This register can only be written to if write protection is disabled, see ”Write Protect Mode Register” on page 410. • FP: Fractional Part 0: Fractional divider is disabled. 1 - 7: Baud rate resolution, defined by FP x 1/8. • CD: Clock Divider Table 19-17. SYNC = 1 or MODE = SPI (Master or Slave) SYNC = 0 CD OVER = 0 0 1 to 65535 OVER = 1 Baud Rate Clock Disabled Baud Rate = Selected Clock/16/CD Baud Rate = Selected Clock/8/CD Baud Rate = Selected Clock /CD 404 32099I–01/2012 AT32UC3L016/32/64 19.7.10 Name: Receiver Time-out Register RTOR Access Type: Read-write Offset: 0x24 Reset Value: 0x00000000 31 – 30 – 29 – 28 – 27 – 26 – 25 – 24 – 23 – 22 – 21 – 20 – 19 – 18 – 17 – 16 TO[16] 15 14 13 12 11 10 9 8 3 2 1 0 TO[15:8] 7 6 5 4 TO[7:0] This register can only be written to if write protection is disabled, see ”Write Protect Mode Register” on page 410. • TO: Time-out Value 0: The receiver Time-out is disabled. 1 - 131071: The receiver Time-out is enabled and the time-out delay is TO x bit period. Note that the size of the TO counter is device dependent, see the Module Configuration section. 405 32099I–01/2012 AT32UC3L016/32/64 19.7.11 Name: Transmitter Timeguard Register TTGR Access Type: Read-write Offset: 0x28 Reset Value: 0x00000000 31 – 30 – 29 – 28 – 27 – 26 – 25 – 24 – 23 – 22 – 21 – 20 – 19 – 18 – 17 – 16 – 15 – 14 – 13 – 12 – 11 – 10 – 9 – 8 – 7 6 5 4 3 2 1 0 TG This register can only be written to if write protection is disabled, see ”Write Protect Mode Register” on page 410. • TG: Timeguard Value 0: The transmitter Timeguard is disabled. 1 - 255: The transmitter timeguard is enabled and the timeguard delay is TG x bit period. 406 32099I–01/2012 AT32UC3L016/32/64 19.7.12 Name: LIN Mode Register LINMR Access Type: Read-write Offset: 0x54 Reset Value: 0x00000000 31 – 30 – 29 – 28 – 27 – 26 – 25 – 24 – 23 – 22 – 21 – 20 – 19 – 18 – 17 – 16 PDCM 15 14 13 12 11 10 9 8 3 CHKDIS 2 PARDIS 1 DLC 7 WKUPTYP 6 FSDIS 5 DLM 4 CHKTYP 0 NACT • PDCM: Peripheral DMA Controller Mode 0: The LIN mode register is not written by the Peripheral DMA Controller. 1: The LIN mode register is, except for this bit, written by the Peripheral DMA Controller. • DLC: Data Length Control 0 - 255: If DLM=0 this field defines the response data length to DLC+1 bytes. • WKUPTYP: Wakeup Signal Type 0: Writing a one to CR.LINWKUP will send a LIN 2.0 wakeup signal. 1: Writing a one to CR.LINWKUP will send a LIN 1.3 wakeup signal. • FSDIS: Frame Slot Mode Disable 0: The Frame Slot mode is enabled. 1: The Frame Slot mode is disabled. • DLM: Data Length Mode 0: The response data length is defined by DLC. 1: The response data length is defined by bits 4 and 5 of the Identifier (LINIR.IDCHR). • CHKTYP: Checksum Type 0: LIN 2.0 “Enhanced” checksum 1: LIN 1.3 “Classic” checksum • CHKDIS: Checksum Disable 0: Checksum is automatically computed and sent when master, and checked when slave. 1: Checksum is not computed and sent, nor checked. • PARDIS: Parity Disable 0: Identifier parity is automatically computed and sent when master, and checked when slave. 1: Identifier parity is not computed and sent, nor checked. • NACT: LIN Node Action Table 19-18. NACT 0 Mode Description 0 PUBLISH: The USART transmits the response. 407 32099I–01/2012 AT32UC3L016/32/64 Table 19-18. 0 1 SUBSCRIBE: The USART receives the response. 1 0 IGNORE: The USART does not transmit and does not receive the response. 1 1 Reserved 408 32099I–01/2012 AT32UC3L016/32/64 19.7.13 Name: LIN Identifier Register LINIR Access Type: Read-write or Read-only Offset: 0x58 Reset Value: 0x00000000 31 – 30 – 29 – 28 – 27 – 26 – 25 – 24 – 23 – 22 – 21 – 20 – 19 – 18 – 17 – 16 – 15 – 14 – 13 – 12 – 11 – 10 – 9 – 8 – 7 6 5 4 3 2 1 0 IDCHR • IDCHR: Identifier Character If USART is in LIN master mode, the IDCHR field is read-write, and its value is the Identifier character to be transmitted. If USART is in LIN slave mode, the IDCHR field is read-only, and its value is the last received Identifier character. 409 32099I–01/2012 AT32UC3L016/32/64 19.7.14 Write Protect Mode Register Register Name: WPMR Access Type: Read-write Offset: 0xE4 Reset Value: See Table 19-10 31 30 29 28 27 WPKEY[23:16] 26 25 24 23 22 21 20 19 18 17 16 11 10 9 8 3 — 2 — 1 — 0 WPEN WPKEY[15:8] 15 14 13 12 WPKEY[7:0] 7 — 6 — 5 — 4 — • WPKEY: Write Protect KEY Has to be written to 0x555341 (“USA” in ASCII) in order to successfully write WPEN. Always reads as zero. • WPEN: Write Protect Enable 0 = Write protection disabled. 1 = Write protection enabled. Protects the registers: • ”Mode Register” on page 394 • ”Baud Rate Generator Register” on page 404 • ”Receiver Time-out Register” on page 405 • ”Transmitter Timeguard Register” on page 406 410 32099I–01/2012 AT32UC3L016/32/64 19.7.15 Write Protect Status Register Register Name: WPSR Access Type: Read-only Offset: 0xE8 Reset Value: See Table 19-10 31 — 30 — 29 — 28 — 27 — 26 — 25 — 24 — 23 22 21 20 19 WPVSRC[15:8] 18 17 16 15 14 13 12 11 10 9 8 3 — 2 — 1 — 0 WPVS WPVSRC[7:0] 7 — 6 — 5 — 4 — • WPVSRC: Write Protect Violation Source If WPVS=1 this field indicates which write-protected register was unsuccessfully written to, either by address offset or code. • WPVS: Write Protect Violation Status 0= No write protect violation has occurred since the last WPSR read. 1= A write protect violation has occurred since the last WPSR read. Note: Reading WPSR automatically clears all fields. 411 32099I–01/2012 AT32UC3L016/32/64 19.7.16 Version Register Name: VERSION Access Type: Read-only Offset: 0xFC Reset Value: - 31 – 30 – 29 – 28 – 27 – 26 – 25 – 24 – 23 – 22 – 21 – 20 – 19 18 17 16 15 – 14 – 13 – 12 – 11 10 9 VERSION[11:8] 8 7 6 5 4 3 2 0 MFN 1 VERSION[7:0] • MFN Reserved. No functionality associated. • VERSION Version of the module. No functionality associated. 412 32099I–01/2012 AT32UC3L016/32/64 19.8 Module Configuration The specific configuration for each USART instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 19-19. USART Configuration Feature USART0 USART1 USART2 USART3 17 bit 17 bit 17 bit 17 bit 8 8 8 8 Receiver Time-out Counter Size (Size of the RTOR.TO field) DIV Value for divided CLK_USART Table 19-20. USART Clocks Module Name Clock Name Description USART0 CLK_USART0 Clock for the USART0 bus interface USART1 CLK_USART1 Clock for the USART1 bus interface USART2 CLK_USART2 Clock for the USART2 bus interface USART3 CLK_USART3 Clock for the USART3 bus interface Table 19-21. Register Reset Values Register Reset Value VERSION 0x00000440 413 32099I–01/2012 AT32UC3L016/32/64 20. Serial Peripheral Interface (SPI) Rev: 2.1.1.3 20.1 Features • Compatible with an embedded 32-bit microcontroller • Supports communication with serial external devices – Four chip selects with external decoder support allow communication with up to 15 peripherals – Serial memories, such as DataFlash and 3-wire EEPROMs – Serial peripherals, such as ADCs, DACs, LCD controllers, CAN controllers and Sensors – External co-processors • Master or Slave Serial Peripheral Bus Interface – 4 - to 16-bit programmable data length per chip select – Programmable phase and polarity per chip select – Programmable transfer delays between consecutive transfers and between clock and data per chip select – Programmable delay between consecutive transfers – Selectable mode fault detection • Connection to Peripheral DMA Controller channel capabilities optimizes data transfers – One channel for the receiver, one channel for the transmitter – Next buffer support – Four character FIFO in reception 20.2 Overview The Serial Peripheral Interface (SPI) circuit is a synchronous serial data link that provides communication with external devices in Master or Slave mode. It also enables communication between processors if an external processor is connected to the system. The Serial Peripheral Interface is essentially a shift register that serially transmits data bits to other SPIs. During a data transfer, one SPI system acts as the “master”' which controls the data flow, while the other devices act as “slaves'' which have data shifted into and out by the master. Different CPUs can take turn being masters (Multiple Master Protocol opposite to Single Master Protocol where one CPU is always the master while all of the others are always slaves) and one master may simultaneously shift data into multiple slaves. However, only one slave may drive its output to write data back to the master at any given time. A slave device is selected when the master asserts its NSS signal. If multiple slave devices exist, the master generates a separate slave select signal for each slave (NPCS). The SPI system consists of two data lines and two control lines: • Master Out Slave In (MOSI): this data line supplies the output data from the master shifted into the input(s) of the slave(s). • Master In Slave Out (MISO): this data line supplies the output data from a slave to the input of the master. There may be no more than one slave transmitting data during any particular transfer. • Serial Clock (SPCK): this control line is driven by the master and regulates the flow of the data bits. The master may transmit data at a variety of baud rates; the SPCK line cycles once for each bit that is transmitted. • Slave Select (NSS): this control line allows slaves to be turned on and off by hardware. 414 32099I–01/2012 AT32UC3L016/32/64 20.3 Block Diagram Figure 20-1. SPI Block Diagram Peripheral DMA Controller Peripheral Bus SPCK MISO CLK_SPI MOSI Spi Interface I/O Controller NPCS0/NSS NPCS1 NPCS2 Interrupt Control NPCS3 SPI Interrupt 20.4 Application Block Diagram Figure 20-2. Application Block Diagram: Single Master/Multiple Slave Implementation Spi Master SPCK SPCK MISO MISO MOSI MOSI NPCS0 Slave 0 NSS NPCS1 NPCS2 NPCS3 NC SPCK MISO MOSI Slave 1 NSS SPCK MISO MOSI Slave 2 NSS 415 32099I–01/2012 AT32UC3L016/32/64 20.5 I/O Lines Description Table 20-1. I/O Lines Description Type 20.6 Pin Name Pin Description Master Slave MISO Master In Slave Out Input Output MOSI Master Out Slave In Output Input SPCK Serial Clock Output Input NPCS1-NPCS3 Peripheral Chip Selects Output Unused NPCS0/NSS Peripheral Chip Select/Slave Select Output Input Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 20.6.1 I/O Lines The pins used for interfacing the compliant external devices may be multiplexed with I/O lines. The user must first configure the I/O Controller to assign the SPI pins to their peripheral functions. 20.6.2 Clocks The clock for the SPI bus interface (CLK_SPI) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the SPI before disabling the clock, to avoid freezing the SPI in an undefined state. 20.6.3 Interrupts The SPI interrupt request line is connected to the interrupt controller. Using the SPI interrupt requires the interrupt controller to be programmed first. 20.7 20.7.1 Functional Description Modes of Operation The SPI operates in master mode or in slave mode. Operation in master mode is configured by writing a one to the Master/Slave Mode bit in the Mode Register (MR.MSTR). The pins NPCS0 to NPCS3 are all configured as outputs, the SPCK pin is driven, the MISO line is wired on the receiver input and the MOSI line driven as an output by the transmitter. If the MR.MSTR bit is written to zero, the SPI operates in slave mode. The MISO line is driven by the transmitter output, the MOSI line is wired on the receiver input, the SPCK pin is driven by the transmitter to synchronize the receiver. The NPCS0 pin becomes an input, and is used as a Slave Select signal (NSS). The pins NPCS1 to NPCS3 are not driven and can be used for other purposes. The data transfers are identically programmable for both modes of operations. The baud rate generator is activated only in master mode. 416 32099I–01/2012 AT32UC3L016/32/64 20.7.2 Data Transfer Four combinations of polarity and phase are available for data transfers. The clock polarity is configured with the Clock Polarity bit in the Chip Select Registers (CSRn.CPOL). The clock phase is configured with the Clock Phase bit in the CSRn registers (CSRn.NCPHA). These two bits determine the edges of the clock signal on which data is driven and sampled. Each of the two bits has two possible states, resulting in four possible combinations that are incompatible with one another. Thus, a master/slave pair must use the same parameter pair values to communicate. If multiple slaves are used and fixed in different configurations, the master must reconfigure itself each time it needs to communicate with a different slave. Table 20-2 on page 417 shows the four modes and corresponding parameter settings. Table 20-2. SPI modes SPI Mode CPOL NCPHA 0 0 1 1 0 0 2 1 1 3 1 0 Figure 20-3 on page 417 and Figure 20-4 on page 418 show examples of data transfers. Figure 20-3. SPI Transfer Format (NCPHA = 1, 8 bits per transfer) SPCK cycle (for reference) 2 1 3 4 5 6 7 8 SPCK (CPOL = 0) SPCK (CPOL = 1) MOSI (from master) MISO (from slave) MSB MSB 6 5 4 3 2 1 LSB 6 5 4 3 2 1 LSB *** NSS (to slave) *** Not Defined, but normaly MSB of previous character received 417 32099I–01/2012 AT32UC3L016/32/64 Figure 20-4. SPI Transfer Format (NCPHA = 0, 8 bits per transfer) SPCK cycle (for reference) 1 2 3 4 5 6 7 8 SPCK (CPOL = 0) SPCK (CPOL = 1) MOSI (from master) MISO (from slave) *** MSB 6 5 4 3 2 1 MSB 6 5 4 3 2 1 LSB LSB NSS (to slave) *** Not Defined, but normaly LSB of previous character transmitted 20.7.3 Master Mode Operations When configured in master mode, the SPI uses the internal programmable baud rate generator as clock source. It fully controls the data transfers to and from the slave(s) connected to the SPI bus. The SPI drives the chip select line to the slave and the serial clock signal (SPCK). The SPI features two holding registers, the Transmit Data Register (TDR) and the Receive Data Register (RDR), and a single Shift Register. The holding registers maintain the data flow at a constant rate. After enabling the SPI, a data transfer begins when the processor writes to the TDR register. The written data is immediately transferred in the Shift Register and transfer on the SPI bus starts. While the data in the Shift Register is shifted on the MOSI line, the MISO line is sampled and shifted in the Shift Register. Transmission cannot occur without reception. Before writing to the TDR, the Peripheral Chip Select field in TDR (TDR.PCS) must be written in order to select a slave. If new data is written to TDR during the transfer, it stays in it until the current transfer is completed. Then, the received data is transferred from the Shift Register to RDR, the data in TDR is loaded in the Shift Register and a new transfer starts. The transfer of a data written in TDR in the Shift Register is indicated by the Transmit Data Register Empty bit in the Status Register (SR.TDRE). When new data is written in TDR, this bit is cleared. The SR.TDRE bit is used to trigger the Transmit Peripheral DMA Controller channel. The end of transfer is indicated by the Transmission Registers Empty bit in the SR register (SR.TXEMPTY). If a transfer delay (CSRn.DLYBCT) is greater than zero for the last transfer, SR.TXEMPTY is set after the completion of said delay. The CLK_SPI can be switched off at this time. During reception, received data are transferred from the Shift Register to the reception FIFO. The FIFO can contain up to 4 characters (both Receive Data and Peripheral Chip Select fields). While a character of the FIFO is unread, the Receive Data Register Full bit in SR remains high (SR.RDRF). Characters are read through the RDR register. If the four characters stored in the FIFO are not read and if a new character is stored, this sets the Overrun Error Status bit in the SR register (SR.OVRES). The procedure to follow in such a case is described in Section 20.7.3.8. 418 32099I–01/2012 AT32UC3L016/32/64 Figure 20-5 on page 419shows a block diagram of the SPI when operating in master mode. Figure 20-6 on page 420 shows a flow chart describing how transfers are handled. 20.7.3.1 Master mode block diagram Figure 20-5. Master Mode Block Diagram CSR0..3 SCBR CLK_SPI Baud Rate Generator SPCK SPI Clock RXFIFOEN RDR RDRF OVRES RD CSR0..3 LSB MISO BITS NCPHA 0 CPOL 1 4 – Character FIFO MSB Shift Register MOSI TDR TD TDRE RXFIFOEN RDR CSR0..3 CSNAAT CSAAT PS MR 0 1 4 – Character FIFO NPCS3 PCSDEC PCS 0 TDR NPCS2 Current Peripheral NPCS1 PCS NPCS0 1 MSTR MODF NPCS0 MODFDIS 419 32099I–01/2012 AT32UC3L016/32/64 20.7.3.2 Master mode flow diagram Figure 20-6. Master Mode Flow Diagram SPI Enable - NPCS defines the current Chip Select - CSAAT, DLYBS, DLYBCT refer to the fields of the Chip Select Register corresponding to the Current Chip Select - When NPCS is 0xF, CSAAT is 0. 1 TDRE ? 0 1 CSAAT ? PS ? 0 1 0 Fixed peripheral PS ? 1 Fixed peripheral 0 Variable peripheral Variable peripheral TDR(PCS) = NPCS ? no NPCS = TDR(PCS) NPCS = MR(PCS) yes MR(PCS) = NPCS ? no NPCS = 0xF NPCS = 0xF Delay DLYBCS Delay DLYBCS NPCS = TDR(PCS) NPCS = MR(PCS), TDR(PCS) Delay DLYBS Serializer = TDR(TD) TDRE = 1 Data Transfer RDR(RD) = Serializer RDRF = 1 Delay DLYBCT 0 TDRE ? 1 1 CSAAT ? 0 NPCS = 0xF Delay DLYBCS 420 32099I–01/2012 AT32UC3L016/32/64 20.7.3.3 Clock generation The SPI Baud rate clock is generated by dividing the CLK_SPI , by a value between 1 and 255. This allows a maximum operating baud rate at up to CLK_SPI and a minimum operating baud rate of CLK_SPI divided by 255. Writing the Serial Clock Baud Rate field in the CSRn registers (CSRn.SCBR) to zero is forbidden. Triggering a transfer while CSRn.SCBR is zero can lead to unpredictable results. At reset, CSRn.SCBR is zero and the user has to configure it at a valid value before performing the first transfer. The divisor can be defined independently for each chip select, as it has to be configured in the CSRn.SCBR field. This allows the SPI to automatically adapt the baud rate for each interfaced peripheral without reprogramming. 20.7.3.4 Transfer delays Figure 20-7 on page 421 shows a chip select transfer change and consecutive transfers on the same chip select. Three delays can be configured to modify the transfer waveforms: • The delay between chip selects, programmable only once for all the chip selects by writing to the Delay Between Chip Selects field in the MR register (MR.DLYBCS). Allows insertion of a delay between release of one chip select and before assertion of a new one. • The delay before SPCK, independently programmable for each chip select by writing the Delay Before SPCK field in the CSRn registers (CSRn.DLYBS). Allows the start of SPCK to be delayed after the chip select has been asserted. • The delay between consecutive transfers, independently programmable for each chip select by writing the Delay Between Consecutive Transfers field in the CSRn registers (CSRn.DLYBCT). Allows insertion of a delay between two transfers occurring on the same chip select These delays allow the SPI to be adapted to the interfaced peripherals and their speed and bus release time. Figure 20-7. Programmable Delays Chip Select 1 Chip Select 2 SPCK DLYBCS DLYBS DLYBCT DLYBCT 421 32099I–01/2012 AT32UC3L016/32/64 20.7.3.5 Peripheral selection The serial peripherals are selected through the assertion of the NPCS0 to NPCS3 signals. By default, all the NPCS signals are high before and after each transfer. The peripheral selection can be performed in two different ways: • Fixed Peripheral Select: SPI exchanges data with only one peripheral • Variable Peripheral Select: Data can be exchanged with more than one peripheral Fixed Peripheral Select is activated by writing a zero to the Peripheral Select bit in MR (MR.PS). In this case, the current peripheral is defined by the MR.PCS field and the TDR.PCS field has no effect. Variable Peripheral Select is activated by writing a one to the MR.PS bit . The TDR.PCS field is used to select the current peripheral. This means that the peripheral selection can be defined for each new data. The Fixed Peripheral Selection allows buffer transfers with a single peripheral. Using the Peripheral DMA Controller is an optimal means, as the size of the data transfer between the memory and the SPI is either 4 bits or 16 bits. However, changing the peripheral selection requires the Mode Register to be reprogrammed. The Variable Peripheral Selection allows buffer transfers with multiple peripherals without reprogramming the MR register. Data written to TDR is 32-bits wide and defines the real data to be transmitted and the peripheral it is destined to. Using the Peripheral DMA Controller in this mode requires 32-bit wide buffers, with the data in the LSBs and the PCS and LASTXFER fields in the MSBs, however the SPI still controls the number of bits (8 to16) to be transferred through MISO and MOSI lines with the CSRn registers. This is not the optimal means in term of memory size for the buffers, but it provides a very effective means to exchange data with several peripherals without any intervention of the processor. 20.7.3.6 Peripheral chip select decoding The user can configure the SPI to operate with up to 15 peripherals by decoding the four Chip Select lines, NPCS0 to NPCS3 with an external logic. This can be enabled by writing a one to the Chip Select Decode bit in the MR register (MR.PCSDEC). When operating without decoding, the SPI makes sure that in any case only one chip select line is activated, i.e. driven low at a time. If two bits are defined low in a PCS field, only the lowest numbered chip select is driven low. When operating with decoding, the SPI directly outputs the value defined by the PCS field of either the MR register or the TDR register (depending on PS). As the SPI sets a default value of 0xF on the chip select lines (i.e. all chip select lines at one) when not processing any transfer, only 15 peripherals can be decoded. The SPI has only four Chip Select Registers, not 15. As a result, when decoding is activated, each chip select defines the characteristics of up to four peripherals. As an example, the CRS0 register defines the characteristics of the externally decoded peripherals 0 to 3, corresponding to the PCS values 0x0 to 0x3. Thus, the user has to make sure to connect compatible peripherals on the decoded chip select lines 0 to 3, 4 to 7, 8 to 11 and 12 to 14. 20.7.3.7 Peripheral deselection When operating normally, as soon as the transfer of the last data written in TDR is completed, the NPCS lines all rise. This might lead to runtime error if the processor is too long in responding 422 32099I–01/2012 AT32UC3L016/32/64 to an interrupt, and thus might lead to difficulties for interfacing with some serial peripherals requiring the chip select line to remain active during a full set of transfers. To facilitate interfacing with such devices, the CSRn registers can be configured with the Chip Select Active After Transfer bit written to one (CSRn.CSAAT) . This allows the chip select lines to remain in their current state (low = active) until transfer to another peripheral is required. When the CSRn.CSAAT bit is written to qero, the NPCS does not rise in all cases between two transfers on the same peripheral. During a transfer on a Chip Select, the SR.TDRE bit rises as soon as the content of the TDR is transferred into the internal shifter. When this bit is detected the TDR can be reloaded. If this reload occurs before the end of the current transfer and if the next transfer is performed on the same chip select as the current transfer, the Chip Select is not de-asserted between the two transfers. This might lead to difficulties for interfacing with some serial peripherals requiring the chip select to be de-asserted after each transfer. To facilitate interfacing with such devices, the CSRn registers can be configured with the Chip Select Not Active After Transfer bit (CSRn.CSNAAT) written to one. This allows to de-assert systematically the chip select lines during a time DLYBCS. (The value of the CSRn.CSNAAT bit is taken into account only if the CSRn.CSAAT bit is written to zero for the same Chip Select). Figure 20-8 on page 424 shows different peripheral deselection cases and the effect of the CSRn.CSAAT and CSRn.CSNAAT bits. 20.7.3.8 FIFO management A FIFO has been implemented in Reception FIFO (both in master and in slave mode), in order to be able to store up to 4 characters without causing an overrun error. If an attempt is made to store a fifth character, an overrun error rises. If such an event occurs, the FIFO must be flushed. There are two ways to Flush the FIFO: • By performing four read accesses of the RDR (the data read must be ignored) • By writing a one to the Flush Fifo Command bit in the CR register (CR.FLUSHFIFO). After that, the SPI is able to receive new data. 423 32099I–01/2012 AT32UC3L016/32/64 Figure 20-8. Peripheral Deselection CSAAT = 0 and CSNAAT = 0 TDRE NPCS[0..3] CSAAT = 1 and CSNAAT= 0 / 1 DLYBCT DLYBCT A A A A DLYBCS A DLYBCS PCS = A PCS = A Write TDR TDRE NPCS[0..3] DLYBCT DLYBCT A A A A DLYBCS A DLYBCS PCS=A PCS = A Write TDR TDRE NPCS[0..3] DLYBCT DLYBCT A B B A DLYBCS DLYBCS PCS = B PCS = B Write TDR CSAAT = 0 and CSNAAT = 0 CSAAT = 0 and CSNAAT = 1 DLYBCT DLYBCT TDRE NPCS[0..3] A A A A DLYBCS PCS = A PCS = A Write TDR Figure 20-8 on page 424 shows different peripheral deselection cases and the effect of the CSRn.CSAAT and CSRn.CSNAAT bits. 20.7.3.9 Mode fault detection The SPI is capable of detecting a mode fault when it is configured in master mode and NPCS0, MOSI, MISO, and SPCK are configured as open drain through the I/O Controller with either internal or external pullup resistors. If the I/O Controller does not have open-drain capability, mode fault detection must be disabled by writing a one to the Mode Fault Detection bit in the MR 424 32099I–01/2012 AT32UC3L016/32/64 register (MR.MODFDIS). In systems with open-drain I/O lines, a mode fault is detected when a low level is driven by an external master on the NPCS0/NSS signal. When a mode fault is detected, the Mode Fault Error bit in the SR (SR.MODF) is set until the SR is read and the SPI is automatically disabled until re-enabled by writing a one to the SPI Enable bit in the CR register (CR.SPIEN). By default, the mode fault detection circuitry is enabled. The user can disable mode fault detection by writing a one to the Mode Fault Detection bit in the MR register (MR.MODFDIS). 20.7.4 SPI Slave Mode When operating in slave mode, the SPI processes data bits on the clock provided on the SPI clock pin (SPCK). The SPI waits for NSS to go active before receiving the serial clock from an external master. When NSS falls, the clock is validated on the serializer, which processes the number of bits defined by the Bits Per Transfer field of the Chip Select Register 0 (CSR0.BITS). These bits are processed following a phase and a polarity defined respectively by the CSR0.NCPHA and CSR0.CPOL bits. Note that the BITS, CPOL, and NCPHA bits of the other Chip Select Registers have no effect when the SPI is configured in Slave Mode. The bits are shifted out on the MISO line and sampled on the MOSI line. When all the bits are processed, the received data is transferred in the Receive Data Register and the SR.RDRF bit rises. If the RDR register has not been read before new data is received, the SR.OVRES bit is set. Data is loaded in RDR even if this flag is set. The user has to read the SR register to clear the SR.OVRES bit. When a transfer starts, the data shifted out is the data present in the Shift Register. If no data has been written in the TDR register, the last data received is transferred. If no data has been received since the last reset, all bits are transmitted low, as the Shift Register resets to zero. When a first data is written in TDR, it is transferred immediately in the Shift Register and the SR.TDRE bit rises. If new data is written, it remains in TDR until a transfer occurs, i.e. NSS falls and there is a valid clock on the SPCK pin. When the transfer occurs, the last data written in TDR is transferred in the Shift Register and the SR.TDRE bit rises. This enables frequent updates of critical variables with single transfers. Then, a new data is loaded in the Shift Register from the TDR. In case no character is ready to be transmitted, i.e. no character has been written in TDR since the last load from TDR to the Shift Register, the Shift Register is not modified and the last received character is retransmitted. In this case the Underrun Error Status bit is set in SR (SR.UNDES). Figure 20-9 on page 426 shows a block diagram of the SPI when operating in slave mode. 425 32099I–01/2012 AT32UC3L016/32/64 Figure 20-9. Slave Mode Functional Block Diagram SPCK SPI Clock NSS SPIEN RXFIFOEN SPIENS RDR SPIDIS RDRF OVRES RD CSR0 BITS NCPHA CPOL MOSI LSB 0 1 4 - Character FIFO MSB Shift Register MISO UNDES TDR TD TDRE 426 32099I–01/2012 AT32UC3L016/32/64 20.8 User Interface Table 20-3. Note: SPI Register Memory Map Offset Register Register Name Access Reset 0x00 Control Register CR Write-only 0x00000000 0x04 Mode Register MR Read/Write 0x00000000 0x08 Receive Data Register RDR Read-only 0x00000000 0x0C Transmit Data Register TDR Write-only 0x00000000 0x10 Status Register SR Read-only 0x00000000 0x14 Interrupt Enable Register IER Write-only 0x00000000 0x18 Interrupt Disable Register IDR Write-only 0x00000000 0x1C Interrupt Mask Register IMR Read-only 0x00000000 0x30 Chip Select Register 0 CSR0 Read/Write 0x00000000 0x34 Chip Select Register 1 CSR1 Read/Write 0x00000000 0x38 Chip Select Register 2 CSR2 Read/Write 0x00000000 0x3C Chip Select Register 3 CSR3 Read/Write 0x00000000 0x E4 Write Protection Control Register WPCR Read/Write 0X00000000 0xE8 Write Protection Status Register WPSR Read-only 0x00000000 0xF8 Features Register FEATURES Read-only - (1) 0xFC Version Register VERSION Read-only - (1) 1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter. 427 32099I–01/2012 AT32UC3L016/32/64 20.8.1 Name: Control Register CR Access Type: Write-only Offset: 0x00 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - LASTXFER 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - FLUSHFIFO 7 6 5 4 3 2 1 0 SWRST - - - - - SPIDIS SPIEN • LASTXFER: Last Transfer 1: The current NPCS will be deasserted after the character written in TD has been transferred. When CSRn.CSAAT is one, this allows to close the communication with the current serial peripheral by raising the corresponding NPCS line as soon as TD transfer has completed. 0: Writing a zero to this bit has no effect. • FLUSHFIFO: Flush Fifo Command 1: If The FIFO Mode is enabled (MR.FIFOEN written to one) and if an overrun error has been detected, this command allows to empty the FIFO. 0: Writing a zero to this bit has no effect. • SWRST: SPI Software Reset 1: Writing a one to this bit will reset the SPI. A software-triggered hardware reset of the SPI interface is performed. The SPI is in slave mode after software reset. Peripheral DMA Controller channels are not affected by software reset. 0: Writing a zero to this bit has no effect. • SPIDIS: SPI Disable 1: Writing a one to this bit will disable the SPI. As soon as SPIDIS is written to one, the SPI finishes its transfer, all pins are set in input mode and no data is received or transmitted. If a transfer is in progress, the transfer is finished before the SPI is disabled. If both SPIEN and SPIDIS are equal to one when the CR register is written, the SPI is disabled. 0: Writing a zero to this bit has no effect. • SPIEN: SPI Enable 1: Writing a one to this bit will enable the SPI to transfer and receive data. 0: Writing a zero to this bit has no effect. 428 32099I–01/2012 AT32UC3L016/32/64 20.8.2 Name: Mode Register MR Access Type: Read/Write Offset: 0x04 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 DLYBCS 23 22 21 20 - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 LLB RXFIFOEN - MODFDIS - PCSDEC PS MSTR PCS • DLYBCS: Delay Between Chip Selects This field defines the delay from NPCS inactive to the activation of another NPCS. The DLYBCS time guarantees nonoverlapping chip selects and solves bus contentions in case of peripherals having long data float times. If DLYBCS is less than or equal to six, six CLK_SPI periods will be inserted by default. Otherwise, the following equation determines the delay: DLYBCS Delay Between Chip Selects = ----------------------CLKSPI • PCS: Peripheral Chip Select This field is only used if Fixed Peripheral Select is active (PS = 0). If PCSDEC = 0: PCS = xxx0NPCS[3:0] = 1110 PCS = xx01NPCS[3:0] = 1101 PCS = x011NPCS[3:0] = 1011 PCS = 0111NPCS[3:0] = 0111 PCS = 1111forbidden (no peripheral is selected) (x = don’t care) If PCSDEC = 1: NPCS[3:0] output signals = PCS. • LLB: Local Loopback Enable 1: Local loopback path enabled. LLB controls the local loopback on the data serializer for testing in master mode only (MISO is internally connected on MOSI). 0: Local loopback path disabled. • RXFIFOEN: FIFO in Reception Enable 1: The FIFO is used in reception (four characters can be stored in the SPI). 429 32099I–01/2012 AT32UC3L016/32/64 0: The FIFO is not used in reception (only one character can be stored in the SPI). • MODFDIS: Mode Fault Detection 1: Mode fault detection is disabled. If the I/O controller does not have open-drain capability, mode fault detection must be disabled for proper operation of the SPI. 0: Mode fault detection is enabled. • PCSDEC: Chip Select Decode 0: The chip selects are directly connected to a peripheral device. 1: The four chip select lines are connected to a 4- to 16-bit decoder. When PCSDEC equals one, up to 15 Chip Select signals can be generated with the four lines using an external 4- to 16-bit decoder. The CSRn registers define the characteristics of the 15 chip selects according to the following rules: CSR0 defines peripheral chip select signals 0 to 3. CSR1 defines peripheral chip select signals 4 to 7. CSR2 defines peripheral chip select signals 8 to 11. CSR3 defines peripheral chip select signals 12 to 14. • PS: Peripheral Select 1: Variable Peripheral Select. 0: Fixed Peripheral Select. • MSTR: Master/Slave Mode 1: SPI is in master mode. 0: SPI is in slave mode. 430 32099I–01/2012 AT32UC3L016/32/64 20.8.3 Name: Receive Data Register RDR Access Type: Read-only Offset: 0x08 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 RD[15:8] 7 6 5 4 RD[7:0] • RD: Receive Data Data received by the SPI Interface is stored in this register right-justified. Unused bits read zero. 431 32099I–01/2012 AT32UC3L016/32/64 20.8.4 Name: Transmit Data Register TDR Access Type: Write-only Offset: 0x0C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - LASTXFER 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 PCS 11 10 9 8 3 2 1 0 TD[15:8] 7 6 5 4 TD[7:0] • LASTXFER: Last Transfer 1: The current NPCS will be deasserted after the character written in TD has been transferred. When CSRn.CSAAT is one, this allows to close the communication with the current serial peripheral by raising the corresponding NPCS line as soon as TD transfer has completed. 0: Writing a zero to this bit has no effect. This field is only used if Variable Peripheral Select is active (MR.PS = 1). • PCS: Peripheral Chip Select If PCSDEC = 0: PCS = xxx0NPCS[3:0] = 1110 PCS = xx01NPCS[3:0] = 1101 PCS = x011NPCS[3:0] = 1011 PCS = 0111NPCS[3:0] = 0111 PCS = 1111forbidden (no peripheral is selected) (x = don’t care) If PCSDEC = 1: NPCS[3:0] output signals = PCS This field is only used if Variable Peripheral Select is active (MR.PS = 1). • TD: Transmit Data Data to be transmitted by the SPI Interface is stored in this register. Information to be transmitted must be written to the TDR register in a right-justified format. 432 32099I–01/2012 AT32UC3L016/32/64 20.8.5 Name: Status Register SR Access Type: Read-only Offset: 0x10 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - SPIENS 15 14 13 12 11 10 9 8 - - - - - UNDES TXEMPTY NSSR 7 6 5 4 3 2 1 0 - - - - OVRES MODF TDRE RDRF • SPIENS: SPI Enable Status 1: This bit is set when the SPI is enabled. 0: This bit is cleared when the SPI is disabled. • UNDES: Underrun Error Status (Slave Mode Only) 1: This bit is set when a transfer begins whereas no data has been loaded in the TDR register. 0: This bit is cleared when the SR register is read. • TXEMPTY: Transmission Registers Empty 1: This bit is set when TDR and internal shifter are empty. If a transfer delay has been defined, TXEMPTY is set after the completion of such delay. 0: This bit is cleared as soon as data is written in TDR. • NSSR: NSS Rising 1: A rising edge occurred on NSS pin since last read. 0: This bit is cleared when the SR register is read. • OVRES: Overrun Error Status 1: This bit is set when an overrun has occurred. An overrun occurs when RDR is loaded at least twice from the serializer since the last read of the RDR. 0: This bit is cleared when the SR register is read. • MODF: Mode Fault Error 1: This bit is set when a Mode Fault occurred. 0: This bit is cleared when the SR register is read. • TDRE: Transmit Data Register Empty 1: This bit is set when the last data written in the TDR register has been transferred to the serializer. 0: This bit is cleared when data has been written to TDR and not yet transferred to the serializer. TDRE equals zero when the SPI is disabled or at reset. The SPI enable command sets this bit to one. • RDRF: Receive Data Register Full 1: Data has been received and the received data has been transferred from the serializer to RDR since the last read of RDR. 0: No data has been received since the last read of RDR 433 32099I–01/2012 AT32UC3L016/32/64 20.8.6 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x14 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - UNDES TXEMPTY NSSR 7 6 5 4 3 2 1 0 - - - - OVRES MODF TDRE RDRF Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 434 32099I–01/2012 AT32UC3L016/32/64 20.8.7 Name: Interrupt Disable Register IDR Access Type: Write-only Offset: 0x18 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - UNDES TXEMPTY NSSR 7 6 5 4 3 2 1 0 - - - - OVRES MODF TDRE RDRF Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 435 32099I–01/2012 AT32UC3L016/32/64 20.8.8 Name: Interrupt Mask Register IMR Access Type: Read-only Offset: 0x1C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - UNDES TXEMPTY NSSR 7 6 5 4 3 2 1 0 - - - - OVRES MODF TDRE RDRF 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. A bit in this register is cleared when the corresponding bit in IDR is written to one. A bit in this register is set when the corresponding bit in IER is written to one. 436 32099I–01/2012 AT32UC3L016/32/64 20.8.9 Name: Chip Select Register 0 CSR0 Access Type: Read/Write Offset: 0x30 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 CSAAT CSNAAT NCPHA CPOL DLYBCT 23 22 21 20 DLYBS 15 14 13 12 SCBR 7 6 5 BITS 4 • DLYBCT: Delay Between Consecutive Transfers This field defines the delay between two consecutive transfers with the same peripheral without removing the chip select. The delay is always inserted after each transfer and before removing the chip select if needed. When DLYBCT equals zero, no delay between consecutive transfers is inserted and the clock keeps its duty cycle over the character transfers. Otherwise, the following equation determines the delay: 32 × DLYBCT Delay Between Consecutive Transfers = -----------------------------------CLKSPI • DLYBS: Delay Before SPCK This field defines the delay from NPCS valid to the first valid SPCK transition. When DLYBS equals zero, the NPCS valid to SPCK transition is 1/2 the SPCK clock period. Otherwise, the following equations determine the delay: DLYBS Delay Before SPCK = --------------------CLKSPI • SCBR: Serial Clock Baud Rate In Master Mode, the SPI Interface uses a modulus counter to derive the SPCK baud rate from the CLK_SPI. The Baud rate is selected by writing a value from 1 to 255 in the SCBR field. The following equations determine the SPCK baud rate: CLKSPI SPCK Baudrate = --------------------SCBR Writing the SCBR field to zero is forbidden. Triggering a transfer while SCBR is zero can lead to unpredictable results. At reset, SCBR is zero and the user has to write it to a valid value before performing the first transfer. If a clock divider (SCBRn) field is set to one and the other SCBR fields differ from one, access on CSn is correct but no correct access will be possible on other CS. 437 32099I–01/2012 AT32UC3L016/32/64 • BITS: Bits Per Transfer The BITS field determines the number of data bits transferred. Reserved values should not be used. BITS Bits Per Transfer 0000 8 0001 9 0010 10 0011 11 0100 12 0101 13 0110 14 0111 15 1000 16 1001 4 1010 5 1011 6 1100 7 1101 Reserved 1110 Reserved 1111 Reserved • CSAAT: Chip Select Active After Transfer 1: The Peripheral Chip Select does not rise after the last transfer is achieved. It remains active until a new transfer is requested on a different chip select. 0: The Peripheral Chip Select Line rises as soon as the last transfer is achieved. • CSNAAT: Chip Select Not Active After Transfer (Ignored if CSAAT = 1) 0: The Peripheral Chip Select does not rise between two transfers if the TDR is reloaded before the end of the first transfer and if the two transfers occur on the same Chip Select. 1: The Peripheral Chip Select rises systematically between each transfer performed on the same slave for a minimal duration of: DLYBCS ----------------------- (if DLYBCT field is different from 0) CLKSPI DLYBCS + 1 (if DLYBCT field equals 0) --------------------------------CLKSPI • NCPHA: Clock Phase 1: Data is captured after the leading (inactive-to-active) edge of SPCK and changed on the trailing (active-to-inactive) edge of SPCK. 0: Data is changed on the leading (inactive-to-active) edge of SPCK and captured after the trailing (active-to-inactive) edge of SPCK. NCPHA determines which edge of SPCK causes data to change and which edge causes data to be captured. NCPHA is used with CPOL to produce the required clock/data relationship between master and slave devices. • CPOL: Clock Polarity 1: The inactive state value of SPCK is logic level one. 0: The inactive state value of SPCK is logic level zero. 438 32099I–01/2012 AT32UC3L016/32/64 CPOL is used to determine the inactive state value of the serial clock (SPCK). It is used with NCPHA to produce the required clock/data relationship between master and slave devices. 439 32099I–01/2012 AT32UC3L016/32/64 20.8.10 Name: Chip Select Register 1 CSR1 Access Type: Read/Write Offset: 0x34 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 CSAAT CSNAAT NCPHA CPOL DLYBCT 23 22 21 20 DLYBS 15 14 13 12 SCBR 7 6 5 BITS 4 • DLYBCT: Delay Between Consecutive Transfers This field defines the delay between two consecutive transfers with the same peripheral without removing the chip select. The delay is always inserted after each transfer and before removing the chip select if needed. When DLYBCT equals zero, no delay between consecutive transfers is inserted and the clock keeps its duty cycle over the character transfers. Otherwise, the following equation determines the delay: 32 × DLYBCT Delay Between Consecutive Transfers = -----------------------------------CLKSPI • DLYBS: Delay Before SPCK This field defines the delay from NPCS valid to the first valid SPCK transition. When DLYBS equals zero, the NPCS valid to SPCK transition is 1/2 the SPCK clock period. Otherwise, the following equations determine the delay: DLYBS Delay Before SPCK = --------------------CLKSPI • SCBR: Serial Clock Baud Rate In Master Mode, the SPI Interface uses a modulus counter to derive the SPCK baud rate from the CLK_SPI. The Baud rate is selected by writing a value from 1 to 255 in the SCBR field. The following equations determine the SPCK baud rate: CLKSPI SPCK Baudrate = --------------------SCBR Writing the SCBR field to zero is forbidden. Triggering a transfer while SCBR is zero can lead to unpredictable results. At reset, SCBR is zero and the user has to write it to a valid value before performing the first transfer. If a clock divider (SCBRn) field is set to one and the other SCBR fields differ from one, access on CSn is correct but no correct access will be possible on other CS. 440 32099I–01/2012 AT32UC3L016/32/64 • BITS: Bits Per Transfer The BITS field determines the number of data bits transferred. Reserved values should not be used. BITS Bits Per Transfer 0000 8 0001 9 0010 10 0011 11 0100 12 0101 13 0110 14 0111 15 1000 16 1001 4 1010 5 1011 6 1100 7 1101 Reserved 1110 Reserved 1111 Reserved • CSAAT: Chip Select Active After Transfer 1: The Peripheral Chip Select does not rise after the last transfer is achieved. It remains active until a new transfer is requested on a different chip select. 0: The Peripheral Chip Select Line rises as soon as the last transfer is achieved. • CSNAAT: Chip Select Not Active After Transfer (Ignored if CSAAT = 1) 0: The Peripheral Chip Select does not rise between two transfers if the TDR is reloaded before the end of the first transfer and if the two transfers occur on the same Chip Select. 1: The Peripheral Chip Select rises systematically between each transfer performed on the same slave for a minimal duration of: DLYBCS ----------------------- (if DLYBCT field is different from 0) CLKSPI DLYBCS + 1 (if DLYBCT field equals 0) --------------------------------CLKSPI • NCPHA: Clock Phase 1: Data is captured after the leading (inactive-to-active) edge of SPCK and changed on the trailing (active-to-inactive) edge of SPCK. 0: Data is changed on the leading (inactive-to-active) edge of SPCK and captured after the trailing (active-to-inactive) edge of SPCK. NCPHA determines which edge of SPCK causes data to change and which edge causes data to be captured. NCPHA is used with CPOL to produce the required clock/data relationship between master and slave devices. • CPOL: Clock Polarity 1: The inactive state value of SPCK is logic level one. 0: The inactive state value of SPCK is logic level zero. 441 32099I–01/2012 AT32UC3L016/32/64 CPOL is used to determine the inactive state value of the serial clock (SPCK). It is used with NCPHA to produce the required clock/data relationship between master and slave devices. 442 32099I–01/2012 AT32UC3L016/32/64 20.8.11 Name: Chip Select Register 2 CSR2 Access Type: Read/Write Offset: 0x38 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 CSAAT CSNAAT NCPHA CPOL DLYBCT 23 22 21 20 DLYBS 15 14 13 12 SCBR 7 6 5 BITS 4 • DLYBCT: Delay Between Consecutive Transfers This field defines the delay between two consecutive transfers with the same peripheral without removing the chip select. The delay is always inserted after each transfer and before removing the chip select if needed. When DLYBCT equals zero, no delay between consecutive transfers is inserted and the clock keeps its duty cycle over the character transfers. Otherwise, the following equation determines the delay: 32 × DLYBCT Delay Between Consecutive Transfers = -----------------------------------CLKSPI • DLYBS: Delay Before SPCK This field defines the delay from NPCS valid to the first valid SPCK transition. When DLYBS equals zero, the NPCS valid to SPCK transition is 1/2 the SPCK clock period. Otherwise, the following equations determine the delay: DLYBS Delay Before SPCK = --------------------CLKSPI • SCBR: Serial Clock Baud Rate In Master Mode, the SPI Interface uses a modulus counter to derive the SPCK baud rate from the CLK_SPI. The Baud rate is selected by writing a value from 1 to 255 in the SCBR field. The following equations determine the SPCK baud rate: CLKSPI SPCK Baudrate = --------------------SCBR Writing the SCBR field to zero is forbidden. Triggering a transfer while SCBR is zero can lead to unpredictable results. At reset, SCBR is zero and the user has to write it to a valid value before performing the first transfer. If a clock divider (SCBRn) field is set to one and the other SCBR fields differ from one, access on CSn is correct but no correct access will be possible on other CS. 443 32099I–01/2012 AT32UC3L016/32/64 • BITS: Bits Per Transfer The BITS field determines the number of data bits transferred. Reserved values should not be used. BITS Bits Per Transfer 0000 8 0001 9 0010 10 0011 11 0100 12 0101 13 0110 14 0111 15 1000 16 1001 4 1010 5 1011 6 1100 7 1101 Reserved 1110 Reserved 1111 Reserved • CSAAT: Chip Select Active After Transfer 1: The Peripheral Chip Select does not rise after the last transfer is achieved. It remains active until a new transfer is requested on a different chip select. 0: The Peripheral Chip Select Line rises as soon as the last transfer is achieved. • CSNAAT: Chip Select Not Active After Transfer (Ignored if CSAAT = 1) 0: The Peripheral Chip Select does not rise between two transfers if the TDR is reloaded before the end of the first transfer and if the two transfers occur on the same Chip Select. 1: The Peripheral Chip Select rises systematically between each transfer performed on the same slave for a minimal duration of: DLYBCS ----------------------- (if DLYBCT field is different from 0) CLKSPI DLYBCS + 1 (if DLYBCT field equals 0) --------------------------------CLKSPI • NCPHA: Clock Phase 1: Data is captured after the leading (inactive-to-active) edge of SPCK and changed on the trailing (active-to-inactive) edge of SPCK. 0: Data is changed on the leading (inactive-to-active) edge of SPCK and captured after the trailing (active-to-inactive) edge of SPCK. NCPHA determines which edge of SPCK causes data to change and which edge causes data to be captured. NCPHA is used with CPOL to produce the required clock/data relationship between master and slave devices. • CPOL: Clock Polarity 1: The inactive state value of SPCK is logic level one. 0: The inactive state value of SPCK is logic level zero. 444 32099I–01/2012 AT32UC3L016/32/64 CPOL is used to determine the inactive state value of the serial clock (SPCK). It is used with NCPHA to produce the required clock/data relationship between master and slave devices. 445 32099I–01/2012 AT32UC3L016/32/64 20.8.12 Name: Chip Select Register 3 CSR3 Access Type: Read/Write Offset: 0x3C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 CSAAT CSNAAT NCPHA CPOL DLYBCT 23 22 21 20 DLYBS 15 14 13 12 SCBR 7 6 5 BITS 4 • DLYBCT: Delay Between Consecutive Transfers This field defines the delay between two consecutive transfers with the same peripheral without removing the chip select. The delay is always inserted after each transfer and before removing the chip select if needed. When DLYBCT equals zero, no delay between consecutive transfers is inserted and the clock keeps its duty cycle over the character transfers. Otherwise, the following equation determines the delay: 32 × DLYBCT Delay Between Consecutive Transfers = -----------------------------------CLKSPI • DLYBS: Delay Before SPCK This field defines the delay from NPCS valid to the first valid SPCK transition. When DLYBS equals zero, the NPCS valid to SPCK transition is 1/2 the SPCK clock period. Otherwise, the following equations determine the delay: DLYBS Delay Before SPCK = --------------------CLKSPI • SCBR: Serial Clock Baud Rate In Master Mode, the SPI Interface uses a modulus counter to derive the SPCK baud rate from the CLK_SPI. The Baud rate is selected by writing a value from 1 to 255 in the SCBR field. The following equations determine the SPCK baud rate: CLKSPI SPCK Baudrate = --------------------SCBR Writing the SCBR field to zero is forbidden. Triggering a transfer while SCBR is zero can lead to unpredictable results. At reset, SCBR is zero and the user has to write it to a valid value before performing the first transfer. If a clock divider (SCBRn) field is set to one and the other SCBR fields differ from one, access on CSn is correct but no correct access will be possible on other CS. 446 32099I–01/2012 AT32UC3L016/32/64 • BITS: Bits Per Transfer The BITS field determines the number of data bits transferred. Reserved values should not be used. BITS Bits Per Transfer 0000 8 0001 9 0010 10 0011 11 0100 12 0101 13 0110 14 0111 15 1000 16 1001 4 1010 5 1011 6 1100 7 1101 Reserved 1110 Reserved 1111 Reserved • CSAAT: Chip Select Active After Transfer 1: The Peripheral Chip Select does not rise after the last transfer is achieved. It remains active until a new transfer is requested on a different chip select. 0: The Peripheral Chip Select Line rises as soon as the last transfer is achieved. • CSNAAT: Chip Select Not Active After Transfer (Ignored if CSAAT = 1) 0: The Peripheral Chip Select does not rise between two transfers if the TDR is reloaded before the end of the first transfer and if the two transfers occur on the same Chip Select. 1: The Peripheral Chip Select rises systematically between each transfer performed on the same slave for a minimal duration of: DLYBCS ----------------------- (if DLYBCT field is different from 0) CLKSPI DLYBCS + 1 (if DLYBCT field equals 0) --------------------------------CLKSPI • NCPHA: Clock Phase 1: Data is captured after the leading (inactive-to-active) edge of SPCK and changed on the trailing (active-to-inactive) edge of SPCK. 0: Data is changed on the leading (inactive-to-active) edge of SPCK and captured after the trailing (active-to-inactive) edge of SPCK. NCPHA determines which edge of SPCK causes data to change and which edge causes data to be captured. NCPHA is used with CPOL to produce the required clock/data relationship between master and slave devices. • CPOL: Clock Polarity 1: The inactive state value of SPCK is logic level one. 0: The inactive state value of SPCK is logic level zero. 447 32099I–01/2012 AT32UC3L016/32/64 CPOL is used to determine the inactive state value of the serial clock (SPCK). It is used with NCPHA to produce the required clock/data relationship between master and slave devices. 448 32099I–01/2012 AT32UC3L016/32/64 20.8.13 Write Protection Control Register Register Name: WPCR Access Type: Read-write Offset: 0xE4 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 18 17 16 10 9 8 SPIWPKEY[23:16] 23 22 21 20 19 SPIWPKEY[15:8] 15 14 13 12 11 SPIWPKEY[7:0] 7 6 5 4 3 2 1 0 - - - - - - - SPIWPEN • SPIWPKEY: SPI Write Protection Key Password If a value is written in SPIWPEN, the value is taken into account only if SPIWPKEY is written with “SPI” (SPI written in ASCII Code, i.e. 0x535049 in hexadecimal). • SPIWPEN: SPI Write Protection Enable 1: The Write Protection is Enabled 0: The Write Protection is Disabled 449 32099I–01/2012 AT32UC3L016/32/64 20.8.14 Write Protection Status Register Register Name: WPSR Access Type: Read-only Offset: 0xE8 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 2 1 0 SPIWPVSRC 7 6 5 4 3 - - - - - SPIWPVS • SPIWPVSRC: SPI Write Protection Violation Source This Field indicates the Peripheral Bus Offset of the register concerned by the violation (MR or CSRx) • SPIWPVS: SPI Write Protection Violation Status SPIWPVS value Violation Type 1 The Write Protection has blocked a Write access to a protected register (since the last read). 2 Software Reset has been performed while Write Protection was enabled (since the last read or since the last write access on MR, IER, IDR or CSRx). 3 Both Write Protection violation and software reset with Write Protection enabled have occurred since the last read. 4 Write accesses have been detected on MR (while a chip select was active) or on CSRi (while the Chip Select “i” was active) since the last read. 5 The Write Protection has blocked a Write access to a protected register and write accesses have been detected on MR (while a chip select was active) or on CSRi (while the Chip Select “i” was active) since the last read. 6 Software Reset has been performed while Write Protection was enabled (since the last read or since the last write access on MR, IER, IDR or CSRx) and some write accesses have been detected on MR (while a chip select was active) or on CSRi (while the Chip Select “i” was active) since the last read. 7 - The Write Protection has blocked a Write access to a protected register. and - Software Reset has been performed while Write Protection was enabled. and - Write accesses have been detected on MR (while a chip select was active) or on CSRi (while the Chip Select “i” was active) since the last read. 450 32099I–01/2012 AT32UC3L016/32/64 20.8.15 Features Register Register Name: FEATURES Access Type: Read-only Offset: 0xF8 Reset Value: – 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - SWIMPL FIFORIMPL BRPBHSB CSNAATIMPL EXTDEC 15 14 13 12 11 10 9 8 LENNCONF 7 6 5 4 PHZNCONF PHCONF PPNCONF PCONF LENCONF 3 2 1 0 NCS • SWIMPL: Spurious Write Protection Implemented • • • • • • • • 0: Spurious write protection is not implemented. 1: Spurious write protection is implemented. FIFORIMPL: FIFO in Reception Implemented 0: FIFO in reception is not implemented. 1: FIFO in reception is implemented. BRPBHSB: Bridge Type is PB to HSB 0: Bridge type is not PB to HSB. 1: Bridge type is PB to HSB. CSNAATIMPL: CSNAAT Features Implemented 0: CSNAAT (Chip select not active after transfer) features are not implemented. 1: CSNAAT features are implemented. EXTDEC: External Decoder True 0: External decoder capability is not implemented. 1: External decoder capability is implemented. LENNCONF: Character Length if not Configurable If the character length is not configurable, this field specifies the fixed character length. LENCONF: Character Length Configurable 0: The character length is not configurable. 1: The character length is configurable. PHZNCONF: Phase is Zero if Phase not Configurable 0: If phase is not configurable, phase is non-zero. 1: If phase is not configurable, phase is zero. PHCONF: Phase Configurable 0: Phase is not configurable. 1: Phase is configurable. 451 32099I–01/2012 AT32UC3L016/32/64 • PPNCONF: Polarity Positive if Polarity not Configurable 0: If polarity is not configurable, polarity is negative. 1: If polarity is not configurable, polarity is positive. • PCONF: Polarity Configurable 0: Polarity is not configurable. 1: Polarity is configurable. • NCS: Number of Chip Selects This field indicates the number of chip selects implemented. 452 32099I–01/2012 AT32UC3L016/32/64 20.8.16 Version Register Register Name: VERSION Access Type: Read-only Offset: 0xFC Reset Value: – 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 MFN 11 10 VERSION[11:8] 7 6 5 4 3 2 1 0 VERSION[7:0] • MFN Reserved. No functionality associated. • VERSION Version number of the module. No functionality associated. 453 32099I–01/2012 AT32UC3L016/32/64 20.9 Module Configuration The specific configuration for each SPI instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 20-4. SPI Clock Name Module Name Clock Name Description SPI CLK_SPI Clock for the SPI bus interface Table 20-5. Register Reset Value FEATURES 0x001F0154 VERSION 0x00000211 454 32099I–01/2012 AT32UC3L016/32/64 21. Two-wire Master Interface (TWIM) Rev.: 1.0.1.1 21.1 Features • Compatible with I²C standard • • • • • • 21.2 – Multi-master support – Transfer speeds of 100 and 400 kbit/s – 7- and 10-bit and General Call addressing Compatible with SMBus standard – Hardware Packet Error Checking (CRC) generation and verification with ACK control – SMBus ALERT interface – 25 ms clock low timeout delay – 10 ms master cumulative clock low extend time – 25 ms slave cumulative clock low extend time Compatible with PMBus Compatible with Atmel Two-wire Interface Serial Memories DMA interface for reducing CPU load Arbitrary transfer lengths, including 0 data bytes Optional clock stretching if transmit or receive buffers not ready for data transfer Overview The Atmel Two-wire Master Interface (TWIM) interconnects components on a unique two-wire bus, made up of one clock line and one data line with speeds of up to 400 kbit/s, based on a byte-oriented transfer format. It can be used with any Atmel Two-wire Interface bus serial EEPROM and I²C compatible device such as a real time clock (RTC), dot matrix/graphic LCD controller, and temperature sensor, to name a few. The TWIM is always a bus master and can transfer sequential or single bytes. Multiple master capability is supported. Arbitration of the bus is performed internally and relinquishes the bus automatically if the bus arbitration is lost. A configurable baud rate generator permits the output data rate to be adapted to a wide range of core clock frequencies.Table 21-1 lists the compatibility level of the Atmel Two-wire Interface in Master Mode and a full I²C compatible device. Table 21-1. Atmel TWIM Compatibility with I²C Standard I²C Standard Atmel TWIM Standard-mode (100 kbit/s) Supported Fast-mode (400 kbit/s) Supported Fast-mode Plus (1 Mbit/s) Supported 7- or 10-bits Slave Addressing Supported (1) START BYTE Not Supported Repeated Start (Sr) Condition Supported ACK and NACK Management Supported Slope Control and Input Filtering (Fast mode) Supported Clock Stretching Supported Note: 1. START + b000000001 + Ack + Sr 455 32099I–01/2012 AT32UC3L016/32/64 Table 21-2 lists the compatibility level of the Atmel Two-wire Master Interface and a full SMBus compatible master. Table 21-2. 21.3 SMBus Standard Atmel TWIM Bus Timeouts Supported Address Resolution Protocol Supported Alert Supported Host Functionality Supported Packet Error Checking Supported List of Abbreviations Table 21-3. 21.4 Atmel TWIM Compatibility with SMBus Standard Abbreviations Abbreviation Description TWI Two-wire Interface A Acknowledge NA Non Acknowledge P Stop S Start Sr Repeated Start SADR Slave Address ADR Any address except SADR R Read W Write Block Diagram Figure 21-1. Block Diagram Peripheral Bus Bridge TWCK I/O Controller Two-wire Interface Power Manager TWD TWALM CLK_TWIM INTC TWI Interrupt 456 32099I–01/2012 AT32UC3L016/32/64 21.5 Application Block Diagram Figure 21-2. Application Block Diagram VDD Rp Rp Rp TWD TWI Master TWCK TWALM Atmel TWI serial EEPROM I2C RTC I2C LCD controller I2C temp sensor Slave 1 Slave 2 Slave 3 Slave 4 Rp: pull-up value as given by the I2C Standard 21.6 I/O Lines Description Table 21-4. I/O Lines Description Pin Name Pin Description TWD Two-wire Serial Data Input/Output TWCK Two-wire Serial Clock Input/Output TWALM SMBus SMBALERT Input/Output 21.7 Type Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 21.7.1 I/O Lines TWD and TWCK are bidirectional lines, connected to a positive supply voltage via a current source or pull-up resistor (see Figure 21-4 on page 459). When the bus is free, both lines are high. The output stages of devices connected to the bus must have an open-drain or open-collector to perform the wired-AND function. TWALM is used to implement the optional SMBus SMBALERT signal. The TWALM, TWD, and TWCK pins may be multiplexed with I/O Controller lines. To enable the TWIM, the user must perform the following steps: • Program the I/O Controller to: – Dedicate TWD, TWCK, and optionally TWALM as peripheral lines. – Define TWD, TWCK, and optionally TWALM as open-drain. 21.7.2 Power Management If the CPU enters a sleep mode that disables clocks used by the TWIM, the TWIM will stop functioning and resume operation after the system wakes up from sleep mode. 457 32099I–01/2012 AT32UC3L016/32/64 21.7.3 Clocks The clock for the TWIM bus interface (CLK_TWIM) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the TWIM before disabling the clock, to avoid freezing the TWIM in an undefined state. 21.7.4 DMA The TWIM DMA handshake interface is connected to the Peripheral DMA Controller. Using the TWIM DMA functionality requires the Peripheral DMA Controller to be programmed after setting up the TWIM. 21.7.5 Interrupts The TWIM interrupt request lines are connected to the interrupt controller. Using the TWIM interrupts requires the interrupt controller to be programmed first. 21.7.6 Debug Operation When an external debugger forces the CPU into debug mode, the TWIM continues normal operation. If the TWIM is configured in a way that requires it to be periodically serviced by the CPU through interrupts or similar, improper operation or data loss may result during debugging. 458 32099I–01/2012 AT32UC3L016/32/64 21.8 21.8.1 Functional Description Transfer Format The data put on the TWD line must be 8 bits long. Data is transferred MSB first; each byte must be followed by an acknowledgement. The number of bytes per transfer is unlimited (see Figure 21-4). Each transfer begins with a START condition and terminates with a STOP condition (see Figure 21-4). • A high-to-low transition on the TWD line while TWCK is high defines the START condition. • A low-to-high transition on the TWD line while TWCK is high defines a STOP condition. Figure 21-3. START and STOP Conditions TWD TWCK Start Stop Figure 21-4. Transfer Format TWD TWCK Start 21.8.2 Address R/W Ack Data Ack Data Ack Stop Operation The TWIM has two modes of operation: • Master transmitter mode • Master receiver mode The master is the device which starts and stops a transfer and generates the TWCK clock. These modes are described in the following chapters. 459 32099I–01/2012 AT32UC3L016/32/64 21.8.2.1 Clock Generation The Clock Waveform Generator Register (CWGR) is used to control the waveform of the TWCK clock. CWGR must be written so that the desired TWI bus timings are generated. CWGR describes bus timings as a function of cycles of a prescaled clock. The clock prescaling can be selected through the Clock Prescaler field in CWGR (CWGR.EXP). f CLK_TWIM f PRESCALER = ------------------------( EXP + 1 ) 2 CWGR has the following fields: LOW: Prescaled clock cycles in clock low count. Used to time TLOW and TBUF. HIGH: Prescaled clock cycles in clock high count. Used to time THIGH. STASTO: Prescaled clock cycles in clock high count. Used to time THD_STA, TSU_STA, TSU_STO. DATA: Prescaled clock cycles for data setup and hold count. Used to time THD_DAT, TSU_DAT. EXP: Specifies the clock prescaler setting. Note that the total clock low time generated is the sum of THD_DAT + TSU_DAT + TLOW. Any slave or other bus master taking part in the transfer may extend the TWCK low period at any time. The TWIM hardware monitors the state of the TWCK line as required by the I²C specification. The clock generation counters are started when a high/low level is detected on the TWCK line, not when the TWIM hardware releases/drives the TWCK line. This means that the CWGR settings alone do not determine the TWCK frequency. The CWGR settings determine the clock low time and the clock high time, but the TWCK rise and fall times are determined by the external circuitry (capacitive load, etc.). Figure 21-5. Bus Timing Diagram t HIGH t LOW S t HD:STA t LOW t SU:DAT t HD:DAT t t SU:DAT t SU:STA SU:STO P Sr 460 32099I–01/2012 AT32UC3L016/32/64 21.8.2.2 Setting up and Performing a Transfer Operation of the TWIM is mainly controlled by the Control Register (CR) and the Command Register (CMDR). TWIM status is provided in the Status Register (SR). The following list presents the main steps in a typical communication: 1. Before any transfers can be performed, bus timings must be configured by writing to the Clock Waveform Generator Register (CWGR). If operating in SMBus mode, the SMBus Timing Register (SMBTR) register must also be configured. 2. If the Peripheral DMA Controller is to be used for the transfers, it must be set up. 3. CMDR or NCMDR must be written with a value describing the transfer to be performed. The interrupt system can be set up to give interrupt requests on specific events or error conditions in the SR, for example when the transfer is complete or if arbitration is lost. The Interrupt Enable Register (IER) and Interrupt Disable Register (IDR) can be written to specify which bits in the SR will generate interrupt requests. The SR.BUSFREE bit is set when activity is completed on the two-wire bus. The SR.CRDY bit is set when CMDR and/or NCMDR is ready to receive one or more commands. The controller will refuse to start a new transfer while ANAK, DNAK, or ARBLST in the Status Register (SR) is one. This is necessary to avoid a race when the software issues a continuation of the current transfer at the same time as one of these errors happen. Also, if ANAK or DNAK occurs, a STOP condition is sent automatically. The user will have to restart the transmission by clearing the error bits in SR after resolving the cause for the NACK. After a data or address NACK from the slave, a STOP will be transmitted automatically. Note that the VALID bit in CMDR is NOT cleared in this case. If this transfer is to be discarded, the VALID bit can be cleared manually allowing any command in NCMDR to be copied into CMDR. When a data or address NACK is returned by the slave while the master is transmitting, it is possible that new data has already been written to the THR register. This data will be transferred out as the first data byte of the next transfer. If this behavior is to be avoided, the safest approach is to perform a software reset of the TWIM. 21.8.3 Master Transmitter Mode A START condition is transmitted and master transmitter mode is initiated when the bus is free and CMDR has been written with START=1 and READ=0. START and SADR+W will then be transmitted. During the address acknowledge clock pulse (9th pulse), the master releases the data line (HIGH), enabling the slave to pull it down in order to acknowledge the address. The master polls the data line during this clock pulse and sets the Address Not Acknowledged bit (ANAK) in the Status Register if no slave acknowledges the address. After the address phase, the following is repeated: while (NBYTES>0) 1. Wait until THR contains a valid data byte, stretching low period of TWCK. SR.TXRDY indicates the state of THR. Software or the Peripheral DMA Controller must write the data byte to THR. 2. Transmit this data byte 3. Decrement NBYTES 4. If (NBYTES==0) and STOP=1, transmit STOP condition Writing CMDR with START=STOP=1 and NBYTES=0 will generate a transmission with no data bytes, ie START, SADR+W, STOP. 461 32099I–01/2012 AT32UC3L016/32/64 TWI transfers require the slave to acknowledge each received data byte. During the acknowledge clock pulse (9th pulse), the master releases the data line (HIGH), enabling the slave to pull it down in order to generate the acknowledge. The master polls the data line during this clock pulse and sets the Data Acknowledge bit (DNACK) in the Status Register if the slave does not acknowledge the data byte. As with the other status bits, an interrupt can be generated if enabled in the Interrupt Enable Register (IER). TXRDY is used as Transmit Ready for the Peripheral DMA Controller transmit channel. The end of a command is marked when the TWIM sets the SR.CCOMP bit. See Figure 21-6 and Figure 21-7. Figure 21-6. Master Write with One Data Byte TWD S DADR W A DATA A P SR.IDLE TXRDY Write THR (DATA) NBYTES set to 1 STOP sent automatically (ACK received and NBYTES=0) Figure 21-7. Master Write with Multiple Data Bytes TWD S DADR W A DATAn A DATAn+5 A DATAn+m A P SR.IDLE TXRDY Write THR (DATAn) NBYTES set to n 21.8.4 Write THR (DATAn+1) Write THR (DATAn+m) Last data sent STOP sent automatically (ACK received and NBYTES=0) Master Receiver Mode A START condition is transmitted and master receiver mode is initiated when the bus is free and CMDR has been written with START=1 and READ=1. START and SADR+R will then be transmitted. During the address acknowledge clock pulse (9th pulse), the master releases the data line (HIGH), enabling the slave to pull it down in order to acknowledge the address. The master polls the data line during this clock pulse and sets the Address Not Acknowledged bit (ANAK) in the Status Register if no slave acknowledges the address. After the address phase, the following is repeated: while (NBYTES>0) 462 32099I–01/2012 AT32UC3L016/32/64 1. Wait until RHR is empty, stretching low period of TWCK. SR.RXRDY indicates the state of RHR. Software or the Peripheral DMA Controller must read any data byte present in RHR. 2. Release TWCK generating a clock that the slave uses to transmit a data byte. 3. Place the received data byte in RHR, set RXRDY. 4. If NBYTES=0, generate a NAK after the data byte, otherwise generate an ACK. 5. Decrement NBYTES 6. If (NBYTES==0) and STOP=1, transmit STOP condition. Writing CMDR with START=STOP=1 and NBYTES=0 will generate a transmission with no data bytes, ie START, DADR+R, STOP The TWI transfers require the master to acknowledge each received data byte. During the acknowledge clock pulse (9th pulse), the slave releases the data line (HIGH), enabling the master to pull it down in order to generate the acknowledge. All data bytes except the last are acknowledged by the master. Not acknowledging the last byte informs the slave that the transfer is finished. RXRDY is used as Receive Ready for the Peripheral DMA Controller receive channel. Figure 21-8. Master Read with One Data Byte TWD S DADR R A DATA N P SR.IDLE RXRDY Write START & STOP bit NBYTES set to 1 Read RHR Figure 21-9. Master Read with Multiple Data Bytes TWD S DADR R A DATAn A DATAn+1 DATAn+m-1 A DATAn+m N P SR.IDLE RXRDY Write START + STOP bit NBYTES set to m Read RHR DATAn Read RHR DATAn+m-2 Read RHR DATAn+m-1 Read RHR DATAn+m Send STOP When NBYTES=0 463 32099I–01/2012 AT32UC3L016/32/64 21.8.5 Using the Peripheral DMA Controller The use of the Peripheral DMA Controller significantly reduces the CPU load. The user can set up ring buffers for the Peripheral DMA Controller, containing data to transmit or free buffer space to place received data. To assure correct behavior, respect the following programming sequences: 21.8.5.1 Data Transmit with the Peripheral DMA Controller 1. Initialize the transmit Peripheral DMA Controller (memory pointers, size, etc.). 2. Configure the TWIM (ADR, NBYTES, etc.). 3. Start the transfer by enabling the Peripheral DMA Controller to transmit. 4. Wait for the Peripheral DMA Controller end-of-transmit flag. 5. Disable the Peripheral DMA Controller. 21.8.5.2 Data Receive with the Peripheral DMA Controller 1. Initialize the receive Peripheral DMA Controller (memory pointers, size, etc.). 2. Configure the TWIM (ADR, NBYTES, etc.). 3. Start the transfer by enabling the Peripheral DMA Controller to receive. 4. Wait for the Peripheral DMA Controller end-of-receive flag. 5. Disable the Peripheral DMA Controller. 21.8.6 Multi-master Mode More than one master may access the bus at the same time without data corruption by using arbitration. Arbitration starts as soon as two or more masters place information on the bus at the same time, and stops (arbitration is lost) for the master that intends to send a logical one while the other master sends a logical zero. As soon as arbitration is lost by a master, it stops sending data and listens to the bus in order to detect a STOP. The SR.ARBLST flag will be set. When the STOP is detected, the master who lost arbitration may reinitiate the data transfer. Arbitration is illustrated in Figure 21-11. If the user starts a transfer and if the bus is busy, the TWIM automatically waits for a STOP condition on the bus before initiating the transfer (see Figure 21-10). Note: The state of the bus (busy or free) is not indicated in the user interface. 464 32099I–01/2012 AT32UC3L016/32/64 Figure 21-10. User Sends Data While the Bus is Busy TWCK START sent by the TWI STOP sent by the master DATA sent by a master TWD DATA sent by the TWI Bus is busy Bus is free Transfer is kept TWI DATA transfer A transfer is programmed (DADR + W + START + Write THR) Bus is considered as free Transfer is initiated Figure 21-11. Arbitration Cases TWCK TWD TWCK Data from a Master S 1 0 0 1 1 Data from TWI S 1 0 TWD S 1 0 0 1 P Arbitration is lost TWI stops sending data 1 1 Data from the master P Arbitration is lost S 1 0 S 1 0 0 1 1 S 1 0 1 1 The master stops sending data 0 1 Data from the TWI ARBLST Bus is busy Transfer is kept TWI DATA transfer A transfer is programmed (DADR + W + START + Write THR) 21.8.7 Bus is free Transfer is stopped Transfer is programmed again (DADR + W + START + Write THR) Bus is considered as free Transfer is initiated Combined Transfers CMDR and NCMDR may be used to generate longer sequences of connected transfers, since generation of START and/or STOP conditions is programmable on a per-command basis. Writing NCMDR with START=1 when the previous transfer was written with STOP=0 will cause a REPEATED START on the bus. The ability to generate such connected transfers allows arbitrary transfer lengths, since it is legal to write CMDR with both START=0 and STOP=0. If this is done in master receiver mode, the CMDR.ACKLAST bit must also be controlled. 465 32099I–01/2012 AT32UC3L016/32/64 As for single data transfers, the TXRDY and RXRDY bits in the Status Register indicates when data to transmit can be written to THR, or when received data can be read from RHR. Transfer of data to THR and from RHR can also be done automatically by DMA, see Section 21.8.5 21.8.7.1 Write Followed by Write Consider the following transfer: START, DADR+W, DATA+A, DATA+A, REPSTART, DADR+W, DATA+A, DATA+A, STOP. To generate this transfer: 1. Write CMDR with START=1, STOP=0, DADR, NBYTES=2 and READ=0. 2. Write NCMDR with START=1, STOP=1, DADR, NBYTES=2 and READ=0. 3. Wait until SR.TXRDY==1, then write first data byte to transfer to THR. 4. Wait until SR.TXRDY==1, then write second data byte to transfer to THR. 5. Wait until SR.TXRDY==1, then write third data byte to transfer to THR. 6. Wait until SR.TXRDY==1, then write fourth data byte to transfer to THR. 21.8.7.2 Read Followed by Read Consider the following transfer: START, DADR+R, DATA+A, DATA+NA, REPSTART, DADR+R, DATA+A, DATA+NA, STOP. To generate this transfer: 1. Write CMDR with START=1, STOP=0, DADR, NBYTES=2 and READ=1. 2. Write NCMDR with START=1, STOP=1, DADR, NBYTES=2 and READ=1. 3. Wait until SR.RXRDY==1, then read first data byte received from RHR. 4. Wait until SR.RXRDY==1, then read second data byte received from RHR. 5. Wait until SR.RXRDY==1, then read third data byte received from RHR. 6. Wait until SR.RXRDY==1, then read fourth data byte received from RHR. If combining several transfers, without any STOP or REPEATED START between them, remember to write a one to the ACKLAST bit in CMDR to keep from ending each of the partial transfers with a NACK. 21.8.7.3 Write Followed by Read Consider the following transfer: START, DADR+W, DATA+A, DATA+A, REPSTART, DADR+R, DATA+A, DATA+NA, STOP. 466 32099I–01/2012 AT32UC3L016/32/64 Figure 21-12. Combining a Write and Read Transfer THR DATA0 DATA1 RHR TWD DATA2 S DADR W A DATA0 A DATA1 NA Sr R DADR A DATA2 A DATA3 DATA3 A P SR.IDLE 1 TXRDY RXRDY To generate this transfer: 1. Write CMDR with START=1, STOP=0, DADR, NBYTES=2 and READ=0. 2. Write NCMDR with START=1, STOP=1, DADR, NBYTES=2 and READ=1. 3. Wait until SR.TXRDY==1, then write first data byte to transfer to THR. 4. Wait until SR.TXRDY==1, then write second data byte to transfer to THR. 5. Wait until SR.RXRDY==1, then read first data byte received from RHR. 6. Wait until SR.RXRDY==1, then read second data byte received from RHR. 21.8.7.4 Read Followed by Write Consider the following transfer: START, DADR+R, DATA+A, DATA+NA, REPSTART, DADR+W, DATA+A, DATA+A, STOP. Figure 21-13. Combining a Read and Write Transfer THR DATA2 RHR TWD DATA0 S SADR R A DATA0 A DATA3 DATA1 A DATA3 1 Sr DADR W A DATA2 A DATA3 SR.IDLE TXRDY RXRDY NA P 2 Read TWI_RHR To generate this transfer: 1. Write CMDR with START=1, STOP=0, DADR, NBYTES=2 and READ=1. 2. Write NCMDR with START=1, STOP=1, DADR, NBYTES=2 and READ=0. 3. Wait until SR.RXRDY==1, then read first data byte received from RHR. 4. Wait until SR.RXRDY==1, then read second data byte received from RHR. 5. Wait until SR.TXRDY==1, then write first data byte to transfer to THR. 6. Wait until SR.TXRDY==1, then write second data byte to transfer to THR. 467 32099I–01/2012 AT32UC3L016/32/64 21.8.8 Ten Bit Addressing Writing a one to CMDR.TENBIT enables 10-bit addressing in hardware. Performing transfers with 10-bit addressing is similar to transfers with 7-bit addresses, except that bits 9:7 of CMDR.SADR must be written appropriately. In Figure 21-14 and Figure 21-15, the grey boxes represent signals driven by the master, the white boxes are driven by the slave. 21.8.8.1 Master Transmitter To perform a master transmitter transfer: 1. Write CMDR with TENBIT=1, REPSAME=0, READ=0, START=1, STOP=1 and the desired address and NBYTES value. Figure 21-14. A Write Transfer with 10-bit Addressing 1 1 S 21.8.8.2 1 1 0 X SLAVE ADDRESS 1st 7 bits X 0 RW A1 SLAVE ADDRESS 2nd byte A2 DATA A DATA AA P Master Receiver When using master receiver mode with 10-bit addressing, CMDR.REPSAME must also be controlled. CMDR.REPSAME must be written to one when the address phase of the transfer should consist of only 1 address byte (the 11110xx byte) and not 2 address bytes. The I²C standard specifies that such addressing is required when addressing a slave for reads using 10-bit addressing. To perform a master receiver transfer: 1. Write CMDR with TENBIT=1, REPSAME=0, READ=0, START=1, STOP=0, NBYTES=0 and the desired address. 2. Write NCMDR with TENBIT=1, REPSAME=1, READ=1, START=1, STOP=1 and the desired address and NBYTES value. Figure 21-15. A Read Transfer with 10-bit Addressing 1 S 21.8.9 1 1 1 0 X X SLAVE ADDRESS 1st 7 bits 1 0 RW A1 SLAVE ADDRESS 2nd byte A2 Sr 1 1 1 0 X SLAVE ADDRESS 1st 7 bits X 1 RW A3 DATA A DATA A P SMBus Mode SMBus mode is enabled and disabled by writing to the SMEN and SMDIS bits in CR. SMBus mode operation is similar to I²C operation with the following exceptions: • Only 7-bit addressing can be used. • The SMBus standard describes a set of timeout values to ensure progress and throughput on the bus. These timeout values must be written into SMBTR. • Transmissions can optionally include a CRC byte, called Packet Error Check (PEC). • A dedicated bus line, SMBALERT, allows a slave to get a master’s attention. • A set of addresses have been reserved for protocol handling, such as Alert Response Address (ARA) and Host Header (HH) Address. 468 32099I–01/2012 AT32UC3L016/32/64 21.8.9.1 Packet Error Checking Each SMBus transfer can optionally end with a CRC byte, called the PEC byte. Writing a one to CMDR.PECEN enables automatic PEC handling in the current transfer. Transfers with and without PEC can freely be intermixed in the same system, since some slaves may not support PEC. The PEC LFSR is always updated on every bit transmitted or received, so that PEC handling on combined transfers will be correct. In master transmitter mode, the master calculates a PEC value and transmits it to the slave after all data bytes have been transmitted. Upon reception of this PEC byte, the slave will compare it to the PEC value it has computed itself. If the values match, the data was received correctly, and the slave will return an ACK to the master. If the PEC values differ, data was corrupted, and the slave will return a NACK value. The DNAK bit in SR reflects the state of the last received ACK/NACK value. Some slaves may not be able to check the received PEC in time to return a NACK if an error occurred. In this case, the slave should always return an ACK after the PEC byte, and some other mechanism must be implemented to verify that the transmission was received correctly. In master receiver mode, the slave calculates a PEC value and transmits it to the master after all data bytes have been transmitted. Upon reception of this PEC byte, the master will compare it to the PEC value it has computed itself. If the values match, the data was received correctly. If the PEC values differ, data was corrupted, and SR.PECERR is set. In master receiver mode, the PEC byte is always followed by a NACK transmitted by the master, since it is the last byte in the transfer. The PEC byte is automatically inserted in a master transmitter transmission if PEC is enabled when NBYTES reaches zero. The PEC byte is identified in a master receiver transmission if PEC is enabled when NBYTES reaches zero. NBYTES must therefore be written with the total number of data bytes in the transmission, including the PEC byte. In combined transfers, the PECEN bit should only be written to one in the last of the combined transfers. Consider the following transfer: S, ADR+W, COMMAND_BYTE, ACK, SR, ADR+R, DATA_BYTE, ACK, PEC_BYTE, NACK, P This transfer is generated by writing two commands to the command registers. The first command is a write with NBYTES=1 and PECEN=0, and the second is a read with NBYTES=2 and PECEN=1. Writing a one to the STOP bit in CR will place a STOP condition on the bus after the current byte. No PEC byte will be sent in this case. 21.8.9.2 Timeouts The TLOWS and TLOWM fields in SMBTR configure the SMBus timeout values. If a timeout occurs, the master will transmit a STOP condition and leave the bus. The SR.TOUT bit is set. 21.8.9.3 SMBus ALERT Signal A slave can get the master’s attention by pulling the TWALM line low. The TWIM will then set the SR.SMBALERT bit. This can be set up to trigger an interrupt, and software can then take the appropriate action, as defined in the SMBus standard. 469 32099I–01/2012 AT32UC3L016/32/64 21.8.10 Identifying Bus Events This chapter lists the different bus events, and how they affect bits in the TWIM registers. This is intended to help writing drivers for the TWIM. Table 21-5. Bus Events Event Effect Master transmitter has sent a data byte SR.THR is cleared. Master receiver has received a data byte SR.RHR is set. Start+Sadr sent, no ack received from slave SR.ANAK is set. SR.CCOMP not set. CMDR.VALID remains set. STOP automatically transmitted on bus. Data byte sent to slave, no ack received from slave SR.DNAK is set. SR.CCOMP not set. CMDR.VALID remains set. STOP automatically transmitted on bus. Arbitration lost SR.ARBLST is set. SR.CCOMP not set. CMDR.VALID remains set. TWCK and TWD immediately released to a pulled-up state. SMBus Alert received SR.SMBALERT is set. SMBus timeout received SR.SMBTOUT is set. SR.CCOMP not set. CMDR.VALID remains set. STOP automatically transmitted on bus. Master transmitter receives SMBus PEC Error SR.DNAK is set. SR.CCOMP not set. CMDR.VALID remains set. STOP automatically transmitted on bus. Master receiver discovers SMBus PEC Error SR.PECERR is set. SR.CCOMP not set. CMDR.VALID remains set. STOP automatically transmitted on bus. CR.STOP is written by user SR.STOP is set. SR.CCOMP set. CMDR.VALID remains set. STOP transmitted on bus after current byte transfer has finished. 470 32099I–01/2012 AT32UC3L016/32/64 21.9 User Interface Table 21-6. Note: TWIM Register Memory Map Offset Register Register Name Access Reset 0x00 Control Register CR Write-only 0x00000000 0x04 Clock Waveform Generator Register CWGR Read/Write 0x00000000 0x08 SMBus Timing Register SMBTR Read/Write 0x00000000 0x0C Command Register CMDR Read/Write 0x00000000 0x10 Next Command Register NCMDR Read/Write 0x00000000 0x14 Receive Holding Register RHR Read-only 0x00000000 0x18 Transmit Holding Register THR Write-only 0x00000000 0x1C Status Register SR Read-only 0x00000002 0x20 Interrupt Enable Register IER Write-only 0x00000000 0x24 Interrupt Disable Register IDR Write-only 0x00000000 0x28 Interrupt Mask Register IMR Read-only 0x00000000 0x2C Status Clear Register SCR Write-only 0x00000000 0x30 Parameter Register PR Read-only -(1) 0x34 Version Register VR Read-only -(1) 1. The reset values for these registers are device specific. Please refer to the Module Configuration section at the end of this chapter. 471 32099I–01/2012 AT32UC3L016/32/64 21.9.1 Name: Control Register CR Access Type: Write-only Offset: 0x00 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - STOP 7 6 5 4 3 2 1 0 SWRST - SMDIS SMEN - - MDIS MEN • STOP: Stop the Current Transfer Writing a one to this bit terminates the current transfer, sending a STOP condition after the shifter has become idle. If there are additional pending transfers, they will have to be explicitly restarted by software after the STOP condition has been successfully sent. Writing a zero to this bit has no effect. • SWRST: Software Reset If the TWIM master interface is enabled, writing a one to this bit resets the TWIM. All transfers are halted immediately, possibly violating the bus semantics. If the TWIM master interface is not enabled, it must first be enabled before writing a one to this bit. Writing a zero to this bit has no effect. • SMDIS: SMBus Disable Writing a one to this bit disables SMBus mode. Writing a zero to this bit has no effect. • SMEN: SMBus Enable Writing a one to this bit enables SMBus mode. Writing a zero to this bit has no effect. • MDIS: Master Disable Writing a one to this bit disables the master interface. Writing a zero to this bit has no effect. • MEN: Master Enable Writing a one to this bit enables the master interface. Writing a zero to this bit has no effect. 472 32099I–01/2012 AT32UC3L016/32/64 21.9.2 Name: Clock Waveform Generator Register CWGR Access Type: Read/Write Offset: 0x04 Reset Value: 0x00000000 31 30 - 23 29 28 27 26 EXP 22 21 25 24 DATA 20 19 18 17 16 11 10 9 8 3 2 1 0 STASTO 15 14 13 12 HIGH 7 6 5 4 LOW • EXP: Clock Prescaler Used to specify how to prescale the TWCK clock. Counters are prescaled according to the following formula f CLK_TWIM f PRESCALER = ------------------------( EXP + 1 ) 2 • DATA: Data Setup and Hold Cycles Clock cycles for data setup and hold count. Prescaled by CWGR.EXP. Used to time THD_DAT, TSU_DAT. • STASTO: START and STOP Cycles Clock cycles in clock high count. Prescaled by CWGR.EXP. Used to time THD_STA, TSU_STA, TSU_STO • HIGH: Clock High Cycles Clock cycles in clock high count. Prescaled by CWGR.EXP. Used to time THIGH. • LOW: Clock Low Cycles Clock cycles in clock low count. Prescaled by CWGR.EXP. Used to time TLOW, TBUF. 473 32099I–01/2012 AT32UC3L016/32/64 21.9.3 Name: SMBus Timing Register SMBTR Access Type: Read/Write Offset: 0x08 Reset Value: 0x00000000 31 30 29 28 EXP 23 22 21 20 27 26 25 24 - - - - 19 18 17 16 11 10 9 8 3 2 1 0 THMAX 15 14 13 12 TLOWM 7 6 5 4 TLOWS • EXP: SMBus Timeout Clock Prescaler Used to specify how to prescale the TIM and TLOWM counters in SMBTR. Counters are prescaled according to the following formula f CLKTWIM f prescaled, SMBus = -----------------------( EXP + 1 ) 2 • THMAX: Clock High Maximum Cycles Clock cycles in clock high maximum count. Prescaled by SMBTR.EXP. Used for bus free detection. Used to time THIGH:MAX. NOTE: Uses the prescaler specified by CWGR, NOT the prescaler specified by SMBTR. • TLOWM: Master Clock Stretch Maximum Cycles Clock cycles in master maximum clock stretch count. Prescaled by SMBTR.EXP. Used to time TLOW:MEXT • TLOWS: Slave Clock Stretch Maximum Cycles Clock cycles in slave maximum clock stretch count. Prescaled by SMBTR.EXP. Used to time TLOW:SEXT. 474 32099I–01/2012 AT32UC3L016/32/64 21.9.4 Name: Command Register CMDR Access Type: Read/Write Offset: 0x0C Reset Value: 0x00000000 31 30 - 23 29 28 - 22 21 20 27 26 25 24 - - ACKLAST PECEN 19 18 17 16 10 9 8 NBYTES 15 14 13 12 11 VALID STOP START REPSAME TENBIT 7 6 5 4 3 SADR[6:0] SADR[9:7] 2 1 0 READ • ACKLAST: ACK Last Master RX Byte 0: Causes the last byte in master receive mode (when NBYTES has reached 0) to be NACKed. This is the standard way of ending a master receiver transfer. 1: Causes the last byte in master receive mode (when NBYTES has reached 0) to be ACKed. Used for performing linked transfers in master receiver mode with no STOP or REPEATED START between the subtransfers. This is needed when more than 255 bytes are to be received in one single transmission. • PECEN: Packet Error Checking Enable 0: Causes the transfer not to use PEC byte verification. The PEC LFSR is still updated for every bit transmitted or received. Must be used if SMBus mode is disabled. 1: Causes the transfer to use PEC. PEC byte generation (if master transmitter) or PEC byte verification (if master receiver) will be performed. • NBYTES: Number of Data Bytes in Transfer The number of data bytes in the transfer. After the specified number of bytes have been transferred, a STOP condition is transmitted if CMDR.STOP is one. In SMBus mode, if PEC is used, NBYTES includes the PEC byte, i.e. there are NBYTES-1 data bytes and a PEC byte. • VALID: CMDR Valid 0: Indicates that CMDR does not contain a valid command. 1: Indicates that CMDR contains a valid command. This bit is cleared when the command is finished. • STOP: Send STOP Condition 0: Do not transmit a STOP condition after the data bytes have been transmitted. 1: Transmit a STOP condition after the data bytes have been transmitted. • START: Send START Condition 0: The transfer in CMDR should not commence with a START or REPEATED START condition. 1: The transfer in CMDR should commence with a START or REPEATED START condition. If the bus is free when the command is executed, a START condition is used. If the bus is busy, a REPEATED START is used. • REPSAME: Transfer is to Same Address as Previous Address Only used in 10-bit addressing mode, always write to 0 in 7-bit addressing mode. 475 32099I–01/2012 AT32UC3L016/32/64 Write this bit to one if the command in CMDR performs a repeated start to the same slave address as addressed in the previous transfer in order to enter master receiver mode. Write this bit to zero otherwise. • TENBIT: Ten Bit Addressing Mode 0: Use 7-bit addressing mode. 1: Use 10-bit addressing mode. Must not be used when the TWIM is in SMBus mode. • SADR: Slave Address Address of the slave involved in the transfer. Bits 9-7 are don’t care if 7-bit addressing is used. • READ: Transfer Direction 0: Allow the master to transmit data. 1: Allow the master to receive data. 476 32099I–01/2012 AT32UC3L016/32/64 21.9.5 Name: Next Command Register NCMDR Access Type: Read/Write Offset: 0x10 Reset Value: 0x00000000 31 30 - 29 28 - 23 22 21 20 27 26 25 24 - - ACKLAST PECEN 19 18 17 16 10 9 8 NBYTES 15 14 13 12 11 VALID STOP START REPSAME TENBIT 7 6 5 4 3 SADR[6:0] SADR[9:7] 2 1 0 READ This register is identical to CMDR. When the VALID bit in CMDR becomes 0, the content of NCMDR is copied into CMDR, clearing the VALID bit in NCMDR. If the VALID bit in CMDR is cleared when NCMDR is written, the content is copied immediately. 477 32099I–01/2012 AT32UC3L016/32/64 21.9.6 Name: Receive Holding Register RHR Access Type: Read-only Offset: 0x14 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 RXDATA • RXDATA: Received Data When the RXRDY bit in the Status Register (SR) is one, this field contains a byte received from the TWI bus. 478 32099I–01/2012 AT32UC3L016/32/64 21.9.7 Name: Transmit Holding Register THR Access Type: Write-only Offset: 0x18 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 TXDATA • TXDATA: Data to Transmit Write data to be transferred on the TWI bus here. 479 32099I–01/2012 AT32UC3L016/32/64 21.9.8 Name: Status Register SR Access Type: Read-only Offset: 0x1C Reset Value: 0x00000002 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - MENB 15 14 13 12 11 10 9 8 - STOP PECERR TOUT SMBALERT ARBLST DNAK ANAK 7 6 5 4 3 2 1 0 - - BUSFREE IDLE CCOMP CRDY TXRDY RXRDY • MENB: Master Interface Enable 0: Master interface is disabled. 1: Master interface is enabled. • STOP: Stop Request Accepted This bit is one when a STOP request caused by writing a one to CR.STOP has been accepted, and transfer has stopped. This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR). • PECERR: PEC Error This bit is one when a SMBus PEC error occurred. This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR). • TOUT: Timeout This bit is one when a SMBus timeout occurred. This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR). • SMBALERT: SMBus Alert This bit is one when an SMBus Alert was received. This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR). • ARBLST: Arbitration Lost This bit is one when the actual state of the SDA line did not correspond to the data driven onto it, indicating a higher-priority transmission in progress by a different master. This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR). • DNAK: NAK in Data Phase Received This bit is one when no ACK was received form slave during data transmission. This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR). • ANAK: NAK in Address Phase Received This bit is one when no ACK was received from slave during address phase This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR). • BUSFREE: Two-wire Bus is Free This bit is one when activity has completed on the two-wire bus. Otherwise, this bit is cleared. 480 32099I–01/2012 AT32UC3L016/32/64 • IDLE: Master Interface is Idle This bit is one when no command is in progress, and no command waiting to be issued. Otherwise, this bit is cleared. • CCOMP: Command Complete This bit is one when the current command has completed successfully. This bit is zero if the command failed due to conditions such as a NAK receved from slave. This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR). • CRDY: Ready for More Commands This bit is one when CMDR and/or NCMDR is ready to receive one or more commands. This bit is cleared when this is no longer true. • TXRDY: THR Data Ready This bit is one when THR is ready for one or more data bytes. This bit is cleared when this is no longer true (i.e. THR is full or transmission has stopped). • RXRDY: RHR Data Ready This bit is one when RX data are ready to be read from RHR. This bit is cleared when this is no longer true. 481 32099I–01/2012 AT32UC3L016/32/64 21.9.9 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x20 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - STOP PECERR TOUT SMBALERT ARBLST DNAK ANAK 7 6 5 4 3 2 1 0 - - BUSFREE IDLE CCOMP CRDY TXRDY RXRDY Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR 482 32099I–01/2012 AT32UC3L016/32/64 21.9.10 Name: Interrupt Disable Register IDR Access Type: Write-only Offset: 0x24 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - STOP PECERR TOUT SMBALERT ARBLST DNAK ANAK 7 6 5 4 3 2 1 0 - - BUSFREE IDLE CCOMP CRDY TXRDY RXRDY Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR 483 32099I–01/2012 AT32UC3L016/32/64 21.9.11 Name: Interrupt Mask Register IMR Access Type: Read-only Offset: 0x28 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - STOP PECERR TOUT SMBALERT ARBLST DNAK ANAK 7 6 5 4 3 2 1 0 - - BUSFREE IDLE CCOMP CRDY TXRDY RXRDY 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. This bit is cleared when the corresponding bit in IDR is written to one. This bit is set when the corresponding bit in IER is written to one. 484 32099I–01/2012 AT32UC3L016/32/64 21.9.12 Name: Status Clear Register SCR Access Type : Write-only Offset: 0x2C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - STOP PECERR TOUT SMBALERT ARBLST DNAK ANAK 7 6 5 4 3 2 1 0 - - - - CCOMP - - - Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in SR and the corresponding interrupt request. 485 32099I–01/2012 AT32UC3L016/32/64 21.9.13 Name: Parameter Register (PR) PR Access Type: Read-only Offset: 0x30 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - - 486 32099I–01/2012 AT32UC3L016/32/64 21.9.14 Name: Version Register (VR) VR Access Type: Read-only Offset: 0x34 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION [11:8] 3 2 1 0 VERSION [7:0] • VARIANT: Variant Number Reserved. No functionality associated. • VERSION: Version Number Version number of the module. No functionality associated. 487 32099I–01/2012 AT32UC3L016/32/64 21.10 Module Configuration The specific configuration for each TWIM instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 21-7. Module Clock Name Module Name Clock Name Description TWIM0 CLK_TWIM0 Clock for the TWIM0 bus interface TWIM1 CLK_TWIM1 Clock for the TWIM1 bus interface Table 21-8. Register Reset Values Register Reset Value VERSION 0x0000 0101 PARAMETER 0x0000 0000 488 32099I–01/2012 AT32UC3L016/32/64 22. Two-wire Slave Interface (TWIS) Rev.: 1.1.2.1 22.1 Features • Compatible with I²C standard • • • • • • 22.2 – Transfer speeds of 100 and 400 kbit/s – 7 and 10-bit and General Call addressing Compatible with SMBus standard – Hardware Packet Error Checking (CRC) generation and verification with ACK response – SMBALERT interface – 25 ms clock low timeout delay – 25 ms slave cumulative clock low extend time Compatible with PMBus DMA interface for reducing CPU load Arbitrary transfer lengths, including 0 data bytes Optional clock stretching if transmit or receive buffers not ready for data transfer 32-bit Peripheral Bus interface for configuration of the interface Overview The Atmel Two-wire Slave Interface (TWIS) interconnects components on a unique two-wire bus, made up of one clock line and one data line with speeds of up to 400 kbit/s, based on a byte-oriented transfer format. It can be used with any Atmel Two-wire Interface bus, I²C, or SMBus-compatible master. The TWIS is always a bus slave and can transfer sequential or single bytes. Below, Table 22-1 lists the compatibility level of the Atmel Two-wire Slave Interface and a full I²C compatible device. Table 22-1. Atmel TWIS Compatibility with I²C Standard I²C Standard Atmel TWIS Standard-mode (100 kbit/s) Supported Fast-mode (400 kbit/s) Supported 7 or 10 bits Slave Addressing Supported START BYTE(1) Not Supported Repeated Start (Sr) Condition Supported ACK and NAK Management Supported Slope control and input filtering (Fast mode) Supported Clock stretching Supported Note: 1. START + b000000001 + Ack + Sr 489 32099I–01/2012 AT32UC3L016/32/64 Below, Table 22-2 lists the compatibility level of the Atmel Two-wire Slave Interface and a full SMBus compatible device. Table 22-2. 22.3 SMBus Standard Atmel TWIS Bus Timeouts Supported Address Resolution Protocol Supported Alert Supported Packet Error Checking Supported List of Abbreviations Table 22-3. 22.4 Atmel TWIS Compatibility with SMBus Standard Abbreviations Abbreviation Description TWI Two-wire Interface A Acknowledge NA Non Acknowledge P Stop S Start Sr Repeated Start SADR Slave Address ADR Any address except SADR R Read W Write Block Diagram Figure 22-1. Block Diagram Peripheral Bus Bridge TWCK I/O Controller Two-wire Interface Power Manager TWD TWALM Interrupt Controller CLK_TWIS TWI Interrupt 490 32099I–01/2012 AT32UC3L016/32/64 22.5 Application Block Diagram Figure 22-2. Application Block Diagram VDD Rp Host with TWI Interface Rp TWD TWCK Atmel TWI serial EEPROM Slave 1 I²C RTC I²C LCD controller I²C temp. sensor Slave 2 Slave 3 Slave 4 Rp: Pull up value as given by the I²C Standard 22.6 I/O Lines Description Table 22-4. I/O Lines Description Pin Name Pin Description TWD Two-wire Serial Data Input/Output TWCK Two-wire Serial Clock Input/Output TWALM SMBus SMBALERT Input/Output 22.7 Type Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 22.7.1 I/O Lines TWDand TWCK are bidirectional lines, connected to a positive supply voltage via a current source or pull-up resistor (see Figure 22-5 on page 493). When the bus is free, both lines are high. The output stages of devices connected to the bus must have an open-drain or open-collector to perform the wired-AND function. TWALM is used to implement the optional SMBus SMBALERT signal. TWALM, TWD, and TWCK pins may be multiplexed with I/O Controller lines. To enable the TWIS, the user must perform the following steps: • Program the I/O Controller to: – Dedicate TWD, TWCK, and optionally TWALM as peripheral lines. – Define TWD, TWCK, and optionally TWALM as open-drain. 491 32099I–01/2012 AT32UC3L016/32/64 22.7.2 Power Management If the CPU enters a sleep mode that disables clocks used by the TWIS, the TWIS will stop functioning and resume operation after the system wakes up from sleep mode. The TWIS is able to wake the system from sleep mode upon address match, see Section 22.8.7 on page 500. 22.7.3 Clocks The clock for the TWIS bus interface (CLK_TWIS) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the TWIS before disabling the clock, to avoid freezing the TWIS in an undefined state. 22.7.4 DMA The TWIS DMA handshake interface is connected to the Peripheral DMA Controller. Using the TWIS DMA functionality requires the Peripheral DMA Controller to be programmed after setting up the TWIS. 22.7.5 Interrupts The TWIS interrupt request lines are connected to the interrupt controller. Using the TWIS interrupts requires the interrupt controller to be programmed first. 22.7.6 22.8 22.8.1 Debug Operation When an external debugger forces the CPU into debug mode, the TWIS continues normal operation. If the TWIS is configured in a way that requires it to be periodically serviced by the CPU through interrupts or similar, improper operation or data loss may result during debugging. Functional Description Transfer Format The data put on the TWD line must be 8 bits long. Data is transferred MSB first; each byte must be followed by an acknowledgement. The number of bytes per transfer is unlimited (see Figure 22-4 on page 493). Each transfer begins with a START condition and terminates with a STOP condition (see Figure 22-3). • A high-to-low transition on the TWD line while TWCK is high defines the START condition. • A low-to-high transition on the TWD line while TWCK is high defines a STOP condition. Figure 22-3. START and STOP Conditions TWD TWCK Start Stop 492 32099I–01/2012 AT32UC3L016/32/64 Figure 22-4. Transfer Format TWD TWCK Start 22.8.2 Address R/W Ack Data Ack Data Ack Stop Operation The TWIS has two modes of operation: • Slave transmitter mode • Slave receiver mode A master is a device which starts and stops a transfer and generates the TWCK clock. A slave is assigned an address and responds to requests from the master. These modes are described in the following chapters. Figure 22-5. Typical Application Block Diagram VDD Rp Host with TWI Interface Rp TWD TWCK Atmel TWI Serial EEPROM Slave 1 I²C RTC I²C LCD Controller I²C Temp. Sensor Slave 2 Slave 3 Slave 4 Rp: Pull up value as given by the I²C Standard 22.8.2.1 Bus Timing The Timing Register (TR) is used to control the timing of bus signals driven by the TWIS. TR describes bus timings as a function of cycles of the prescaled CLK_TWIS. The clock prescaling can be selected through TR.EXP. f CLK_TWIS f PRESCALED = -----------------------( EXP + 1 ) 2 TR has the following fields: TLOWS: Prescaled clock cycles used to time SMBUS timeout TLOW:SEXT. 493 32099I–01/2012 AT32UC3L016/32/64 TTOUT: Prescaled clock cycles used to time SMBUS timeout TTIMEOUT. SUDAT: Non-prescaled clock cycles for data setup and hold count. Used to time TSU_DAT. EXP: Specifies the clock prescaler setting used for the SMBUS timeouts. Figure 22-6. Bus Timing Diagram t HIGH t LOW S t HD:STA t LOW t SU:DAT t HD:DAT t t 22.8.2.2 t SU:DAT SU:STA SU:STO P Sr Setting Up and Performing a Transfer Operation of the TWIS is mainly controlled by the Control Register (CR). The following list presents the main steps in a typical communication: 3. Before any transfers can be performed, bus timings must be configured by writing to the Timing Register (TR).If the Peripheral DMA Controller is to be used for the transfers, it must be set up. 4. The Control Register (CR) must be configured with information such as the slave address, SMBus mode, Packet Error Checking (PEC), number of bytes to transfer, and which addresses to match. The interrupt system can be set up to generate interrupt request on specific events or error conditions, for example when a byte has been received. The NBYTES register is only used in SMBus mode, when PEC is enabled. In I²C mode or in SMBus mode when PEC is disabled, the NBYTES register is not used, and should be written to zero. NBYTES is updated by hardware, so in order to avoid hazards, software updates of NBYTES can only be done through writes to the NBYTES register. 22.8.2.3 Address Matching The TWIS can be set up to match several different addresses. More than one address match may be enabled simultaneously, allowing the TWIS to be assigned to several addresses. The address matching phase is initiated after a START or REPEATED START condition. When the TWIS receives an address that generates an address match, an ACK is automatically returned to the master. 494 32099I–01/2012 AT32UC3L016/32/64 In I²C mode: • The address in CR.ADR is checked for address match if CR.SMATCH is one. • The General Call address is checked for address match if CR.GCMATCH is one. In SMBus mode: • The address in CR.ADR is checked for address match if CR.SMATCH is one. • The Alert Response Address is checked for address match if CR.SMAL is one. • The Default Address is checked for address match if CR.SMDA is one. • The Host Header Address is checked for address match if CR.SMHH is one. 22.8.2.4 Clock Stretching Any slave or bus master taking part in a transfer may extend the TWCK low period at any time. The TWIS may extend the TWCK low period after each byte transfer if CR.STREN is one and: • Module is in slave transmitter mode, data should be transmitted, but THR is empty, or • Module is in slave receiver mode, a byte has been received and placed into the internal shifter, but the Receive Holding Register (RHR) is full, or • Stretch-on-address-match bit CR.SOAM=1 and slave was addressed. Bus clock remains stretched until all address match bits in the Status Register (SR) have been cleared. If CR.STREN is zero and: • Module is in slave transmitter mode, data should be transmitted but THR is empty: Transmit the value present in THR (the last transmitted byte or reset value), and set SR.URUN. • Module is in slave receiver mode, a byte has been received and placed into the internal shifter, but RHR is full: Discard the received byte and set SR.ORUN. 22.8.2.5 Bus Errors If a bus error (misplaced START or STOP) condition is detected, the SR.BUSERR bit is set and the TWIS waits for a new START condition. 22.8.3 Slave Transmitter Mode If the TWIS matches an address in which the R/W bit in the TWI address phase transfer is set, it will enter slave transmitter mode and set the SR.TRA bit (note that SR.TRA is set one CLK_TWIS cycle after the relevant address match bit in the same register is set). After the address phase, the following actions are performed: 1. If SMBus mode and PEC is used, NBYTES must be set up with the number of bytes to transmit. This is necessary in order to know when to transmit the PEC byte. NBYTES can also be used to count the number of bytes received if using DMA. 2. Byte to transmit depends on I²C/SMBus mode and CR.PEC: – If in I²C mode or CR.PEC is zero or NBYTES is non-zero: The TWIS waits until THR contains a valid data byte, possibly stretching the low period of TWCK. After THR contains a valid data byte, the data byte is transferred to a shifter, and then SR.TXRDY is changed to one because the THR is empty again. – SMBus mode and CR.PEC is one: If NBYTES is zero, the generated PEC byte is automatically transmitted instead of a data byte from THR. TWCK will not be stretched by the TWIS. 3. The data byte in the shifter is transmitted. 495 32099I–01/2012 AT32UC3L016/32/64 4. NBYTES is updated. If CR.CUP is one, NBYTES is incremented, otherwise NBYTES is decremented. 5. After each data byte has been transmitted, the master transmits an ACK (Acknowledge) or NAK (Not Acknowledge) bit. If a NAK bit is received by the TWIS, the SR.NAK bit is set. Note that this is done two CLK_TWIS cycles after TWCK has been sampled by the TWIS to be HIGH (see Figure 22-9). The NAK indicates that the transfer is finished, and the TWIS will wait for a STOP or REPEATED START. If an ACK bit is received, the SR.NAK bit remains LOW. The ACK indicates that more data should be transmitted, jump to step 2. At the end of the ACK/NAK clock cycle, the Byte Transfer Finished (SR.BTF) bit is set. Note that this is done two CLK_TWIS cycles after TWCK has been sampled by the TWIS to be LOW (see Figure 22-9). Also note that in the event that SR.NAK bit is set, it must not be cleared before the SR.BTF bit is set to ensure correct TWIS behavior. 6. If STOP is received, SR.TCOMP and SR.STO will be set. 7. If REPEATED START is received, SR.REP will be set. The TWI transfers require the receiver to acknowledge each received data byte. During the acknowledge clock pulse (9th pulse), the slave releases the data line (HIGH), enabling the master to pull it down in order to generate the acknowledge. The slave polls the data line during this clock pulse and sets the NAK bit in SR if the master does not acknowledge the data byte. A NAK means that the master does not wish to receive additional data bytes. As with the other status bits, an interrupt can be generated if enabled in the Interrupt Enable Register (IER). SR.TXRDY is used as Transmit Ready for the Peripheral DMA Controller transmit channel. The end of the complete transfer is marked by the SR.TCOMP bit changing from zero to one. See Figure 22-7 and Figure 22-8. Figure 22-7. Slave Transmitter with One Data Byte TWD S DADR R A DATA N P TCOMP TXRDY Write THR (DATA) NBYTES set to 1 STOP sent by master 496 32099I–01/2012 AT32UC3L016/32/64 Figure 22-8. Slave Transmitter with Multiple Data Bytes TWD S DADR R A DATA n A DATA n+5 A DATA n+m N P TCOMP TXRDY Write THR (Data n) NBYTES set to m Write THR (Data n+1) Write THR (Data n+m) Last data sent STOP sent by master Figure 22-9. Timing Relationship between TWCK, SR.NAK, and SR.BTF TWD DATA (LSB) N P TWCK SR.NAK SR.BTF t1 t1 t1: (CLK_TWIS period) x 2 22.8.4 Slave Receiver Mode If the TWIS matches an address in which the R/W bit in the TWI address phase transfer is cleared, it will enter slave receiver mode and clear SR.TRA (note that SR.TRA is cleared one CLK_TWIS cycle after the relevant address match bit in the same register is set). After the address phase, the following is repeated: 1. If SMBus mode and PEC is used, NBYTES must be set up with the number of bytes to receive. This is necessary in order to know which of the received bytes is the PEC byte. NBYTES can also be used to count the number of bytes received if using DMA. 2. Receive a byte. Set SR.BTF when done. 3. Update NBYTES. If CR.CUP is written to one, NBYTES is incremented, otherwise NBYTES is decremented. NBYTES is usually configured to count downwards if PEC is used. 4. After a data byte has been received, the slave transmits an ACK or NAK bit. For ordinary data bytes, the CR.ACK field controls if an ACK or NAK should be returned. If PEC is enabled and the last byte received was a PEC byte (indicated by NBYTES equal to zero), The TWIS will automatically return an ACK if the PEC value was correct, otherwise a NAK will be returned. 5. If STOP is received, SR.TCOMP will be set. 6. If REPEATED START is received, SR.REP will be set. The TWI transfers require the receiver to acknowledge each received data byte. During the acknowledge clock pulse (9th pulse), the master releases the data line (HIGH), enabling the 497 32099I–01/2012 AT32UC3L016/32/64 slave to pull it down in order to generate the acknowledge. The master polls the data line during this clock pulse. The SR.RXRDY bit indicates that a data byte is available in the RHR. The RXRDY bit is also used as Receive Ready for the Peripheral DMA Controller receive channel. Figure 22-10. Slave Receiver with One Data Byte TWD S DADR W A DATA A P TCOMP RXRDY Read RHR Figure 22-11. Slave Receiver with Multiple Data Bytes TWD S DADR W A DATA n A DATA (n+1) A DAT A (n+m)-1 A DATA (n+m) P A TCOMP RXRDY Read RHR DATA n 22.8.5 Read RHR DATA (n+1) Read RHR DAT A (n+m)-1 Read RHR DATA (n+m) Using the Peripheral DMA Controller The use of the Peripheral DMA Controller significantly reduces the CPU load. The user can set up ring buffers for the Peripheral DMA Controller, containing data to transmit or free buffer space to place received data. By initializing NBYTES to zero before a transfer, and writing a one to CR.CUP, NBYTES is incremented by one each time a data has been transmitted or received. This allows the user to detect how much data was actually transferred by the DMA system. To assure correct behavior, respect the following programming sequences: 22.8.5.1 Data Transmit with the Peripheral DMA Controller 1. Initialize the transmit Peripheral DMA Controller (memory pointers, size, etc.). 2. Configure the TWIS (ADR, NBYTES, etc.). 3. Start the transfer by enabling the Peripheral DMA Controller to transmit. 4. Wait for the Peripheral DMA Controller end-of-transmit flag. 5. Disable the Peripheral DMA Controller. 22.8.5.2 Data Receive with the Peripheral DMA Controller 1. Initialize the receive Peripheral DMA Controller (memory pointers, size - 1, etc.). 2. Configure the TWIS (ADR, NBYTES, etc.). 498 32099I–01/2012 AT32UC3L016/32/64 3. Start the transfer by enabling the Peripheral DMA Controller to receive. 4. Wait for the Peripheral DMA Controller end-of-receive flag. 5. Disable the Peripheral DMA Controller. 22.8.6 SMBus Mode SMBus mode is enabled by writing a one to the SMBus Mode Enable (SMEN) bit in CR. SMBus mode operation is similar to I²C operation with the following exceptions: • Only 7-bit addressing can be used. • The SMBus standard describes a set of timeout values to ensure progress and throughput on the bus. These timeout values must be written to TR. • Transmissions can optionally include a CRC byte, called Packet Error Check (PEC). • A dedicated bus line, SMBALERT, allows a slave to get a master’s attention. • A set of addresses have been reserved for protocol handling, such as Alert Response Address (ARA) and Host Header (HH) Address. Address matching on these addresses can be enabled by configuring CR appropriately. 22.8.6.1 Packet Error Checking (PEC) Each SMBus transfer can optionally end with a CRC byte, called the PEC byte. Writing a one to the Packet Error Checking Enable (PECEN) bit in CR enables automatic PEC handling in the current transfer. The PEC generator is always updated on every bit transmitted or received, so that PEC handling on following linked transfers will be correct. In slave receiver mode, the master calculates a PEC value and transmits it to the slave after all data bytes have been transmitted. Upon reception of this PEC byte, the slave will compare it to the PEC value it has computed itself. If the values match, the data was received correctly, and the slave will return an ACK to the master. If the PEC values differ, data was corrupted, and the slave will return a NAK value. The SR.SMBPECERR bit is set automatically if a PEC error occurred. In slave transmitter mode, the slave calculates a PEC value and transmits it to the master after all data bytes have been transmitted. Upon reception of this PEC byte, the master will compare it to the PEC value it has computed itself. If the values match, the data was received correctly. If the PEC values differ, data was corrupted, and the master must take appropriate action. The PEC byte is automatically inserted in a slave transmitter transmission if PEC enabled when NBYTES reaches zero. The PEC byte is identified in a slave receiver transmission if PEC enabled when NBYTES reaches zero. NBYTES must therefore be set to the total number of data bytes in the transmission, including the PEC byte. 22.8.6.2 Timeouts The Timing Register (TR) configures the SMBus timeout values. If a timeout occurs, the slave will leave the bus. The SR.SMBTOUT bit is also set. 22.8.6.3 SMBALERT A slave can get the master’s attention by pulling the SMBALERT line low. This is done by writing a one to the SMBus Alert (SMBALERT) bit in CR. This will also enable address match on the Alert Response Address (ARA). 499 32099I–01/2012 AT32UC3L016/32/64 22.8.7 Wakeup from Sleep Modes by TWI Address Match The TWIS is able to wake the device up from a sleep mode upon an address match, including sleep modes where CLK_TWIS is stopped. After detecting the START condition on the bus, The TWIS will stretch TWCK until CLK_TWIS has started. The time required for starting CLK_TWIS depends on which sleep mode the device is in. After CLK_TWIS has started, the TWIS releases its TWCK stretching and receives one byte of data on the bus. At this time, only a limited part of the device, including the TWIS, receives a clock, thus saving power. The TWIS goes on to receive the slave address. If the address phase causes a TWIS address match, the entire device is wakened and normal TWIS address matching actions are performed. Normal TWI transfer then follows. If the TWIS is not addressed, CLK_TWIS is automatically stopped and the device returns to its original sleep mode. 22.8.8 Identifying Bus Events This chapter lists the different bus events, and how these affects the bits in the TWIS registers. This is intended to help writing drivers for the TWIS. Table 22-5. Bus Events Event Effect Slave transmitter has sent a data byte SR.THR is cleared. SR.BTF is set. The value of the ACK bit sent immediately after the data byte is given by CR.ACK. Slave receiver has received a data byte SR.RHR is set. SR.BTF is set. SR.NAK updated according to value of ACK bit received from master. Start+Sadr on bus, but address is to another slave None. Start+Sadr on bus, current slave is addressed, but address match enable bit in CR is not set None. Start+Sadr on bus, current slave is addressed, corresponding address match enable bit in CR set Correct address match bit in SR is set. SR.TRA updated according to transfer direction (updating is done one CLK_TWIS cycle after address match bit is set) Slave enters appropriate transfer direction mode and data transfer can commence. Start+Sadr on bus, current slave is addressed, corresponding address match enable bit in CR set, SR.STREN and SR.SOAM are set. Correct address match bit in SR is set. SR.TRA updated according to transfer direction (updating is done one CLK_TWIS cycle after address match bit is set). Slave stretches TWCK immediately after transmitting the address ACK bit. TWCK remains stretched until all address match bits in SR have been cleared. Slave enters appropriate transfer direction mode and data transfer can commence. Repeated Start received after being addressed SR.REP set. SR.TCOMP unchanged. Stop received after being addressed SR.STO set. SR.TCOMP set. 500 32099I–01/2012 AT32UC3L016/32/64 Table 22-5. Bus Events Event Effect Start, Repeated Start, or Stop received in illegal position on bus SR.BUSERR set. SR.STO and SR.TCOMP may or may not be set depending on the exact position of an illegal stop. Data is to be received in slave receiver mode, SR.STREN is set, and RHR is full TWCK is stretched until RHR has been read. Data is to be transmitted in slave receiver mode, SR.STREN is set, and THR is empty TWCK is stretched until THR has been written. Data is to be received in slave receiver mode, SR.STREN is cleared, and RHR is full TWCK is not stretched, read data is discarded. SR.ORUN is set. Data is to be transmitted in slave receiver mode, SR.STREN is cleared, and THR is empty TWCK is not stretched, previous contents of THR is written to bus. SR.URUN is set. SMBus timeout received SR.SMBTOUT is set. TWCK and TWD are immediately released. Slave transmitter in SMBus PEC mode has transmitted a PEC byte, that was not identical to the PEC calculated by the master receiver. Slave receiver discovers SMBus PEC Error Master receiver will transmit a NAK as usual after the last byte of a master receiver transfer. Master receiver will retry the transfer at a later time. SR.SMBPECERR is set. NAK returned after the data byte. 501 32099I–01/2012 AT32UC3L016/32/64 22.9 User Interface Table 22-6. TWIS Register Memory Map Offset Register Register Name Access Reset 0x00 Control Register CR Read/Write 0x00000000 0x04 NBYTES Register NBYTES Read/Write 0x00000000 0x08 Timing Register TR Read/Write 0x00000000 0x0C Receive Holding Register RHR Read-only 0x00000000 0x10 Transmit Holding Register THR Write-only 0x00000000 0x14 Packet Error Check Register PECR Read-only 0x00000000 0x18 Status Register SR Read-only 0x00000002 0x1C Interrupt Enable Register IER Write-only 0x00000000 0x20 Interrupt Disable Register IDR Write-only 0x00000000 0x24 Interrupt Mask Register IMR Read-only 0x00000000 0x28 Status Clear Register SCR Write-only 0x00000000 0x2C Parameter Register PR Read-only -(1) 0x30 Version Register VR Read-only -(1) Note: 1. The reset values for these registers are device specific. Please refer to the Module Configuration section at the end of this chapter. 502 32099I–01/2012 AT32UC3L016/32/64 22.9.1 Name: Control Register CR Access Type: Read/Write Offset: 0x00 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - TENBIT 23 22 21 20 19 18 17 16 ADR[9:8] ADR[7:0] 15 14 13 12 11 10 9 8 - SOAM CUP ACK PECEN SMHH SMDA SMBALERT 7 6 5 4 3 2 1 0 SWRST - - STREN GCMATCH SMATCH SMEN SEN • TENBIT: Ten Bit Address Match 0: Disables Ten Bit Address Match. 1: Enables Ten Bit Address Match. • ADR: Slave Address Slave address used in slave address match. Bits 9:0 are used if in 10-bit mode, bits 6:0 otherwise. • SOAM: Stretch Clock on Address Match 0: Does not stretch bus clock after address match. 1: Stretches bus clock after address match. • CUP: NBYTES Count Up 0: Causes NBYTES to count down (decrement) per byte transferred. 1: Causes NBYTES to count up (increment) per byte transferred. • ACK: Slave Receiver Data Phase ACK Value 0: Causes a low value to be returned in the ACK cycle of the data phase in slave receiver mode. 1: Causes a high value to be returned in the ACK cycle of the data phase in slave receiver mode. • PECEN: Packet Error Checking Enable 0: Disables SMBus PEC (CRC) generation and check. 1: Enables SMBus PEC (CRC) generation and check. • SMHH: SMBus Host Header 0: Causes the TWIS not to acknowledge the SMBus Host Header. 1: Causes the TWIS to acknowledge the SMBus Host Header. • SMDA: SMBus Default Address 0: Causes the TWIS not to acknowledge the SMBus Default Address. 1: Causes the TWIS to acknowledge the SMBus Default Address. • SMBALERT: SMBus Alert 0: Causes the TWIS to release the SMBALERT line and not to acknowledge the SMBus Alert Response Address (ARA). 1: Causes the TWIS to pull down the SMBALERT line and to acknowledge the SMBus Alert Response Address (ARA). • SWRST: Software Reset This bit will always read as 0. Writing a zero to this bit has no effect. 503 32099I–01/2012 AT32UC3L016/32/64 Writing a one to this bit resets the TWIS. • STREN: Clock Stretch Enable 0: Disables clock stretching if RHR/THR buffer full/empty. May cause over/underrun. 1: Enables clock stretching if RHR/THR buffer full/empty. • GCMATCH: General Call Address Match 0: Causes the TWIS not to acknowledge the General Call Address. 1: Causes the TWIS to acknowledge the General Call Address. • SMATCH: Slave Address Match 0: Causes the TWIS not to acknowledge the Slave Address. 1: Causes the TWIS to acknowledge the Slave Address. • SMEN: SMBus Mode Enable 0: Disables SMBus mode. 1: Enables SMBus mode. • SEN: Slave Enable 0: Disables the slave interface. 1: Enables the slave interface. 504 32099I–01/2012 AT32UC3L016/32/64 22.9.2 Name: NBYTES Register NBYTES Access Type: Read/Write Offset: 0x04 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 NBYTES • NBYTES: Number of Bytes to Transfer Writing to this field updates the NBYTES counter. The field can also be read to learn the progress of the transfer. NBYTES can be incremented or decremented automatically by hardware. 505 32099I–01/2012 AT32UC3L016/32/64 22.9.3 Name: Timing Register TR Access Type: Read/Write Offset: 0x08 Reset Value: 0x00000000 31 30 29 28 EXP 23 22 21 20 27 26 25 24 - - - - 19 18 17 16 11 10 9 8 3 2 1 0 SUDAT 15 14 13 12 TTOUT 7 6 5 4 TLOWS • EXP: Clock Prescaler Used to specify how to prescale the SMBus TLOWS counter. The counter is prescaled according to the following formula: f CLK_TWIS f PRESCALED = -----------------------( EXP + 1 ) 2 • SUDAT: Data Setup Cycles Non-prescaled clock cycles for data setup count. Used to time TSU_DAT. Data is driven SUDAT cycles after TWCK low detected. This timing is used for timing the ACK/NAK bits, and any data bits driven in slave transmitter mode. • TTOUT: SMBus TTIMEOUT Cycles Prescaled clock cycles used to time SMBus TTIMEOUT. • TLOWS: SMBus TLOW:SEXT Cycles Prescaled clock cycles used to time SMBus TLOW:SEXT. 506 32099I–01/2012 AT32UC3L016/32/64 22.9.4 Name: Receive Holding Register RHR Access Type: Read-only Offset: 0x0C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 RXDATA • RXDATA: Received Data Byte When the RXRDY bit in the Status Register (SR) is one, this field contains a byte received from the TWI bus. 507 32099I–01/2012 AT32UC3L016/32/64 22.9.5 Name: Transmit Holding Register THR Access Type: Write-only Offset: 0x10 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 TXDATA • TXDATA: Data Byte to Transmit Write data to be transferred on the TWI bus here. 508 32099I–01/2012 AT32UC3L016/32/64 22.9.6 Name: Packet Error Check Register PECR Access Type: Read-only Offset: 0x14 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 PEC • PEC: Calculated PEC Value The calculated PEC value. Updated automatically by hardware after each byte has been transferred. Reset by hardware after a STOP condition. Provided if the user manually wishes to control when the PEC byte is transmitted, or wishes to access the PEC value for other reasons. In ordinary operation, the PEC handling is done automatically by hardware. 509 32099I–01/2012 AT32UC3L016/32/64 22.9.7 Name: Status Register SR Access Type: Read-only Offset: 0x18 Reset Value: 0x000000002 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 BTF REP STO SMBDAM SMBHHM SMBALERTM GCM SAM 15 14 13 12 11 10 9 8 - BUSERR SMBPECERR SMBTOUT - - - NAK 7 6 5 4 3 2 1 0 ORUN URUN TRA - TCOMP SEN TXRDY RXRDY • BTF: Byte Transfer Finished This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when byte transfer has completed. • REP: Repeated Start Received This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when a REPEATED START condition is received. • STO: Stop Received This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when the STOP condition is received. • SMBDAM: SMBus Default Address Match This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when the received address matched the SMBus Default Address. • SMBHHM: SMBus Host Header Address Match This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when the received address matched the SMBus Host Header Address. • SMBALERTM: SMBus Alert Response Address Match This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when the received address matched the SMBus Alert Response Address. • GCM: General Call Match This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when the received address matched the General Call Address. • SAM: Slave Address Match This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when the received address matched the Slave Address. • BUSERR: Bus Error This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when a misplaced START or STOP condition has occurred. 510 32099I–01/2012 AT32UC3L016/32/64 • SMBPECERR: SMBus PEC Error This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when a SMBus PEC error has occurred. • SMBTOUT: SMBus Timeout This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when a SMBus timeout has occurred. • NAK: NAK Received This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when a NAK was received from the master during slave transmitter operation. • ORUN: Overrun This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when an overrun has occurred in slave receiver mode. Can only occur if CR.STREN is zero. • URUN: Underrun This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when an underrun has occurred in slave transmitter mode. Can only occur if CR.STREN is zero. • TRA: Transmitter Mode 0: The slave is in slave receiver mode. 1: The slave is in slave transmitter mode. • TCOMP: Transmission Complete This bit is cleared when the corresponding bit in SCR is written to one. This bit is set when transmission is complete. Set after receiving a STOP after being addressed. • SEN: Slave Enabled 0: The slave interface is disabled. 1: The slave interface is enabled. • TXRDY: TX Buffer Ready 0: The TX buffer is full and should not be written to. 1: The TX buffer is empty, and can accept new data. • RXRDY: RX Buffer Ready 0: No RX data ready in RHR. 1: RX data is ready to be read from RHR. 511 32099I–01/2012 AT32UC3L016/32/64 22.9.8 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x1C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 BTF REP STO SMBDAM SMBHHM SMBALERTM GCM SAM 15 14 13 12 11 10 9 8 - BUSERR SMBPECERR SMBTOUT - - - NAK 7 6 5 4 3 2 1 0 ORUN URUN - - TCOMP - TXRDY RXRDY Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will write a one to the corresponding bit in IMR. 512 32099I–01/2012 AT32UC3L016/32/64 22.9.9 Name: Interrupt Disable Register IDR Access Type: Write-only Offset: 0x20 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 BTF REP STO SMBDAM SMBHHM SMBALERTM GCM SAM 15 14 13 12 11 10 9 8 - BUSERR SMBPECERR SMBTOUT - - - NAK 7 6 5 4 3 2 1 0 ORUN URUN - - TCOMP - TXRDY RXRDY Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 513 32099I–01/2012 AT32UC3L016/32/64 22.9.10 Name: Interrupt Mask Register IMR Access Type: Read-only Offset: 0x24 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 BTF REP STO SMBDAM SMBHHM SMBALERTM GCM SAM 15 14 13 12 11 10 9 8 - BUSERR SMBPECERR SMBTOUT - - - NAK 7 6 5 4 3 2 1 0 ORUN URUN - - TCOMP - TXRDY RXRDY 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. This bit is cleared when the corresponding bit in IDR is written to one. This bit is set when the corresponding bit in IER is written to one. 514 32099I–01/2012 AT32UC3L016/32/64 22.9.11 Name: Status Clear Register SCR Access Type: Write-only Offset: 0x28 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 BTF REP STO SMBDAM SMBHHM SMBALERTM GCM SAM 15 14 13 12 11 10 9 8 - BUSERR SMBPECERR SMBTOUT - - - NAK 7 6 5 4 3 2 1 0 ORUN URUN - - TCOMP - - - Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in SR and the corresponding interrupt request. 515 32099I–01/2012 AT32UC3L016/32/64 22.9.12 Name: Parameter Register PR Access Type: Read-only Offset: 0x2C Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - - 516 32099I–01/2012 AT32UC3L016/32/64 22.9.13 Name: Version Register (VR) VR Access Type: Read-only Offset: 0x30 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION [11:8] 3 2 1 0 VERSION [7:0] • VARIANT: Variant Number Reserved. No functionality associated. • VERSION: Version Number Version number of the module. No functionality associated. 517 32099I–01/2012 AT32UC3L016/32/64 22.10 Module Configuration The specific configuration for each TWIS instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 22-7. Module Clock Name Module Name Clock Name Description TWIS0 CLK_TWIS0 Clock for the TWIS0 bus interface TWIS1 CLK_TWIS1 Clock for the TWIS1 bus interface Table 22-8. Register Reset Values Register Reset Value VERSION 0x00000112 PARAMETER 0x00000000 518 32099I–01/2012 AT32UC3L016/32/64 23. Pulse Width Modulation Controller (PWMA) Rev 1.0.1.0 23.1 Features • Left-aligned non-inverted 8-bit PWM • Common 8-bit timebase counter • • • • • • • • 23.2 – Asynchronous clock source supported – Spread-spectrum counter to allow a constantly varying duty cycle Separate 8-bit duty cycle register per channel Synchronized channel updates – No glitches when changing the duty cycles Interlinked operation supported – Multiple channels can be updated with the same duty cycle value at a time – Up to four channels can be updated with different duty cycle values at a time Interrupt on PWM timebase overflow Incoming peripheral events supported – Pre-defined channels support incoming (increase/decrease) peripheral events from the Peripheral Event System – Increase event will increase the duty cycle by one – Decrease event will decrease the duty cycle by one One output peripheral event supported – Connected to channel 0 and asserted when the common timebase counter is equal to the programmed duty cycle for channel 0 Output PWM waveform for each channel Open drain driving on selected pins for 5V PWM operation Overview The Pulse Width Modulation Controller (PWMA) controls several pulse width modulation (PWM) channels. The number of channels is specific to the device. Each channel controls one square output PWM waveform. Characteristics of the output PWM waveforms such as period and duty cycle are configured through the user interface. All user interface registers are mapped on the peripheral bus. The duty cycle value for each channel can be set independently, while the period is determined by a common timebase counter (TC). The timebase for the counter is selected by using the allocated asynchronous Generic Clock (GCLK). The user interface for the PWMA contains handshake and synchronizing logic to ensure that no glitches occur on the output PWM waveforms while changing the duty cycle values. PWMA duty cycle values can be changed using two approaches, either an interlinked singlevalue mode or an interlinked multi-value mode. In the interlinked single-value mode, any set of channels, up to 32 channels, can be updated simultaneously with the same value while the other channels remain unchanged. In the interlinked multi-value mode, up to 4 selected channels can be updated with 4 different values while the other channels remain unchanged. Some pins can be driven in open drain mode, allowing the PWMA to generate a 5V waveform using an external pullup resistor. 519 32099I–01/2012 AT32UC3L016/32/64 23.3 Block Diagram Figure 23-1. PWMA Block Diagram PB IRQ PB Clock Domain Control CLK_PWMA TOP TCLR SPREAD BUSY Duty Cycle Interrupt Handling Channel Select Adjust Channel_0 Sync Spread Spectrum Counter Timebase Counter GCLK PWM Blocks Duty Cycle Register TOFL ETV COMP Channel 0 Channel 1 Channel m GCLK Domain PWMA[m:0] 23.4 I/O Lines Description Each channel outputs one PWM waveform on one external I/O line. Table 23-1. 23.5 I/O Line Description Pin Name Pin Description Type PWMA[n] Output PWM waveform for one channel n Output PWMMOD[n] Output PWM waveform for one channel n, open drain mode Output Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 520 32099I–01/2012 AT32UC3L016/32/64 23.5.1 I/O Lines The pins used for interfacing the PWMA may be multiplexed with I/O Controller lines. The programmer must first program the I/O Controller to assign the desired PWMA pins to their peripheral function. It is only required to enable the PWMA outputs actually in use. 23.5.2 Power Management If the CPU enters a sleep mode that disables clocks used by the PWMA, the PWMA will stop functioning and resume operation after the system wakes up from sleep mode. 23.5.3 Clocks The clock for the PWMA bus interface (CLK_PWMA) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the PWMA before disabling the clock, to avoid freezing the PWMA in an undefined state. Additionally, the PWMA depends on a dedicated Generic Clock (GCLK). The GCLK can be set to a wide range of frequencies and clock sources and must be enabled in the System Control Interface (SCIF) before the PWMA can be used. 23.5.4 Interrupts The PWMA interrupt request lines are connected to the interrupt controller. Using the PWMA interrupts requires the interrupt controller to be programmed first. 23.5.5 Peripheral Events The PWMA peripheral events are connected via the Peripheral Event System. Refer to the Peripheral Event System chapter for details. 23.5.6 Debug Operation When an external debugger forces the CPU into debug mode, the PWMA continues normal operation. If the PWMA is configured in a way that requires it to be periodically serviced by the CPU through interrupts, improper operation or data loss may result during debugging. 23.6 Functional Description The PWMA embeds a number of PWM channel submodules, each providing an output PWM waveform. Each PWM channel contains a duty cycle register and a comparator. A common timebase counter for all channels determines the frequency and the period for all the PWM waveforms. 23.6.1 Enabling the PWMA Once the GCLK has been enabled, the PWMA is enabled by writing a one to the EN bit in the Control Register (CR). 23.6.2 Timebase Counter The top value of the timebase counter defines the period of the PWMA output waveform. The timebase counter starts at zero when the PWMA is enabled and counts upwards until it reaches its effective top value (ETV). The effective top value is defined by specifying the desired number of GCLK clock cycles in the TOP field of CR (CR.TOP) in normal operation (CR.SPREAD is 521 32099I–01/2012 AT32UC3L016/32/64 zero). When the timebase counter reaches its effective top value, it restarts counting from zero. The period of the PWMA output waveform is then: T PWMA = ( ETV + 1 ) ⋅ T GCLK The timebase counter can be reset by writing a one to the Timebase Clear bit in CR (CR.TCLR). Note that this can cause a glitch to the output PWM waveforms in use. 23.6.3 Spread Spectrum Counter The spread spectrum counter allows the generation of constantly varying duty cycles on the output PWM waveforms. This is achieved by varying the effective top value of the timebase counter in a range defined by the spread spectrum counter value. When CR.SPREAD is not zero, the spread spectrum counter is enabled. Its range is defined by CR.SPREAD. It starts to count from -CR.SPREAD when the PWMA is enabled or after reset and counts upwards. When it reaches CR.SPREAD, it restarts to count from -CR.SPREAD again. The spread spectrum counter will cause the effective top value (ETV) to vary from TOPSPREAD to TOP+SPREAD. Figure 23-2 on page 522 illustrates this. This leads to a constantly varying duty cycle on the PWM output waveforms though the duty cycle values stored are unchanged. Figure 23-2. PWMA Adjusting Top Value for Timebase Counter 0xFF SPREAD TOP Adjusting top value range for the timerbase counter -SPREAD Duty Cycle 0x0 23.6.3.1 Special considerations The maximum value of the timebase counter is 255. If SPREAD is written to a value that will cause the ETV to exceed this value, the spread spectrum counter’s range will be limited to prevent the timebase counter to exceed its maximum value. If SPREAD is written to a value causing (TOP-SPREAD) to be below zero, the spread specturm counter’s range will be limited to prevent the timebase counter to count below zero. In both cases, the SPREAD value read from the Control Register will be the same value as written to the SPREAD field. 522 32099I–01/2012 AT32UC3L016/32/64 When writing a one to CR.TCLR, the timebase counter and the spread spectrum counter are reset at their lower limit values and the effective top value of the timebase counter will also be reset. 23.6.4 Duty Cycle and Waveform Properties Each PWM channel has its own duty cycle value (DCV) which is write-only and cannot be read out. The duty cycle value can be changed in two approaches as described in Section23.6.5. When the duty cycle value is zero, the PWM output is zero. Otherwise, the PWM output is set when the timebase counter is zero, and cleared when the timebase counter reaches the duty cycle value. This is summarized as: ⎧ low when DCV = 0 or TC > DCV PWM Waveform = ⎨ ⎩ high when TC ≤ DCV and DCV ≠ 0 Note that when increasing the duty cycle value for one channel from 0 to 1, the number of GCLK cycles when the PWM waveform is high will jump from 0 to 2. When incrementing the duty cycle value by one for any other values, the number of GCLK cycle when the waveform is high will increase by one. This is summarized in Table 23-2. Table 23-2. PMW Waveform Duty Cycles Duty Cycle Value #Clock Cycles When Waveform is High #Clock Cycles When Waveform is Low 0 0 ETV+1 1 2 ETV-1 2 3 ETV-2 ... ... ... ETV-1 ETV 1 ETV ETV+1 0 Every other output PWM waveform toggles on the negative edge of the GCLK instead of the positive edge. This is to avoid too many I/O toggling simultaneously on the output I/O lines. 23.6.5 Updating Duty Cycle Values 23.6.5.1 Interlinked Single Value PWM Operation The PWM channels can be interlinked to allow multiple channels to be updated simultaneously with the same duty cycle value. This value must be written to the Interlinked Single Value Duty (ISDUTY) register. Each channel has a corresponding enabling bit in the Interlinked Single Value Channel Set (ISCHSETm) register. When a bit is written to one in the ISCHSETm register, the duty cycle register for the corresponding channel will be updated with the value stored in the ISDUTY register. It can only be updated when the READY bit in the Status Register (SR.READY) is one, indicating that the PWMA is ready for writing. Figure 23-3 on page 524 shows the writing procedure. It is thus possible to update the duty cycle values for up to 32 PWM channels within one ISCHSETm register at a time. 523 32099I–01/2012 AT32UC3L016/32/64 Figure 23-3. Interlinked Single Value PWM Operation Flow ISDUTY DUTYm 23.6.5.2 ISCHSET ... Write Enable DUTY1 DUTY0 Interlinked Multiple Value PWM Operation The interlinked multiple value PWM operation allows up to four channels to be updated simultaneously with different duty cycle values. These duty cycle values must be written to the IMDUTY register. The index number of the four channels to be updated is written to the four SEL fields in the Interlinked Multiple Value Channel Select (IMCHSEL) register (IMCHSEL.SEL). When the IMCHSEL register is written, the values stored in the IMDUTY register are synchronized to the duty cycle registers for the channels selected by the SEL fields. Figure 23-4 on page 524 shows the writing procedure. Note that only writes to the implemented channels will be effective. If one of the IMCHSEL.SEL fields points to a non-existing channel, the corresponding value in the IMDUTY register will not be written. If the same channel is specified in multiple IMCHSEL.SEL fields, the channel will be updated with the value stored in the corresponding upper field of the IMDUTY register. Figure 23-4. Interlinked Multiple Value PWM Operation Flow IMDUTY IMCHSEL MUX DUTYm 23.6.6 ... DUTY1 DUTY0 Open Drain Mode Some pins can be used in open drain mode, allowing the PWMA waveform to toggle between 0V and up to 5V on these pins. In this mode the PWMA will drive the pin to zero or leave the output open. An external pullup can be used to pull the pin up to the desired voltage. To enable open drain mode on a pin the PWMAOD function must be selected instead of the PWMA function in the I/O Controller. Please refer to the Module Configuration chapter for information about which pins are available in open drain mode. 23.6.7 Synchronization Both the timebase counter and the spread spectrum counter can be reset and the duty cycle registers can be written through the user interface of the module. This requires a synchronization between the PB and GCLK clock domains, which takes a few clock cycles of each clock domain. The BUSY bit in SR indicates when the synchronization is ongoing. Writing to the module while the BUSY bit is set will result in discarding the new value. 524 32099I–01/2012 AT32UC3L016/32/64 Note that the duty cycle registers will not be updated with the new values until the timebase counter reaches its top value, in order to avoid glitches. The BUSY bit in SR will always be set during this updating and synchronization period. 23.6.8 Interrupts When the timebase counter overflows, the Timebase Overflow bit in the Status Register (SR.TOFL) is set. If the corresponding bit in the Interrupt Mask Register (IMR) is set, an interrupt request will be generated. Since the user needs to wait until the user interface is available between each write due to synchronization, a READY bit is provided in SR, which can be used to generate an interrupt request. The interrupt request will be generated if the corresponding bit in IMR is set. Bits in IMR are set by writing a one to the corresponding bit in the Interrupt Enable Register (IER), and cleared by writing a one to the corresponding bit in the Interrupt Disable Register (IDR). The interrupt request remains active until the corresponding bit in SR is cleared by writing a one to the corresponding bit in the Status Clear Register (SCR). 23.6.9 Peripheral Events 23.6.9.1 Input Peripheral Events The pre-defined channels support input peripheral events from the Peripheral Event System. An increase event (event_incr) will increase the duty cycle value by one, and a decrease event (event_decr) will decrease the duty cycle value by one. If an increase event and a decrease event occur at the same time, the duty cycle value will not be changed. The number of channels supporting input peripheral events is device specific. Please refer to the Module Configuration section at the end of this chapter for details. Input peripheral events must be enabled by writing a one to the corresponding bit in the Chanel Event Enable Register (CHEERm) before peripheral events can be used to control the duty cycle value. Each bit in the register corresponds to one channel, where bit 0 corresponds to channel 0 and so on. Both the increase and decrease events are enabled for the corresponding channel when a bit in the CHEERm register is written to one. 23.6.9.2 Output Peripheral Event The PWMA also supports one output peripheral event (event_ch0) to the Peripheral Event System. This output peripheral event is connected to channel 0 and will be asserted when the timebase counter reaches the duty cycle value for channel 0. This output event is always enabled. 525 32099I–01/2012 AT32UC3L016/32/64 23.7 User Interface Table 23-3. Note: PWMA Register Memory Map Offset Register Register Name Access Reset 0x00 Control Register CR Read/Write 0x00000000 0x04 Interlinked Single Value Duty Register ISDUTY Write-only 0x00000000 0x08 Interlinked Multiple Value Duty Register IMDUTY Write-only 0x00000000 0x0C Interlinked Multiple Value Channel Select IMCHSEL Write-only 0x00000000 0x10 Interrupt Enable Register IER Write-only 0x00000000 0x14 Interrupt Disable Register IDR Write-only 0x00000000 0x18 Interrupt Mask Register IMR Read-only 0x00000000 0x1C Status Register SR Read-only 0x00000000 0x20 Status Clear Register SCR Write-only 0x00000000 0x24 Parameter Register PARAMETER Read-only - (1) 0x28 Version Register VERSION Read-only - (1) 0x30 Interlinked Single Value Channel Set 0 ISCHSET0 Write-only 0x00000000 0x38 Channel Event Enable Register 0 CHEER0 Write-only 0x00000000 0x40 Interlinked Single Value Channel Set 1 ISCHSET1 Write-only 0x00000000 0x48 Channel Event Enable Register 1 CHEER1 Write-only 0x00000000 1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter. 526 32099I–01/2012 AT32UC3L016/32/64 23.7.1 Name: Control Register CR Access Type: Read/Write Offset: 0x00 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - 15 14 13 SPREAD 12 11 10 9 8 TOP 7 6 5 4 3 2 1 0 - - - - - - TCLR EN • SPREAD: Spread Spectrum Limit Value The spread spectrum limit value, together with the TOP field, defines the range for the spread spectrum counter. It is introduced in order to achieve constant varying duty cycles on the output PWM waveforms. Refer to Section23.6.3 for more information. • TOP: Timebase Counter Top Value The top value for the timebase counter. The effective top value of the timebase counter is defined by both the TOP and the SPREAD fields. Refer to Section23.6.2 for more information. • TCLR: Timebase Clear Writing a zero to this bit has no effect. Writing a one to this bit will clear the timebase counter. This bit is always read as zero. • EN: Module Enable 0: The PWMA is disabled 1: The PWMA is enabled 527 32099I–01/2012 AT32UC3L016/32/64 23.7.2 Name: Interlinked Single Value Duty Register ISDUTY Access Type: Write-only Offset: 0x04 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 DUTY • DUTY: Duty Cycle Value The duty cycle value written to this field is written simultaneously to all channels selected in the ISCHSET registers. If the value zero is written to DUTY all affected channels will be disabled. In this state the output waveform will be zero all the time. 528 32099I–01/2012 AT32UC3L016/32/64 23.7.3 Name: Interlinked Multiple Value Duty Register IMDUTY Access Type: Write-only Offset: 0x08 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 DUTY3 23 22 21 20 DUTY2 15 14 13 12 DUTY1 7 6 5 4 DUTY0 • DUTYn: Duty Cycle The value written to DUTY field n will be written to the PWMA channel selected by the corresponding SEL field in the IMCHSEL register. If the value zero is written to DUTY all affected channels will be disabled. In this state the output waveform will be zero all the time. 529 32099I–01/2012 AT32UC3L016/32/64 23.7.4 Name: Interlinked Multiple Value Channel Select IMCHSEL Access Type: Write-only Offset: 0x0C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 SEL3 23 22 21 20 SEL2 15 14 13 12 SEL1 7 6 5 4 SEL0 • SELn: Channel Select The duty cycle of the PWMA channel SELn will be updated with the value in the DUTYn field of the IMDUTY register when IMCHSEL is written. If SELn points to a non-implemented channel, the write will be discarded. Note: The duty registers will be updated with the value stored in the IMDUTY register when the IMCHSEL register is written. Synchronization takes place immeidately when an IMCHSEL register is written. The duty cycle registers will, however, not be updated until the synchronization is completed and the timebase counter reaches its top value in order to avoid glitches. 530 32099I–01/2012 AT32UC3L016/32/64 23.7.5 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x10 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - READY - TOFL Writing a zero to a bit in this register has no effect Writing a one to a bit in this register will set the corresponding bit in IMR. 531 32099I–01/2012 AT32UC3L016/32/64 23.7.6 Interrupt Disable Register Name: IDR Access Type: Write-only Offset: 0x14 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - READY - TOFL Writing a zero to a bit in this register has no effect Writing a one to a bit in this register will clear the corresponding bit in IMR. 532 32099I–01/2012 AT32UC3L016/32/64 23.7.7 Interrupt Mask Register Name: IMR Access Type: Read-only Offset: 0x18 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - READY - TOFL 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. A bit in this register is cleared when the corresponding bit in IDR is written to one. A bit in this register is set when the corresponding bit in IER is written to one. 533 32099I–01/2012 AT32UC3L016/32/64 23.7.8 Name: Status Register SR Access Type: Read-only Offset: 0x1C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - BUSY READY - TOFL • BUSY: Interface Busy This bit is automatically cleared when the interface is no longer busy. This bit is set when the user interface is busy and will not respond to new write operations. • READY: Interface Ready This bit is cleared by writing a one to the corresponding bit in the SCR register. This bit is set when the BUSY bit has a 1-to-0 transition. • TOFL: Timebase Overflow This bit is cleared by writing a one to corresponding bit in the SCR register. This bit is set when the timebase counter has wrapped at its top value. 534 32099I–01/2012 AT32UC3L016/32/64 23.7.9 Name: Status Clear Register SCR Access Type: Write-only Offset: 0x20 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 1 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - READY - TOFL Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in SR and the corresponding interrupt request. This register always reads as zero. 535 32099I–01/2012 AT32UC3L016/32/64 23.7.10 Name: Parameter Register PARAMETER Access Type: Read-only Offset: 0x24 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 CHANNELS • CHANNELS: Channels Implemented This field contains the number of channels implemented on the device. 536 32099I–01/2012 AT32UC3L016/32/64 23.7.11 Name: Version Register VERSION Access Type: Read-only Offset: 0x28 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant Number Reserved. No functionality associated. • VERSION: Version Number Version number of the module. No functionality associated. 537 32099I–01/2012 AT32UC3L016/32/64 23.7.12 Name: Interlinked Single Value Channel Set ISCHSETm Access Type: Write-only Offset: 0x30+m*0x10 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 SET 23 22 21 20 SET 15 14 13 12 SET 7 6 5 4 SET • SET: Single Value Channel Set If the bit n in SET is one, the duty cycle of PWMA channel n will be updated with the value written to ISDUTY. If more than one ISCHSET register is present, ISCHSET0 controls channels 31 to 0 and ISCHSET1 controls channels 63 to 32. Note: The duty registers will be updated with the value stored in the ISDUTY register when any ISCHSETm register is written. Synchronization takes place immeidately when an ISCHSET register is written. The duty cycle registers will, however, not be updated until the synchronization is completed and the timebase counter reaches its top value in order to avoid glitches. 538 32099I–01/2012 AT32UC3L016/32/64 23.7.13 Name: Channel Event Enable Register CHEERm Access Type: Write-only Offset: 0x38+m*0x10 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 CHEE 23 22 21 20 CHEE 15 14 13 12 CHEE 7 6 5 4 CHEE • CHEE: Channel Event Enable 0: The input peripheral event for the corresponding channel is disabled. 1: The input peripheral event for the corresponding channel is enabled. Both increase and decrease events for channel n are enabled if bit n is one. If more than one CHEER register is present, CHEER0 controls channels 31-0 and CHEER1 controls channels 64-32 and so on. 539 32099I–01/2012 AT32UC3L016/32/64 23.8 Module Configuration The specific configuration for each PWMA instance is listed in the following tables. The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 23-4. PWMA Configuration Feature PWMA Number of PWM channels 36 Channels supporting incoming peripheral events 0, 6, 8, 9, 11, 14, 19, and 20 PWMA channels with Open Drain mode 21, 27, and 28 Table 23-5. PWMA Clocks Clock Name Descripton CLK_PWMA Clock for the PWMA bus interface GCLK The generic clock used for the PWMA is GCLK3 Table 23-6. Register Reset Values Register Reset Value VERSION 0x00000101 PARAMETER 0x00000024 540 32099I–01/2012 AT32UC3L016/32/64 24. Timer/Counter (TC) Rev: 2.2.3.1.3 24.1 Features • Three 16-bit Timer Counter channels • A wide range of functions including: • • • • 24.2 – Frequency measurement – Event counting – Interval measurement – Pulse generation – Delay timing – Pulse width modulation – Up/down capabilities Each channel is user-configurable and contains: – Three external clock inputs – Five internal clock inputs – Two multi-purpose input/output signals Internal interrupt signal Two global registers that act on all three TC channels Peripheral event input on all A lines in capture mode Overview The Timer Counter (TC) includes three identical 16-bit Timer Counter channels. Each channel can be independently programmed to perform a wide range of functions including frequency measurement, event counting, interval measurement, pulse generation, delay timing, and pulse width modulation. Each channel has three external clock inputs, five internal clock inputs, and two multi-purpose input/output signals which can be configured by the user. Each channel drives an internal interrupt signal which can be programmed to generate processor interrupts. The TC block has two global registers which act upon all three TC channels. The Block Control Register (BCR) allows the three channels to be started simultaneously with the same instruction. The Block Mode Register (BMR) defines the external clock inputs for each channel, allowing them to be chained. 541 32099I–01/2012 AT32UC3L016/32/64 24.3 Block Diagram Figure 24-1. TC Block Diagram I/O Contr oller TIMER_CLOCK1 TCLK0 TIMER_CLOCK2 TIOA1 TIMER_CLOCK4 XC0 TIOA2 TIMER_CLOCK3 TCLK1 XC1 TCLK2 XC2 TIMER_CLOCK5 Timer/Counter Channel 0 TIOA TIOB A0 TIOA0 B0 TIOB0 TC0XC0S SYNC CLK0 CLK1 CLK2 INT0 TCLK0 XC0 TCLK1 TIOA0 XC1 TIOA2 XC2 TCLK2 Timer/Counter Channel 1 XC0 TCLK1 XC1 TCLK2 XC2 TIOA0 TIOA1 TC2XC2S TIOB A1 TIOA1 B1 TIOB1 SYNC TC1XC1S TCLK0 TIOA Timer/Counter Channel 2 INT1 TIOA TIOB SYNC A2 TIOA2 B2 TIOB2 INT2 Timer Count er Interrupt Controller 24.4 I/O Lines Description Table 24-1. 24.5 I/O Lines Description Pin Name Description Type CLK0-CLK2 External Clock Input Input A0-A2 I/O Line A Input/Output B0-B2 I/O Line B Input/Output Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 24.5.1 I/O Lines The pins used for interfacing the compliant external devices may be multiplexed with I/O lines. The user must first program the I/O Controller to assign the TC pins to their peripheral functions. 542 32099I–01/2012 AT32UC3L016/32/64 When using the TIOA lines as inputs the user must make sure that no peripheral events are generated on the line. Refer to the Peripheral Event System chapter for details. 24.5.2 Power Management If the CPU enters a sleep mode that disables clocks used by the TC, the TC will stop functioning and resume operation after the system wakes up from sleep mode. 24.5.3 Clocks The clock for the TC bus interface (CLK_TC) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the TC before disabling the clock, to avoid freezing the TC in an undefined state. 24.5.4 Interrupts The TC interrupt request line is connected to the interrupt controller. Using the TC interrupt requires the interrupt controller to be programmed first. 24.5.5 Peripheral Events The TC peripheral events are connected via the Peripheral Event System. Refer to the Peripheral Event System chapter for details. 24.5.6 Debug Operation The Timer Counter clocks are frozen during debug operation, unless the OCD system keeps peripherals running in debug operation. 24.6 Functional Description 24.6.1 TC Description The three channels of the Timer Counter are independent and identical in operation. The registers for channel programming are listed in Figure 24-3 on page 558. 24.6.1.1 Channel I/O Signals As described in Figure 24-1 on page 542, each Channel has the following I/O signals. Table 24-2. Channel I/O Signals Description Block/Channel Signal Name XC0, XC1, XC2 Channel Signal External Clock Inputs TIOA Capture mode: Timer Counter Input Waveform mode: Timer Counter Output TIOB Capture mode: Timer Counter Input Waveform mode: Timer Counter Input/Output INT SYNC 24.6.1.2 Description Interrupt Signal Output Synchronization Input Signal 16-bit counter Each channel is organized around a 16-bit counter. The value of the counter is incremented at each positive edge of the selected clock. When the counter has reached the value 0xFFFF and passes to 0x0000, an overflow occurs and the Counter Overflow Status bit in the Channel n Status Register (SRn.COVFS) is set. 543 32099I–01/2012 AT32UC3L016/32/64 The current value of the counter is accessible in real time by reading the Channel n Counter Value Register (CVn). The counter can be reset by a trigger. In this case, the counter value passes to 0x0000 on the next valid edge of the selected clock. 24.6.1.3 Clock selection At block level, input clock signals of each channel can either be connected to the external inputs TCLK0, TCLK1 or TCLK2, or be connected to the configurable I/O signals A0, A1 or A2 for chaining by writing to the BMR register. See Figure 24-2 on page 544. Each channel can independently select an internal or external clock source for its counter: • Internal clock signals: TIMER_CLOCK1, TIMER_CLOCK2, TIMER_CLOCK3, TIMER_CLOCK4, TIMER_CLOCK5. See the Module Configuration Chapter for details about the connection of these clock sources. • External clock signals: XC0, XC1 or XC2. See the Module Configuration Chapter for details about the connection of these clock sources. This selection is made by the Clock Selection field in the Channel n Mode Register (CMRn.TCCLKS). The selected clock can be inverted with the Clock Invert bit in CMRn (CMRn.CLKI). This allows counting on the opposite edges of the clock. The burst function allows the clock to be validated when an external signal is high. The Burst Signal Selection field in the CMRn register (CMRn.BURST) defines this signal. Note: In all cases, if an external clock is used, the duration of each of its levels must be longer than the CLK_TC period. The external clock frequency must be at least 2.5 times lower than the CLK_TC. Figure 24-2. Clock Selection TCCLKS TIMER_CLOCK1 TIMER_CLOCK2 CLKI TIMER_CLOCK3 TIMER_CLOCK4 TIMER_CLOCK5 Selected Clock XC0 XC1 XC2 BURST 1 544 32099I–01/2012 AT32UC3L016/32/64 24.6.1.4 Clock control The clock of each counter can be controlled in two different ways: it can be enabled/disabled and started/stopped. See Figure 24-3 on page 545. • The clock can be enabled or disabled by the user by writing to the Counter Clock Enable/Disable Command bits in the Channel n Clock Control Register (CCRn.CLKEN and CCRn.CLKDIS). In Capture mode it can be disabled by an RB load event if the Counter Clock Disable with RB Loading bit in CMRn is written to one (CMRn.LDBDIS). In Waveform mode, it can be disabled by an RC Compare event if the Counter Clock Disable with RC Compare bit in CMRn is written to one (CMRn.CPCDIS). When disabled, the start or the stop actions have no effect: only a CLKEN command in CCRn can re-enable the clock. When the clock is enabled, the Clock Enabling Status bit is set in SRn (SRn.CLKSTA). • The clock can also be started or stopped: a trigger (software, synchro, external or compare) always starts the clock. In Capture mode the clock can be stopped by an RB load event if the Counter Clock Stopped with RB Loading bit in CMRn is written to one (CMRn.LDBSTOP). In Waveform mode it can be stopped by an RC compare event if the Counter Clock Stopped with RC Compare bit in CMRn is written to one (CMRn.CPCSTOP). The start and the stop commands have effect only if the clock is enabled. Figure 24-3. Clock Control Selected Clock Trigger CLKSTA Q Q S CLKEN CLKDIS S R R Counter Clock 24.6.1.5 Stop Event Disable Event TC operating modes Each channel can independently operate in two different modes: • Capture mode provides measurement on signals. • Waveform mode provides wave generation. The TC operating mode selection is done by writing to the Wave bit in the CCRn register (CCRn.WAVE). In Capture mode, TIOA and TIOB are configured as inputs. 545 32099I–01/2012 AT32UC3L016/32/64 In Waveform mode, TIOA is always configured to be an output and TIOB is an output if it is not selected to be the external trigger. 24.6.1.6 Trigger A trigger resets the counter and starts the counter clock. Three types of triggers are common to both modes, and a fourth external trigger is available to each mode. The following triggers are common to both modes: • Software Trigger: each channel has a software trigger, available by writing a one to the Software Trigger Command bit in CCRn (CCRn.SWTRG). • SYNC: each channel has a synchronization signal SYNC. When asserted, this signal has the same effect as a software trigger. The SYNC signals of all channels are asserted simultaneously by writing a one to the Synchro Command bit in the BCR register (BCR.SYNC). • Compare RC Trigger: RC is implemented in each channel and can provide a trigger when the counter value matches the RC value if the RC Compare Trigger Enable bit in CMRn (CMRn.CPCTRG) is written to one. The channel can also be configured to have an external trigger. In Capture mode, the external trigger signal can be selected between TIOA and TIOB. In Waveform mode, an external event can be programmed to be one of the following signals: TIOB, XC0, XC1, or XC2. This external event can then be programmed to perform a trigger by writing a one to the External Event Trigger Enable bit in CMRn (CMRn.ENETRG). If an external trigger is used, the duration of the pulses must be longer than the CLK_TC period in order to be detected. Regardless of the trigger used, it will be taken into account at the following active edge of the selected clock. This means that the counter value can be read differently from zero just after a trigger, especially when a low frequency signal is selected as the clock. 24.6.1.7 24.6.2 Peripheral events on TIOA inputs The TIOA input lines are ored internally with peripheral events from the Peripheral Event System. To capture using events the user must ensure that the corresponding pin functions for the TIOA line are disabled. When capturing on the external TIOA pin the user must ensure that no peripheral events are generated on this pin. Capture Operating Mode This mode is entered by writing a zero to the CMRn.WAVE bit. Capture mode allows the TC channel to perform measurements such as pulse timing, frequency, period, duty cycle and phase on TIOA and TIOB signals which are considered as inputs. Figure 24-4 on page 548 shows the configuration of the TC channel when programmed in Capture mode. 24.6.2.1 Capture registers A and B Registers A and B (RA and RB) are used as capture registers. This means that they can be loaded with the counter value when a programmable event occurs on the signal TIOA. 546 32099I–01/2012 AT32UC3L016/32/64 The RA Loading Selection field in CMRn (CMRn.LDRA) defines the TIOA edge for the loading of the RA register, and the RB Loading Selection field in CMRn (CMRn.LDRB) defines the TIOA edge for the loading of the RB register. RA is loaded only if it has not been loaded since the last trigger or if RB has been loaded since the last loading of RA. RB is loaded only if RA has been loaded since the last trigger or the last loading of RB. Loading RA or RB before the read of the last value loaded sets the Load Overrun Status bit in SRn (SRn.LOVRS). In this case, the old value is overwritten. 24.6.2.2 Trigger conditions In addition to the SYNC signal, the software trigger and the RC compare trigger, an external trigger can be defined. The TIOA or TIOB External Trigger Selection bit in CMRn (CMRn.ABETRG) selects TIOA or TIOB input signal as an external trigger. The External Trigger Edge Selection bit in CMRn (CMRn.ETREDG) defines the edge (rising, falling or both) detected to generate an external trigger. If CMRn.ETRGEDG is zero (none), the external trigger is disabled. 547 32099I–01/2012 32099I–01/2012 TIOA TIOB SYNC MTIOA MTIOB TIMER_CLOCK2 TIMER_CLOCK3 TIMER_CLOCK4 TIMER_CLOCK5 XC0 XC1 XC2 TIMER_CLOCK1 1 Edge Detector ETRGEDG SWTRG CLKI Edge Detector LDRA CLK Trig S R OVF If RA is Loaded CPCTRG 16-bit Counter RESET Q LDBSTOP R S CLKEN Edge Detector LDRB Capture Register A Q CLKSTA LDBDIS Capture Register B CLKDIS SR Timer/Counter Channel If RA is not Loaded or RB is Loaded ABETRG BURST TCCLKS Compare RC = Register C COVFS LDRBS INT AT32UC3L016/32/64 Figure 24-4. Capture Mode LOVRS CPCS LDRAS ETRGS IMR 548 AT32UC3L016/32/64 24.6.3 Waveform Operating Mode Waveform operating mode is entered by writing a one to the CMRn.WAVE bit. In Waveform operating mode the TC channel generates one or two PWM signals with the same frequency and independently programmable duty cycles, or generates different types of oneshot or repetitive pulses. In this mode, TIOA is configured as an output and TIOB is defined as an output if it is not used as an external event. Figure 24-5 on page 550 shows the configuration of the TC channel when programmed in Waveform operating mode. 24.6.3.1 Waveform selection Depending on the Waveform Selection field in CMRn (CMRn.WAVSEL), the behavior of CVn varies. With any selection, RA, RB and RC can all be used as compare registers. RA Compare is used to control the TIOA output, RB Compare is used to control the TIOB output (if correctly configured) and RC Compare is used to control TIOA and/or TIOB outputs. 549 32099I–01/2012 32099I–01/2012 TIOB SYNC XC2 XC1 TIMER_CLOCK5 XC0 TIMER_CLOCK4 TIMER_CLOCK3 TIMER_CLOCK2 TIMER_CLOCK1 1 Edge Detector EEVTEDG SWTRG ENETRG Trig CLK R S Register A Q CLKSTA Compare RA = OVF WAVSEL RESET 16-bit Counter WAVSEL Q SR Timer/Counter Channel EEVT BURST CLKI Compare RC = Compare RB = CPCSTOP CPCDIS Register C CLKDIS Register B R S CLKEN CPAS INT BSWTRG BEEVT BCPB BCPC ASWTRG AEEVT ACPA ACPC O utput Contr oller O utput Cont r oller TCCLKS TIOB MTIOB TIOA MTIOA AT32UC3L016/32/64 Figure 24-5. Waveform Mode CPCS CPBS COVFS ETRGS IMR 550 AT32UC3L016/32/64 24.6.3.2 WAVSEL = 0 When CMRn.WAVSEL is zero, the value of CVn is incremented from 0 to 0xFFFF. Once 0xFFFF has been reached, the value of CVn is reset. Incrementation of CVn starts again and the cycle continues. See Figure 24-6 on page 551. An external event trigger or a software trigger can reset the value of CVn. It is important to note that the trigger may occur at any time. See Figure 24-7 on page 552. RC Compare cannot be programmed to generate a trigger in this configuration. At the same time, RC Compare can stop the counter clock (CMRn.CPCSTOP = 1) and/or disable the counter clock (CMRn.CPCDIS = 1). Figure 24-6. WAVSEL= 0 Without Trigger Counter Value Counter cleared by compare match with 0xFFFF 0xFFFF RC RB RA Waveform Examples Time TIOB TIOA 551 32099I–01/2012 AT32UC3L016/32/64 Figure 24-7. WAVSEL= 0 With Trigger Counter Value Counter cleared by compare match with 0xFFFF 0xFFFF RC Counter cleared by trigger RB RA Waveform Examples Time TIOB TIOA 24.6.3.3 WAVSEL = 2 When CMRn.WAVSEL is two, the value of CVn is incremented from zero to the value of RC, then automatically reset on a RC Compare. Once the value of CVn has been reset, it is then incremented and so on. See Figure 24-8 on page 553. It is important to note that CVn can be reset at any time by an external event or a software trigger if both are programmed correctly. See Figure 24-9 on page 553. In addition, RC Compare can stop the counter clock (CMRn.CPCSTOP) and/or disable the counter clock (CMRn.CPCDIS = 1). 552 32099I–01/2012 AT32UC3L016/32/64 Figure 24-8. WAVSEL = 2 Without Trigger Counter Value 0xFFFF Counter cleared by compare match with RC RC RB RA Waveform Examples Time TIOB TIOA Figure 24-9. WAVSEL = 2 With Trigger Counter Value 0xFFFF Counter cleared by compare match with RC Counter cleared by trigger RC RB RA Waveform Examples Time TIOB TIOA 24.6.3.4 WAVSEL = 1 When CMRn.WAVSEL is one, the value of CVn is incremented from 0 to 0xFFFF. Once 0xFFFF is reached, the value of CVn is decremented to 0, then re-incremented to 0xFFFF and so on. See Figure 24-10 on page 554. 553 32099I–01/2012 AT32UC3L016/32/64 A trigger such as an external event or a software trigger can modify CVn at any time. If a trigger occurs while CVn is incrementing, CVn then decrements. If a trigger is received while CVn is decrementing, CVn then increments. See Figure 24-11 on page 554. RC Compare cannot be programmed to generate a trigger in this configuration. At the same time, RC Compare can stop the counter clock (CMRn.CPCSTOP = 1) and/or disable the counter clock (CMRn.CPCDIS = 1). Figure 24-10. WAVSEL = 1 Without Trigger Counter Value Counter decremented by compare match with 0xFFFF 0xFFFF RC RB RA Time Waveform Examples TIOB TIOA Figure 24-11. WAVSEL = 1 With Trigger Counter Value Counter decremented by compare match with 0xFFFF 0xFFFF Counter decremented by trigger RC RB Counter incremented by trigger RA Waveform Examples Time TIOB TIOA 554 32099I–01/2012 AT32UC3L016/32/64 24.6.3.5 WAVSEL = 3 When CMRn.WAVSEL is three, the value of CVn is incremented from zero to RC. Once RC is reached, the value of CVn is decremented to zero, then re-incremented to RC and so on. See Figure 24-12 on page 555. A trigger such as an external event or a software trigger can modify CVn at any time. If a trigger occurs while CVn is incrementing, CVn then decrements. If a trigger is received while CVn is decrementing, CVn then increments. See Figure 24-13 on page 556. RC Compare can stop the counter clock (CMRn.CPCSTOP = 1) and/or disable the counter clock (CMRn.CPCDIS = 1). Figure 24-12. WAVSEL = 3 Without Trigger Counter Value 0xFFFF Counter cleared by compare match with RC RC RB RA Waveform Examples Time TIOB TIOA 555 32099I–01/2012 AT32UC3L016/32/64 Figure 24-13. WAVSEL = 3 With Trigger Counter Value 0xFFFF Counter decremented by compare match with RC RC Counter decremented by trigger RB Counter incremented by trigger RA Waveform Examples TIOB Time TIOA 24.6.3.6 External event/trigger conditions An external event can be programmed to be detected on one of the clock sources (XC0, XC1, XC2) or TIOB. The external event selected can then be used as a trigger. The External Event Selection field in CMRn (CMRn.EEVT) selects the external trigger. The External Event Edge Selection field in CMRn (CMRn.EEVTEDG) defines the trigger edge for each of the possible external triggers (rising, falling or both). If CMRn.EEVTEDG is written to zero, no external event is defined. If TIOB is defined as an external event signal (CMRn.EEVT = 0), TIOB is no longer used as an output and the compare register B is not used to generate waveforms and subsequently no IRQs. In this case the TC channel can only generate a waveform on TIOA. When an external event is defined, it can be used as a trigger by writing a one to the CMRn.ENETRG bit. As in Capture mode, the SYNC signal and the software trigger are also available as triggers. RC Compare can also be used as a trigger depending on the CMRn.WAVSEL field. 24.6.3.7 Output controller The output controller defines the output level changes on TIOA and TIOB following an event. TIOB control is used only if TIOB is defined as output (not as an external event). The following events control TIOA and TIOB: • software trigger • external event • RC compare RA compare controls TIOA and RB compare controls TIOB. Each of these events can be programmed to set, clear or toggle the output as defined in the following fields in CMRn: • RC Compare Effect on TIOB (CMRn.BCPC) 556 32099I–01/2012 AT32UC3L016/32/64 • RB Compare Effect on TIOB (CMRn.BCPB) • RC Compare Effect on TIOA (CMRn.ACPC) • RA Compare Effect on TIOA (CMRn.ACPA) 557 32099I–01/2012 AT32UC3L016/32/64 24.7 User Interface Table 24-3. TC Register Memory Map Offset Register Register Name Access Reset 0x00 Channel 0 Control Register CCR0 Write-only 0x00000000 0x04 Channel 0 Mode Register CMR0 Read/Write 0x00000000 0x10 Channel 0 Counter Value CV0 Read-only 0x00000000 0x14 Channel 0 Register A RA0 Read/Write(1) 0x00000000 0x18 Channel 0 Register B RB0 Read/Write(1) 0x00000000 0x1C Channel 0 Register C RC0 Read/Write 0x00000000 0x20 Channel 0 Status Register SR0 Read-only 0x00000000 0x24 Interrupt Enable Register IER0 Write-only 0x00000000 0x28 Channel 0 Interrupt Disable Register IDR0 Write-only 0x00000000 0x2C Channel 0 Interrupt Mask Register IMR0 Read-only 0x00000000 0x40 Channel 1 Control Register CCR1 Write-only 0x00000000 0x44 Channel 1 Mode Register CMR1 Read/Write 0x00000000 0x50 Channel 1 Counter Value CV1 Read-only 0x00000000 0x54 Channel 1 Register A RA1 (1) 0x00000000 (1) 0x00000000 Read/Write 0x58 Channel 1 Register B RB1 Read/Write 0x5C Channel 1 Register C RC1 Read/Write 0x00000000 0x60 Channel 1 Status Register SR1 Read-only 0x00000000 0x64 Channel 1 Interrupt Enable Register IER1 Write-only 0x00000000 0x68 Channel 1 Interrupt Disable Register IDR1 Write-only 0x00000000 0x6C Channel 1 Interrupt Mask Register IMR1 Read-only 0x00000000 0x80 Channel 2 Control Register CCR2 Write-only 0x00000000 0x84 Channel 2 Mode Register CMR2 Read/Write 0x00000000 0x90 Channel 2 Counter Value CV2 Read-only 0x00000000 0x94 Channel 2 Register A RA2 Read/Write(1) 0x00000000 0x98 Channel 2 Register B RB2 Read/Write(1) 0x00000000 0x9C Channel 2 Register C RC2 Read/Write 0x00000000 0xA0 Channel 2 Status Register SR2 Read-only 0x00000000 0xA4 Channel 2 Interrupt Enable Register IER2 Write-only 0x00000000 0xA8 Channel 2 Interrupt Disable Register IDR2 Write-only 0x00000000 0xAC Channel 2 Interrupt Mask Register IMR2 Read-only 0x00000000 0xC0 Block Control Register BCR Write-only 0x00000000 0xC4 Block Mode Register BMR Read/Write 0x00000000 0xF8 Features Register FEATURES Read-only -(2) 0xFC Version Register VERSION Read-only -(2) 558 32099I–01/2012 AT32UC3L016/32/64 Notes: 1. Read-only if CMRn.WAVE is zero. 2. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter. 559 32099I–01/2012 AT32UC3L016/32/64 24.7.1 Name: Channel Control Register CCR Access Type: Write-only Offset: 0x00 + n * 0x40 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - SWTRG CLKDIS CLKEN • SWTRG: Software Trigger Command 1: Writing a one to this bit will perform a software trigger: the counter is reset and the clock is started. 0: Writing a zero to this bit has no effect. • CLKDIS: Counter Clock Disable Command 1: Writing a one to this bit will disable the clock. 0: Writing a zero to this bit has no effect. • CLKEN: Counter Clock Enable Command 1: Writing a one to this bit will enable the clock if CLKDIS is not one. 0: Writing a zero to this bit has no effect. 560 32099I–01/2012 AT32UC3L016/32/64 24.7.2 Name: Channel Mode Register: Capture Mode CMR Access Type: Read/Write Offset: 0x04 + n * 0x40 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 11 10 WAVE CPCTRG - - - ABETRG 7 6 5 4 3 2 LDBDIS LDBSTOP BURST LDRB CLKI LDRA 9 8 ETRGEDG 1 0 TCCLKS • LDRB: RB Loading Selection LDRB Edge 0 none 1 rising edge of TIOA 2 falling edge of TIOA 3 each edge of TIOA • LDRA: RA Loading Selection LDRA Edge 0 none 1 rising edge of TIOA 2 falling edge of TIOA 3 each edge of TIOA • WAVE 1: Capture mode is disabled (Waveform mode is enabled). 0: Capture mode is enabled. • CPCTRG: RC Compare Trigger Enable 1: RC Compare resets the counter and starts the counter clock. 0: RC Compare has no effect on the counter and its clock. • ABETRG: TIOA or TIOB External Trigger Selection 1: TIOA is used as an external trigger. 561 32099I–01/2012 AT32UC3L016/32/64 0: TIOB is used as an external trigger. • ETRGEDG: External Trigger Edge Selection ETRGEDG Edge 0 none 1 rising edge 2 falling edge 3 each edge • LDBDIS: Counter Clock Disable with RB Loading 1: Counter clock is disabled when RB loading occurs. 0: Counter clock is not disabled when RB loading occurs. • LDBSTOP: Counter Clock Stopped with RB Loading 1: Counter clock is stopped when RB loading occurs. 0: Counter clock is not stopped when RB loading occurs. • BURST: Burst Signal Selection BURST Burst Signal Selection 0 The clock is not gated by an external signal 1 XC0 is ANDed with the selected clock 2 XC1 is ANDed with the selected clock 3 XC2 is ANDed with the selected clock • CLKI: Clock Invert 1: The counter is incremented on falling edge of the clock. 0: The counter is incremented on rising edge of the clock. • TCCLKS: Clock Selection TCCLKS Clock Selected 0 TIMER_CLOCK1 1 TIMER_CLOCK2 2 TIMER_CLOCK3 3 TIMER_CLOCK4 4 TIMER_CLOCK5 5 XC0 6 XC1 7 XC2 562 32099I–01/2012 AT32UC3L016/32/64 24.7.3 Name: Channel Mode Register: Waveform Mode CMR Access Type: Read/Write Offset: 0x04 + n * 0x40 Reset Value: 0x00000000 31 30 29 BSWTRG 23 27 BEEVT 22 21 ASWTRG 15 28 20 WAVE 13 7 6 19 CPCDIS CPCSTOP 4 BURST BCPB 18 11 ENETRG 5 24 17 16 ACPC 12 WAVSEL 25 BCPC AEEVT 14 26 ACPA 10 9 EEVT 3 CLKI 8 EEVTEDG 2 1 0 TCCLKS • BSWTRG: Software Trigger Effect on TIOB BSWTRG Effect 0 none 1 set 2 clear 3 toggle • BEEVT: External Event Effect on TIOB BEEVT Effect 0 none 1 set 2 clear 3 toggle 563 32099I–01/2012 AT32UC3L016/32/64 • BCPC: RC Compare Effect on TIOB BCPC Effect 0 none 1 set 2 clear 3 toggle • BCPB: RB Compare Effect on TIOB BCPB Effect 0 none 1 set 2 clear 3 toggle • ASWTRG: Software Trigger Effect on TIOA ASWTRG Effect 0 none 1 set 2 clear 3 toggle • AEEVT: External Event Effect on TIOA AEEVT Effect 0 none 1 set 2 clear 3 toggle • ACPC: RC Compare Effect on TIOA ACPC Effect 0 none 1 set 2 clear 3 toggle 564 32099I–01/2012 AT32UC3L016/32/64 • ACPA: RA Compare Effect on TIOA ACPA Effect 0 none 1 set 2 clear 3 toggle • WAVE 1: Waveform mode is enabled. 0: Waveform mode is disabled (Capture mode is enabled). • WAVSEL: Waveform Selection WAVSEL Effect 0 UP mode without automatic trigger on RC Compare 1 UPDOWN mode without automatic trigger on RC Compare 2 UP mode with automatic trigger on RC Compare 3 UPDOWN mode with automatic trigger on RC Compare • ENETRG: External Event Trigger Enable 1: The external event resets the counter and starts the counter clock. 0: The external event has no effect on the counter and its clock. In this case, the selected external event only controls the TIOA output. • EEVT: External Event Selection EEVT Note: Signal selected as external event TIOB Direction 0 TIOB input(1) 1 XC0 output 2 XC1 output 3 XC2 output 1. If TIOB is chosen as the external event signal, it is configured as an input and no longer generates waveforms and subsequently no IRQs. • EEVTEDG: External Event Edge Selection EEVTEDG Edge 0 none 1 rising edge 2 falling edge 3 each edge • CPCDIS: Counter Clock Disable with RC Compare 1: Counter clock is disabled when counter reaches RC. 0: Counter clock is not disabled when counter reaches RC. 565 32099I–01/2012 AT32UC3L016/32/64 • CPCSTOP: Counter Clock Stopped with RC Compare 1: Counter clock is stopped when counter reaches RC. 0: Counter clock is not stopped when counter reaches RC. • BURST: Burst Signal Selection BURST Burst Signal Selection 0 The clock is not gated by an external signal. 1 XC0 is ANDed with the selected clock. 2 XC1 is ANDed with the selected clock. 3 XC2 is ANDed with the selected clock. • CLKI: Clock Invert 1: Counter is incremented on falling edge of the clock. 0: Counter is incremented on rising edge of the clock. • TCCLKS: Clock Selection TCCLKS Clock Selected 0 TIMER_CLOCK1 1 TIMER_CLOCK2 2 TIMER_CLOCK3 3 TIMER_CLOCK4 4 TIMER_CLOCK5 5 XC0 6 XC1 7 XC2 566 32099I–01/2012 AT32UC3L016/32/64 24.7.4 Name: Channel Counter Value Register CV Access Type: Read-only Offset: 0x10 + n * 0x40 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 CV[15:8] 7 6 5 4 CV[7:0] • CV: Counter Value CV contains the counter value in real time. 567 32099I–01/2012 AT32UC3L016/32/64 24.7.5 Name: Channel Register A RA Access Type: Read-only if CMRn.WAVE = 0, Read/Write if CMRn.WAVE = 1 Offset: 0x14 + n * 0X40 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 RA[15:8] 7 6 5 4 RA[7:0] • RA: Register A RA contains the Register A value in real time. 568 32099I–01/2012 AT32UC3L016/32/64 24.7.6 Name: Channel Register B RB Access Type: Read-only if CMRn.WAVE = 0, Read/Write if CMRn.WAVE = 1 Offset: 0x18 + n * 0x40 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 RB[15:8] 7 6 5 4 RB[7:0] • RB: Register B RB contains the Register B value in real time. 569 32099I–01/2012 AT32UC3L016/32/64 24.7.7 Name: Channel Register C RC Access Type: Read/Write Offset: 0x1C + n * 0x40 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 RC[15:8] 7 6 5 4 RC[7:0] • RC: Register C RC contains the Register C value in real time. 570 32099I–01/2012 AT32UC3L016/32/64 24.7.8 Name: Channel Status Register SR Access Type: Read-only Offset: 0x20 + n * 0x40 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - MTIOB MTIOA CLKSTA 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 ETRGS LDRBS LDRAS CPCS CPBS CPAS LOVRS COVFS Note: Reading the Status Register will also clear the interrupt bit for the corresponding interrupts. • MTIOB: TIOB Mirror 1: TIOB is high. If CMRn.WAVE is zero, this means that TIOB pin is high. If CMRn.WAVE is one, this means that TIOB is driven high. 0: TIOB is low. If CMRn.WAVE is zero, this means that TIOB pin is low. If CMRn.WAVE is one, this means that TIOB is driven low. • MTIOA: TIOA Mirror 1: TIOA is high. If CMRn.WAVE is zero, this means that TIOA pin is high. If CMRn.WAVE is one, this means that TIOA is driven high. 0: TIOA is low. If CMRn.WAVE is zero, this means that TIOA pin is low. If CMRn.WAVE is one, this means that TIOA is driven low. • CLKSTA: Clock Enabling Status 1: This bit is set when the clock is enabled. 0: This bit is cleared when the clock is disabled. • ETRGS: External Trigger Status 1: This bit is set when an external trigger has occurred. 0: This bit is cleared when the SR register is read. • LDRBS: RB Loading Status 1: This bit is set when an RB Load has occurred and CMRn.WAVE is zero. 0: This bit is cleared when the SR register is read. • LDRAS: RA Loading Status 1: This bit is set when an RA Load has occurred and CMRn.WAVE is zero. 0: This bit is cleared when the SR register is read. • CPCS: RC Compare Status 1: This bit is set when an RC Compare has occurred. 0: This bit is cleared when the SR register is read. 571 32099I–01/2012 AT32UC3L016/32/64 • CPBS: RB Compare Status 1: This bit is set when an RB Compare has occurred and CMRn.WAVE is one. 0: This bit is cleared when the SR register is read. • CPAS: RA Compare Status 1: This bit is set when an RA Compare has occurred and CMRn.WAVE is one. 0: This bit is cleared when the SR register is read. • LOVRS: Load Overrun Status 1: This bit is set when RA or RB have been loaded at least twice without any read of the corresponding register and CMRn.WAVE is zero. 0: This bit is cleared when the SR register is read. • COVFS: Counter Overflow Status 1: This bit is set when a counter overflow has occurred. 0: This bit is cleared when the SR register is read. 572 32099I–01/2012 AT32UC3L016/32/64 24.7.9 Name: Channel Interrupt Enable Register IER Access Type: Write-only Offset: 0x24 + n * 0x40 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 ETRGS LDRBS LDRAS CPCS CPBS CPAS LOVRS COVFS Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 573 32099I–01/2012 AT32UC3L016/32/64 24.7.10 Name: Channel Interrupt Disable Register IDR Access Type: Write-only Offset: 0x28 + n * 0x40 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 ETRGS LDRBS LDRAS CPCS CPBS CPAS LOVRS COVFS Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 574 32099I–01/2012 AT32UC3L016/32/64 24.7.11 Name: Channel Interrupt Mask Register IMR Access Type: Read-only Offset: 0x2C + n * 0x40 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 ETRGS LDRBS LDRAS CPCS CPBS CPAS LOVRS COVFS 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. A bit in this register is cleared when the corresponding bit in IDR is written to one. A bit in this register is set when the corresponding bit in IER is written to one. 575 32099I–01/2012 AT32UC3L016/32/64 24.7.12 Name: Block Control Register BCR Access Type: Write-only Offset: 0xC0 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - SYNC • SYNC: Synchro Command 1: Writing a one to this bit asserts the SYNC signal which generates a software trigger simultaneously for each of the channels. 0: Writing a zero to this bit has no effect. 576 32099I–01/2012 AT32UC3L016/32/64 24.7.13 Name: Block Mode Register BMR Access Type: Read/Write Offset: 0xC4 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - TC2XC2S TC1XC1S TC0XC0S • TC2XC2S: External Clock Signal 2 Selection TC2XC2S Signal Connected to XC2 0 TCLK2 1 none 2 TIOA0 3 TIOA1 • TC1XC1S: External Clock Signal 1 Selection TC1XC1S Signal Connected to XC1 0 TCLK1 1 none 2 TIOA0 3 TIOA2 577 32099I–01/2012 AT32UC3L016/32/64 • TC0XC0S: External Clock Signal 0 Selection TC0XC0S Signal Connected to XC0 0 TCLK0 1 none 2 TIOA1 3 TIOA2 578 32099I–01/2012 AT32UC3L016/32/64 24.7.14 Name: Features Register FEATURES Access Type: Read-only Offset: 0xF8 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - BRPBHSB UPDNIMPL 7 6 5 4 3 2 1 0 CTRSIZE • BRPBHSB: Bridge type is PB to HSB 1: Bridge type is PB to HSB. 0: Bridge type is not PB to HSB. • UPDNIMPL: Up/down is implemented 1: Up/down counter capability is implemented. 0: Up/down counter capability is not implemented. • CTRSIZE: Counter size This field indicates the size of the counter in bits. 579 32099I–01/2012 AT32UC3L016/32/64 24.7.15 Name: Version Register VERSION Access Type: Read-only Offset: 0xFC Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 580 32099I–01/2012 AT32UC3L016/32/64 24.8 Module Configuration The specific configuration for each Timer/Counter instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 24-4. TC Bus Interface Clocks Clock Name Description CLK_TC0 Clock for the TC0 bus interface CLK_TC1 Clock for the TC1 bus interface Table 24-5. Register Reset Values Register Reset Value FEATURES 0x00000310 VERSION 0x00000223 Table 24-6. Features Feature Implementation Bridge type PB to HSB Up/down counter capability implemented Yes Counter size 16 Table 24-7. Event Sources Instance Source(s) TC0 TIOA TC1 - 581 32099I–01/2012 AT32UC3L016/32/64 24.8.1 Clock Connections There are two Timer/Counter modules, TC0 and TC1, with three channels each, giving a total of six Timer/Counter channels. Each Timer/Counter channel can independently select an internal or external clock source for its counter: Table 24-8. Timer/Counter Clock Connections Module Source Name Connection TC0 Internal TIMER_CLOCK1 32 KHz oscillator clock (OSC32K) TIMER_CLOCK2 PBA Clock / 2 TIMER_CLOCK3 PBA Clock / 8 TIMER_CLOCK4 PBA Clock / 32 TIMER_CLOCK5 PBA Clock / 128 XC0 See Section 3.2 on page 9 External XC1 XC2 TC1 Internal External TIMER_CLOCK1 32 KHz oscillator clock (OSC32K) TIMER_CLOCK2 PBA Clock / 2 TIMER_CLOCK3 PBA Clock / 8 TIMER_CLOCK4 PBA Clock / 32 TIMER_CLOCK5 PBA Clock / 128 XC0 See Section 3.2 on page 9 XC1 XC2 582 32099I–01/2012 AT32UC3L016/32/64 25. Peripheral Event System Rev: 1.0.0.1 25.1 Features • • • • 25.2 Direct peripheral to peripheral communication system Allows peripherals to receive, react to, and send peripheral events without CPU intervention Cycle deterministic event communication Asynchronous interrupts allow advanced peripheral operation in low power sleep modes Overview Several peripheral modules can be configured to emit or respond to signals known as peripheral events. The exact condition to trigger a peripheral event, or the action taken upon receiving a peripheral event, is specific to each module. Peripherals that respond to peripheral events are called peripheral event users and peripherals that emit peripheral events are called peripheral event generators. A single module can be both a peripheral event generator and user. The peripheral event generators and users are interconnected by a network known as the Peripheral Event System. This allows low latency peripheral-to-peripheral signaling without CPU intervention, and without consuming system resources such as bus or RAM bandwidth. This offloads the CPU and system resources compared to a traditional interrupt-based software driven system. 25.3 Peripheral Event System Block Diagram Figure 25-1. Peripheral Event System Block Diagram Generator Generator 25.4 25.4.1 Peripheral Event System User Generator/ User Functional Description Configuration The Peripheral Event System in the AT32UC3L016/32/64 has a fixed mapping of peripheral events between generators and users, as described in Table 25-1 to Table 25-4. Thus, the user does not need to configure the interconnection between the modules, although each peripheral event can be enabled or disabled at the generator or user side as described in the peripheral chapter for each module. 583 32099I–01/2012 AT32UC3L016/32/64 Table 25-1. Peripheral Event Mapping from ACIFB to PWMA Generator Generated Event ACIFB channel 0 AC0 VINP > AC0 VINN User Effect PWMA duty cycle value increased by one PWMA channel 0 AC0 VINN > AC0 VINP ACIFB channel 1 AC1 VINP > AC1 VINN PWMA duty cycle value decreased by one PWMA duty cycle value increased by one PWMA channel 6 AC1 VINN > AC1 VINP ACIFB channel 2 AC2 VINP > AC2 VINN PWMA duty cycle value decreased by one PWMA duty cycle value increased by one PWMA channel 8 AC2 VINN > AC2 VINP ACIFB channel 3 ACIFB channel 4 AC3 VINP > AC3 VINN PWMA duty cycle value decreased by one PWMA duty cycle value increased by one PWMA channel 9 AC3 VINN > AC3 VINP PWMA duty cycle value decreased by one AC4 VINP > AC4 VINN PWMA duty cycle value increased by one No PWMA channel 11 AC4 VINN > AC4 VINP ACIFB channel 5 AC5 VINP > AC5 VINN PWMA duty cycle value decreased by one PWMA duty cycle value increased by one PWMA channel 14 AC5 VINN > AC5 VINP ACIFB channel 6 AC6 VINP > AC6 VINN PWMA duty cycle value decreased by one PWMA duty cycle value increased by one PWMA channel 19 AC6 VINN > AC6 VINP ACIFB channel 7 AC7 VINP > AC7 VINN PWMA duty cycle value decreased by one PWMA duty cycle value increased by one PWMA channel 20 AC7 VINN > AC7 VINP ACIFB channel n Table 25-2. Generator ACn VINN > ACn VINP PWMA duty cycle value decreased by one Automatically used by the CAT when performing QMatrix acquisition. CAT No Peripheral Event Mapping from GPIO to TC Generated Event User Pin change on PA00-PA07 Pin change on PA08-PA15 GPIO Asynchronous Effect Asynchronous A0 capture TC0 Pin change on PA16-PA23 A1 capture A2 capture Pin change on PB00-PB07 No A1 capture TC1 Pin change on PB08-PB15 A2 capture 584 32099I–01/2012 AT32UC3L016/32/64 Table 25-3. Peripheral Event Mapping from AST Generator Generated Event User Effect ACIFB Comparison is triggered if the ACIFB.CONFn register is written to 11 (Event Triggered Single Measurement Mode) and the EVENTEN bit in the ACIFB.CTRL register is written to 1. ADCIFB Conversion is triggered if the TRGMOD bit in the ADCIFB.TRGR register is written to 111 (Peripheral Event Trigger). CAT Trigger one iteration of autonomous touch detection. Overflow event Periodic event Alarm event Overflow event AST Periodic event Alarm event Asynchronous Yes Overflow event Periodic event Alarm event Table 25-4. Peripheral Event Mapping from PWMA Generator PWMA channel 0 Generated Event User Effect Asynchronous ACIFB Comparison is triggered if the ACIFB.CONFn register is written to 11 (Event Triggered Single Measurement Mode) and the EVENTEN bit in the ACIFB.CTRL register is written to 1. No Timebase counter reaches the duty cycle value. ADCIFB 25.4.2 Conversion is triggered if the TRGMOD bit in the ADCIFB.TRGR register is written to 111 (Peripheral Event Trigger). Peripheral Event Connections Each generated peripheral event is connected to one or more users. If a peripheral event is connected to multiple users, the peripheral event can trigger actions in multiple modules. A peripheral event user can likewise be connected to one or more peripheral event generators. If a peripheral event user is connected to multiple generators, the peripheral events are OR’ed together to a single peripheral event. This means that peripheral events from either one of the generators will result in a peripheral event to the user. To configure a peripheral event, the peripheral event must be enabled at both the generator and user side. Even if a generator is connected to multiple users, only the users with the peripheral event enabled will trigger on the peripheral event. 25.4.3 Low Power Operation As the peripheral events do not require CPU intervention, they are available in Idle mode. They are also available in deeper sleep modes if both the generator and user remain clocked in that mode. Certain events are known as asynchronous peripheral events, as identified in Table 25-1 to Table 25-4. These can be issued even when the system clock is stopped, and revive unclocked user peripherals. The clock will be restarted for this module only, without waking the system from sleep mode. The clock remains active only as long as required by the triggered function, before being switched off again, and the system remains in the original sleep mode. The CPU and sys- 585 32099I–01/2012 AT32UC3L016/32/64 tem will only be woken up if the user peripheral generates an interrupt as a result of the operation. This concept is known as SleepWalking and is described in further detail in the Power Manager chapter. Note that asynchronous peripheral events may be associated with a delay due to the need to restart the system clock source if this has been stopped in the sleep mode. 25.5 Application Example This application example shows how the Peripheral Event System can be used to program the ADC Interface to perform ADC conversions at selected intervals. Conversions of the active analog channels are started with a software or a hardware trigger. One of the possible hardware triggers is a peripheral event trigger, allowing the Peripheral Event System to synchronize conversion with some configured peripheral event source. From Table 25-3 and Table 25-4, it can be read that this peripheral event source can be either an AST peripheral event, or an event from the PWM Controller. The AST can generate periodic peripheral events at selected intervals, among other types of peripheral events. The Peripheral Event System can then be used to set up the ADC Interface to sample an analog signal at regular intervals. The user must enable peripheral events in the AST and in the ADC Interface to accomplish this. The periodic peripheral event in the AST is enabled by writing a one to the corresponding bit in the AST Event Enable Register (EVE). To select the peripheral event trigger for the ADC Interface, the user must write the value 0x7 to the Trigger Mode (TRGMOD) field in the ADC Interface Trigger Register (TRGR). When the peripheral events are enabled, the AST will generate peripheral events at the selected intervals, and the Peripheral Event System will route the peripheral events to the ADC Interface, which will perform ADC conversions at the selected intervals. Figure 25-2. Application Example Trigger Periodic peripheral conversion Peripheral event AST Event System ADC Interface Since the AST peripheral event is asynchronous, the description above will also work in sleep modes where the ADC clock is stopped. In this case, the ADC clock (and clock source, if needed) will be restarted during the ADC conversion. After the conversion, the ADC clock and clock source will return to the sleep state, unless the ADC generates an interrupt, which in turn will wake up the system. Using asynchronous interrupts thus allows ADC operation in much lower power states than would otherwise be possible. 586 32099I–01/2012 AT32UC3L016/32/64 26. ADC Interface (ADCIFB) Rev:1.0.1.1 26.1 Features • Multi-channel Analog-to-Digital Converter with up to 12-bit resolution • Enhanced Resolution Mode • • • • • 26.2 – 11-bit resolution obtained by interpolating 4 samples – 12-bit resolution obtained by interpolating 16 samples Glueless interface with resistive touch screen panel, allowing – Resistive Touch Screen position measurement – Pen detection and pen loss detection Integrated enhanced sequencer – ADC Mode – Resistive Touch Screen Mode Numerous trigger sources – Software – Embedded 16-bit timer for periodic trigger – Pen detect trigger – Continuous trigger – External trigger, rising, falling, or any-edge trigger – Peripheral event trigger ADC Sleep Mode for low power ADC applications Programmable ADC timings – Programmable ADC clock – Programmable startup time Overview The ADC Interface (ADCIFB) converts analog input voltages to digital values. The ADCIFB is based on a Successive Approximation Register (SAR) 10-bit Analog-to-Digital Converter (ADC). The conversions extend from 0V to ADVREFP. The ADCIFB supports 8-bit and 10-bit resolution mode, in addition to enhanced resolution mode with 11-bit and 12-bit resolution. Conversion results are reported in a common register for all channels. The 11-bit and 12-bit resolution modes are obtained by interpolating multiple samples to acquire better accuracy. For 11-bit mode 4 samples are used, which gives an effective sample rate of 1/4 of the actual sample frequency. For 12-bit mode 16 samples are used, giving a effective sample rate of 1/16 of actual. This arrangement allows conversion speed to be traded for better accuracy. Conversions can be started for all enabled channels, either by a software trigger, by detection of a level change on the external trigger pin (TRIGGER), or by an integrated programmable timer. When the Resistive Touch Screen Mode is enabled, an integrated sequencer automatically configures the pad control signals and performs resistive touch screen conversions. The ADCIFB also integrates an ADC Sleep Mode, a Pen-Detect Mode, and an Analog Compare Mode, and connects with one Peripheral DMA Controller channel. These features reduce both power consumption and processor intervention. 587 32099I–01/2012 AT32UC3L016/32/64 26.3 Block Diagram Figure 26-1. ADCIFB Block Diagram ADCIFB I/O Controller TRIGGER ADP0 ADP1 Trigger Timer Resisitve Touch Screen Sequencer ADC Control Logic CLK_ADC CLK_ADCIFB User Interface Peripheral Bus AD1 AD2 .... AD3 ADn Analog Multiplexer AD0 Successive Approximation Register Analog-to-Digital Converter DMA Request Interrupt Request ADVREFP 588 32099I–01/2012 AT32UC3L016/32/64 26.4 I/O Lines Description Table 26-1. I/O Lines Description Pin Name Description Type ADVREFP Reference voltage Analog TRIGGER External trigger Digital ADP0 Drive Pin 0 for Resistive Touch Screen top channel (Xp) Digital ADP1 Drive Pin 1 for Resistive Touch Screen right channel (Yp) Digital AD0-ADn Analog input channels 0 to n Analog 26.5 Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 26.5.1 I/O Lines The analog input pins can be multiplexed with I/O Controller lines. The user must make sure the I/O Controller is configured correctly to allow the ADCIFB access to the AD pins before the ADCIFB is instructed to start converting data. If the user fails to do this the converted data may be wrong. The number of analog inputs is device dependent, please refer to the ADCIFB Module Configuration chapter for the number of available AD inputs on the current device. The ADVREFP pin must be connected correctly prior to using the ADCIFB. Failing to do so will result in invalid ADC operation. See the Electrical Characteristics chapter for details. If the TRIGGER, ADP0, and ADP1 pins are to be used in the application, the user must configure the I/O Controller to assign the needed pins to the ADCIFB function. 26.5.2 Power Management If the CPU enters a sleep mode that disables clocks used by the ADCIFB, the ADCIFB will stop functioning and resume operation after the system wakes up from sleep mode. If the Peripheral Event System is configured to send asynchronous peripheral events to the ADCIFB and the clock used by the ADCIFB is stopped, a local and temporary clock will automatically be requested so the event can be processed. Refer to Section 26.6.13, Section 26.6.12, and the Peripheral Event System chapter for details. Before entering a sleep mode where the clock to the ADCIFB is stopped, make sure the Analogto-Digital Converter cell is put in an inactive state. Refer to Section 26.6.13 for more information. 26.5.3 Clocks The clock for the ADCIFB bus interface (CLK_ADCIFB) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the ADCIFB before disabling the clock, to avoid freezing the ADCIFB in an undefined state. 589 32099I–01/2012 AT32UC3L016/32/64 26.5.4 DMA The ADCIFB DMA handshake interface is connected to the Peripheral DMA Controller. Using the ADCIFB DMA functionality requires the Peripheral DMA Controller to be programmed first. When using the ADCIFB DMA functionality CLK_ADCIFB can not be divided. 26.5.5 Interrupts The ADCIFB interrupt request line is connected to the interrupt controller. Using the ADCIFB interrupt request functionality requires the interrupt controller to be programmed first. 26.5.6 Peripheral Events The ADCIFB peripheral events are connected via the Peripheral Event System. Refer to the Peripheral Event System chapter for details 26.5.7 26.6 Debug Operation When an external debugger forces the CPU into debug mode, this module continues normal operation. If this module is configured in a way that requires it to be periodically serviced by the CPU through interrupt requests or similar, improper operation or data loss may result during debugging. Functional Description The ADCIFB embeds a Successive Approximation Register (SAR) Analog-to-Digital Converter (ADC). The ADC supports 8-bit or 10-bit resolution, which can be extended to 11 or 12 bits by the Enhanced Resolution Mode. The conversion is performed on a full range between 0 V and the reference voltage pin ADVREFP. Analog inputs between these voltages converts to digital values (codes) based on a linear conversion. This linear conversion is described in the expression below where M is the number of bits used to represent the analog value, Vin is the voltage of the analog value to convert, Vref is the maximum voltage, and Code is the converted digital value. M 2 ⋅ V in Code = -------------------V ref 26.6.1 Initializing the ADCIFB The ADC Interface is enabled by writing a one to the Enable bit in the Control Register (CR.EN). After the ADC Interface is enabled, the ADC timings needs to be configured by writing the correct values to the RES, PRESCAL, and STARTUP fields in the ADC Configuration Register (ACR). See Section 26.6.5, and Section 26.6.7 for details. Before the ADCIFB can be used, the I/O Controller must be configured correctly and the Reference Voltage (ADVREFP) signal must be connected. Refer to Section 26.5.1 for details. 26.6.2 Basic Operation To convert analog values to digital values the user must first initialize the ADCIFB as described in Section 26.6.1. When the ADCIFB is initialized the channels to convert must be enabled by writing a one the corresponding bits in the Channel Enable Register (CHER). Enabling channel N instructs the ADCIFB to convert the analog voltage applied to AD pin N at each conversion 590 32099I–01/2012 AT32UC3L016/32/64 sequence. Multiple channels can be enabled resulting in multiple AD pins being converted at each conversion sequence. To start converting data the user can either manually start a conversion sequence by writing a one to the START bit in the Control Register (CR.START) or configure an automatic trigger to initiate the conversions. The automatic trigger can be configured to trig on many different conditions. Refer to Section 26.8.1 for details. The result of the conversion is stored in the Last Converted Data Register (LCDR) as they become available, overwriting the result from the previous conversion. To avoid data loss if more than one channel is enabled, the user must read the conversion results as they become available either by using an interrupt handler or by using a Peripheral DMA channel to copy the results to memory. Failing to do so will result in an Overrun Error condition, indicated by the OVRE bit in the Status Register (SR). To use an interrupt handler the user must enable the Data Ready (DRDY) interrupt request by writing a one to the corresponding bit in the Interrupt Enable Register (IER). To clear the interrupt after the conversion result is read, the user must write a one to the corresponding bit in the Interrupt Clear Register (ICR). See Section 26.6.11 for details. To use a Peripheral DMA Controller channel the user must configure the Peripheral DMA Controller appropriately. The Peripheral DMA Controller will, when configured, automatically read converted data as they become available. There is no need to manually clear any bits in the Interrupt Status Register as this is performed by the hardware. If an Overrun Error condition happens during DMA operation, the OVRE bit in the SR will be set. 26.6.3 ADC Resolution The Analog-to-Digital Converter cell supports 8-bit or 10-bit resolution, which can be extended to 11-bit and 12-bit with the Enhanced Resolution Mode. The resolution is selected by writing the selected resolution value to the RES field in the ADC Configuration Register (ACR). See Section 26.9.3. By writing a zero to the RES field, the ADC switches to the lowest resolution and the conversion results can be read in the eight lowest significant bits of the Last Converted Data Register (LCDR). The four highest bits of the Last Converted Data (LDATA) field in the LCDR register reads as zero. Writing a one to the RES field enables 10-bit resolution, the optimal resolution for both sampling speed and accuracy. Writing two or three automatically enables Enhanced Resolution Mode with 11-bit or 12-bit resolution, see Section 26.6.4 for details. When a Peripheral DMA Controller channel is connected to the ADCIFB in 10-bit, 11-bit, or 12bit resolution mode, a transfer size of 16 bits must be used. By writing a zero to the RES field, the destination buffers can be optimized for 8-bit transfers. 26.6.4 Enhanced Resolution Mode The Enhanced Resolution Mode is automatically enabled when 11-bit or 12-bit mode is selected in the ADC Configuration Register (ACR). In this mode the ADCIFB will trade conversion performance for accuracy by averaging multiple samples. To be able to increase the accuracy by averaging multiple samples it is important that some noise is present in the input signal. The noise level should be between one and two LSB peakto-peak to get good averaging performance. 591 32099I–01/2012 AT32UC3L016/32/64 The performance cost of enabling 11-bit mode is 4 ADC samples, which reduces the effective ADC performance by a factor 4. For 12-bit mode this factor is 16. For 12-bit mode the effective sample rate is maximum ADC sample rate divided by 16. 26.6.5 ADC Clock The ADCIFB generates an internal clock named CLK_ADC that is used by the Analog-to-Digital Converter cell to perform conversions. The CLK_ADC frequency is selected by writing to the PRESCAL field in the ADC Configuration Register (ACR). The CLK_ADC range is between CLK_ADCIFB/2, if PRESCAL is 0, and CLK_ADCIFB/128, if PRESCAL is 63 (0x3F). A sensible PRESCAL value must be used in order to provide an ADC clock frequency according to the maximum sampling rate parameter given in the Electrical Characteristics section. Failing to do so may result in incorrect Analog-to-Digital Converter operation. 26.6.6 ADC Sleep Mode The ADC Sleep Mode maximizes power saving by automatically deactivating the Analog-to-Digital Converter cell when it is not being used for conversions. The ADC Sleep Mode is enabled by writing a one to the SLEEP bit in the ADC Configuration Register (ACR). When a trigger occurs while the ADC Sleep Mode is enabled, the Analog-to-Digital Converter cell is automatically activated. As the analog cell requires a startup time, the logic waits during this time and then starts the conversion of the enabled channels. When conversions of all enabled channels are complete, the ADC is deactivated until the next trigger. 26.6.7 Startup Time The Analog-to-Digital Converter cell has a minimal startup time when the cell is activated. This startup time is given in the Electrical Characteristics chapter and must be written to the STARTUP field in the ADC Configuration Register (ACR) to get correct conversion results. The STARTUP field expects the startup time to be represented as the number of CLK_ADC cycles between 8 and 1024 and in steps of 8 that is needed to cover the ADC startup time as specified in the Electrical Characteristics chapter. The Analog-to-Digital Converter cell is activated at the first conversion after reset and remains active if ACR.SLEEP is zero. If ACR.SLEEP is one, the Analog-to-Digital Converter cell is automatically deactivated when idle and thus each conversion sequence will have a initial startup time delay. 26.6.8 Sample and Hold Time A minimal Sample and Hold Time is necessary for the ADCIFB to guarantee the best converted final value when switching between ADC channels. This time depends on the input impedance of the analog input, but also on the output impedance of the driver providing the signal to the analog input, as there is no input buffer amplifier. The Sample and Hold time has to be programmed through the SHTIM field in the ADC Configuration Register (ACR). This field can define a Sample and Hold time between 1 and 16 CLK_ADC cycles. 26.6.9 ADC Conversion ADC conversions are performed on all enabled channels when a trigger condition is detected. For details regarding trigger conditions see Section 26.8.1. The term channel is used to identify a specific analog input pin so it can be included or excluded in an Analog-to-Digital conversion 592 32099I–01/2012 AT32UC3L016/32/64 sequence and to identify which AD pin was used to convert the current value in the Last Converted Data Register (LCDR). Channel number N corresponding to AD pin number N. Channels are enabled by writing a one to the corresponding bit in the Channel Enable Register (CHER), and disabled by writing a one to the corresponding bit in the Channel Disable Register (CHDR). Active channels are listed in the Channel Status Register (CHSR). When a conversion sequence is started, all enabled channels will be converted in one sequence and the result will be placed in the Last Converted Data Register (LCDR) with the channel number used to produce the result. It is important to read out the results while the conversion sequence is ongoing, as new values will automatically overwrite any old value and the old value will be lost if not previously read by the user. If the Analog-to-Digital Converter cell is inactive when starting a conversion sequence, the conversion logic will wait a configurable number of CLK_ADC cycles as defined in the startup time field in the ADC Configuration Register (ACR). After the cell is activated all enabled channels is converted one by one until no more enabled channels exist. The conversion sequence converts each enabled channel in order starting with the channel with the lowest channel number. If the ACR.SLEEP bit is one, the Analog-to-Digital Converter cell is deactivated after the conversion sequence has finished. For each channel converted, the ADCIFB waits a Sample and Hold number of CLK_ADC cycles as defined in the SHTIM field in ACR, and then instructs the Analog-to-Digital Converter cell to start converting the analog voltage. The ADC cell requires 10 CLK_ADC cycles to actually convert the value, so the total time to convert a channel is Sample and Hold + 10 CLK_ADC cycles. 26.6.10 Analog Compare Mode The ADCIFB can test if the converted values, as they become available, are below, above, or inside a specified range and generate interrupt requests based on this information. This is useful for applications where the user wants to monitor some external analog signal and only initiate actions if the value is above, below, or inside some specified range. The Analog Compare mode is enabled by writing a one to the Analog Compare Enable (ACE) bit in the Mode Register (MR). The values to compare must be written to the Low Value (LV) field and the High Value (HV) field in the Compare Value Register (CVR). The Analog Compare mode will, when enabled, check all enabled channels against the pre-programmed high and low values and set status bits. To generate an interrupt request if a converted value is below a limit, write the limit to the CVR.LV field and enable interrupt request on the Compare Lesser Than (CLT) bit by writing a one to the corresponding bit in the Interrupt Enable Register (IER). To generate an interrupt request if a converted value is above a limit, write the limit to the CVR.HV field and enable interrupt for Compare Greater Than (CGT) bit. To generate an interrupt request if a converted value is inside a range, write the low and high limit to the LV and HV fields and enable the Compare Else (CELSE) interrupt. To generate an interrupt request if a value is outside a range, write the LV and HV fields to the low and high limits of the range and enable CGT and CLT interrupts. Note that the values written to LV and HV must match the resolution selected in the ADC Configuration Register (ACR). 26.6.11 Interrupt Operation Interrupt requests are enabled by writing a one to the corresponding bit in the Interrupt Enable Register (IER) and disabled by writing a one to the corresponding bit in the Interrupt Disable 593 32099I–01/2012 AT32UC3L016/32/64 Register (IDR). Enabled interrupts can be read from the Interrupt Mask Register (IMR). Active interrupt requests, but potentially masked, are visible in the Interrupt Status Register (ISR). To clear an active interrupt request, write a one to the corresponding bit in the Interrupt Clear Register (ICR). The source for the interrupt requests are the status bits in the Status Register (SR). The SR shows the ADCIFB status at the time the register is read. The Interrupt Status Register (ISR) shows the status since the last write to the Interrupt Clear Register. The combination of ISR and SR allows the user to react to status change conditions but also allows the user to read the current status at any time. 26.6.12 Peripheral Events The Peripheral Event System can be used together with the ADCIFB to allow any peripheral event generator to be used as a trigger source. To enable peripheral events to trigger a conversion sequence the user must write the Peripheral Event Trigger value (0x7) to the Trigger Mode (TRGMOD) field in the Trigger Register (TRGR). Refer to Table 26-4 on page 606. The user must also configure a peripheral event generator to emit peripheral events for the ADCIFB to trigger on. Refer to the Peripheral Event System chapter for details. 26.6.13 Sleep Mode Before entering sleep modes the user must make sure the ADCIFB is idle and that the Analogto-Digital Converter cell is inactive. To deactivate the Analog-to-Digital Converter cell the SLEEP bit in the ADC Configuration Register (ACR) must be written to one and the ADCIFB must be idle. To make sure the ADCIFB is idle, write a zero the Trigger Mode (TRGMOD) field in the Trigger Register (TRGR) and wait for the READY bit in the Status Register (SR) to be set. Note that by deactivating the Analog-to-Digital Converter cell, a startup time penalty as defined in the STARTUP field in the ADC Configuration Register (ACR) will apply on the next conversion. 26.6.14 26.7 Conversion Performances For performance and electrical characteristics of the ADCIFB, refer to the Electrical Characteristics chapter. Resistive Touch Screen The ADCIFB embeds an integrated Resistive Touch Screen Sequencer that can be used to calculate contact coordinates on a resistive touch screen film. When instructed to start, the integrated Resistive Touch Screen Sequencer automatically applies a sequence of voltage patterns to the resistive touch screen films and the Analog-to-Digital Conversion cell is used to measure the effects. The resulting measurements can be used to calculate the horizontal and vertical contact coordinates. It is recommended to use a high resistance touch screen for optimal resolution. The resistive touch screen film is connected to the ADCIFB using the AD and ADP pins. See Section 26.7.3 for details. Resistive Touch Screen Mode is enabled by writing a one to the Touch Screen ADC Mode field in the Mode Register (MR.TSAMOD). In this mode, channels TSPO+0 though TSPO+3 are automatically enabled where TSPO refers to the Touch Screen Pin Offset field in the Mode Register (MR.TSPO). For each conversion sequence, all enabled channels before TSPO+0 and after TSPO+3 are converted as ordinary ADC channels, producing 1 conversion result each. 594 32099I–01/2012 AT32UC3L016/32/64 When the sequencer enters the TSPO+0 channel the Resistive Touch Screen Sequencer will take over control and convert the next 4 channels as described in Section 26.7.4. 26.7.1 Resistive Touch Screen Principles A resistive touch screen is based on two resistive films, each one fitted with a pair of electrodes, placed at the top and bottom on one film, and on the right and left on the other. Between the two, there is a layer that acts as an insulator, but makes a connection when pressure is applied to the screen. This is illustrated in Figure 26-2 on page 595. Figure 26-2. Resistive Touch Screen Position Measurement Pen Contact XP YM YP XM VDD XP YP XP Volt XM Vertical Position Detection 26.7.2 VDD YP GND Volt YM GND Horizontal Position Detection Position Measurement Method As shown in Figure 26-2 on page 595, to detect the position of a contact, voltage is first applied to XP (top) and Xm (bottom) leaving Yp and Ym tristated. Due to the linear resistance of the film, there is a voltage gradient from top to bottom on the first film. When a contact is performed on the screen, the voltage at the contact point propagates to the second film. If the input impedance on the YP (right) and Ym (left) electrodes are high enough, no current will flow, allowing the voltage at the contact point to be measured at Yp. The value measured represents the vertical position component of the contact point. For the horizontal direction, the same method is used, but by applying voltage from YP (right) to Ym (left) and measuring at XP. In an ideal world (linear, with no loss), the vertical position is equal to: VYP / VDD To compensate for some of the real world imperfections, VXP and VXm can be measured and used to improve accuracy at the cost of two more conversions per axes. The new expression for the vertical position then becomes: (VYP - VXM) / (VXP - VXM) 595 32099I–01/2012 AT32UC3L016/32/64 26.7.3 Resistive Touch Screen Pin Connections Table 26-2. Resistive Touch Screen Pin Connections ADCIFB Pin TS Signal, APOE == 0 TS Signal, APOE == 1 ADP0 Xp through a resistor No Connect ADP1 Yp through a resistor No Connect ADtspo+0 Xp Xp ADtspo+1 Xm Xm ADtspo+2 Yp Yp ADtspo+3 Ym Ym The resistive touch screen film signals connects to the ADCIFB using the AD and ADP pins. The XP (top) and XM (bottom) film signals are connected to ADtspo+0 and ADtspo+1 pins, and the YP (right) and YM (left) signals are connected to ADtspo+2 and ADtspo+3 pins. The tspo index is configurable through the Touch Screen Pin Offset (TSPO) field in the Mode Register (MR) and allows the user to configure which AD pins to use for resistive touch screen applications. Writing a zero to the TSPO field instructs the ADCIFB to use AD0 through AD3, where AD0 is connected to XP, AD1 is connected to XM and so on. Writing a one to the TSPO field instructs the ADCIFB to use AD1 through AD4 for resistive touch screen sequencing, where AD1 is connected to XP and AD0 is free to be used as an ordinary ADC channel. When the Analog Pin Output Enable (APOE) bit in the Mode Register (MR) is zero, the AD pins are used to measure input voltage and drive the GND sequences, while the ADP pins are used to drive the VDD sequences. This arrangement allows the user to reduce the voltage seen at the AD input pins by inserting external resistors between ADP0 and XP and ADP1 and YP signals which are again directly connected to the AD pins. It is important that the voltages observed at the AD pins are not higher than the maximum allowed ADC input voltage. See Figure 26-3 on page 597 for details regarding how to connect the resistive touch screen films to the AD and ADP pins. By adding a resistor between ADP0 and XP, and ADP1 and YP, the maximum voltage observed at the AD pins can be controlled by the following voltage divider expressions: R filmx V ( AD tspo + 0 ) = --------------------------------------------- ⋅ V ( DP 0 ) R filmx + R resistorx The Rfilmx parameter is the film resistance observed when measuring between XP and XM. The Rresistorx parameter is the resistor size inserted between ADP0 and XP. The definition of Rfilmy and Rresistory is the same but for ADP1, YP, and YM instead. 596 32099I–01/2012 AT32UC3L016/32/64 R filmy V ( AD tspo + 2 ) = --------------------------------------------- ⋅ V ( DP 1 ) R filmy + R resistory The ADP pins are used by default, as the APOE bit is zero after reset. Writing a one to the APOE bit instructs the ADCIFB Resistive Touch Screen Sequencer to use the already connected ADtspo+0 and ADtspo+2 pins to drive VDD to XP and YP signals directly. In this mode the ADP pins can be used as general purpose I/O pins. Before writing a one to the APOE bit the user must make sure that the I/O voltage is compatible with the ADC input voltage. If the I/O voltage is higher than the maximum input voltage of the ADC, permanent damage may occur. Refer to the Electrical Characteristics chapter for details. Figure 26-3. Resistive Touch Screen Pin Connections DP0 Analog Pin Output Enable (MR.APOE) == 0 DP1 ADtspo+0 XP YM ADtspo+1 YP ADtspo+2 XM ADtspo+3 Analog Pin Output Enable (MR.APOE) == 1 NC DP0 NC DP1 ADtspo+0 XP YM ADtspo+1 YP ADtspo+2 XM ADtspo+3 597 32099I–01/2012 AT32UC3L016/32/64 26.7.4 Resistive Touch Screen Sequencer The Resistive Touch Screen Sequencer is responsible for applying voltage to the resistive touch screen films as described in Section 26.7.2. This is done by controlling the output enable and the output value of the ADP and AD pins. This allows the Resistive Touch Screen Sequencer to add a voltage gradient on one film while keeping the other film floating so a touch can be measured. The Resistive Touch Screen Sequencer will when measuring the vertical position, apply VDD and GND to the pins connected to XP and XM. The YP and YM pins are put in tristate mode so the measurement of YP can proceed without interference. To compensate for ADC offset errors and non ideal pad drivers, the actual voltage of XP and XM is measured as well, so the real values for VDD and GND can be used in the contact point calculation to increase accuracy. See second formula in Section 26.7.2. When the vertical values are converted the same setup is applies for the second axes, by setting XP and XM in tristate mode and applying VDD and GND to YP and YM. Refer to Section 26.8.3 for details. 26.7.5 Pen Detect If no contact is applied to the resistive touch screen films, any resistive touch screen conversion result will be undefined as the film being measured is floating. This can be avoided by enabling Pen Detect and only trigger resistive touch screen conversions when the Pen Contact (PENCNT) status bit in the Status Register (SR) is one. Pen Detect is enabled by writing a one to the Pen Detect (PENDET) bit in the Mode Register (MR). When Pen Detect is enabled, the ADCIFB grounds the vertical panel by applying GND to XP and XM and polarizes the horizontal panel by enabling pull-up on the pin connected to YP. The YM pin will in this mode be tristated. Since there is no contact, no current is flowing and there is no related power consumption. As soon as a contact occurs, GND will propagate to YM by pulling down YP, allowing the contact to be registered by the ADCIFB. A programmable debouncing filter can be used to filter out false pen detects because of noise. The debouncing filter is programmable from one CLK_ADC period and up to 215 CLK_ADC periods. The debouncer length is set by writing to the PENDBC field in MR. 598 32099I–01/2012 AT32UC3L016/32/64 Figure 26-4. Resistive Touch Screen Pen Detect Resistive Touch Screen Sequencer XP GND XM GND YP Pullup YM Tristate T o the ADC PENDBC D ebouncer Pen Int er r upt The Resistive Touch Screen Pen Detect can be used to generate an ADCIFB interrupt request or it can be used to trig a conversion, so that a position can be measured as soon as a contact is detected. The Pen Detect Mode generates two types of status signals, reported in the Status Register (SR): • The bit PENCNT is set when current flows and remains set until current stops. • The bit NOCNT is set when no current flows and remains set until current flows. Before a current change is reflected in the SR, the new status must be stable for the duration of the debouncing time. Both status conditions can generate an interrupt request if the corresponding bit in the Interrupt Mask Register (IMR) is one. Refer to Section 26.6.11 on page 593. 599 32099I–01/2012 AT32UC3L016/32/64 26.8 Operating Modes The ADCIFB features two operating modes, each defining a separate conversion sequence: • ADC Mode: At each trigger, all the enabled channels are converted. • Resistive Touch Screen Mode: At each trigger, all enabled channels plus the resistive touch screen channels are converted as described in Section 26.8.3. If channels except the dedicated resistive touch screen channels are enabled, they are converted normally before and after the resistive touch screen channels are converted. The operating mode is selected by the TSAMOD field in the Mode Register (MR). 26.8.1 Conversion Triggers A conversion sequence is started either by a software or by a hardware trigger. When a conversion sequence is started, all enabled channels will be converted and made available in the shared Last Converted Register (LCDR). The software trigger is asserted by writing a one to the START field in the Control Register (CR). The hardware trigger can be selected by the TRGMOD field in the Trigger Register (TRGR). Different hardware triggers exist: • External trigger, either rising or falling or any, detected on the external trigger pin TRIGGER • Pen detect trigger, depending the PENDET bit in the Mode Register (MR) • Continuous trigger, meaning the ADCIFB restarts the next sequence as soon as it finishes the current one • Periodic trigger, which is defined by the TRGR.TRGPER field • Peripheral event trigger, allowing the Peripheral Event System to synchronize conversion with some configured peripheral event source. Enabling a hardware trigger does not disable the software trigger functionality. Thus, if a hardware trigger is selected, the start of a conversion can still be initiated by the software trigger. 26.8.2 ADC Mode In the ADC Mode, the active channels are defined by the Channel Status Register (CHSR). A channel is enabled by writing a one to the corresponding bit in the Channel Enable Register (CHER), and disabled by writing a one to the corresponding bit in the Channel Disable Register (CHDR). The conversion results are stored in the Last Converted Data Register (LCDR) as they become available, overwriting old conversions. At each trigger, the following sequence is performed: 1. If ACR.SLEEP is one, wake up the ADC and wait for the startup time. 2. If Channel 0 is enabled, convert Channel 0 and store result in LCDR. 3. If Channel 1 is enabled, convert Channel 1 and store result in LCDR. 4. If Channel N is enabled, convert Channel N and store result in LCDR. 5. If ACR.SLEEP is one, place the ADC cell in a low-power state. If the Peripheral DMA Controller is enabled, all converted values are transferred continuously into the memory buffer. 26.8.3 Resistive Touch Screen Mode Writing a one to the TSAMOD field in the Mode Register (MR) enables Resistive Touch Screen Mode. In this mode the channels TSPO+0 to TSPO+3, corresponding to the resistive touch 600 32099I–01/2012 AT32UC3L016/32/64 screen inputs, are automatically activated. In addition, if any other channels are enabled, they will be converted before and after the resistive touch screen conversion. At each trigger, the following sequence is performed: 1. If ACR.SLEEP is one, wake up the ADC cell and wait for the startup time. 2. Convert all enabled channels before TSPO and store the results in the LCDR. 3. Apply supply on the inputs XP and XM during the Sample and Hold Time. 4. Convert Channel XM and store the result in TMP. 5. Apply supply on the inputs XP and XM during the Sample and Hold Time. 6. Convert Channel XP, subtract TMP from the result and store the subtracted result in LCDR. 7. Apply supply on the inputs XP and XM during the Sample and Hold Time. 8. Convert Channel YP, subtract TMP from the result and store the subtracted result in LCDR. 9. Apply supply on the inputs YP and YM during the Sample and Hold Time. 10. Convert Channel YM and store the result in TMP. 11. Apply supply on the inputs YP and YM during the Sample and Hold Time. 12. Convert Channel YP, subtract TMP from the result and store the subtracted result in LCDR. 13. Apply supply on the inputs YP and YM during the Sample and Hold Time. 14. Convert Channel XP, subtract TMP from the result and store the subtracted result in LCDR. 15. Convert all enabled channels after TSPO + 3 and store results in the LCDR. 16. If ACR.SLEEP is one, place the ADC cell in a low-power state. The resulting buffer structure stored in memory is: 1. XP - XM 2. YP - XM 3. YP - YM 4. XP - YM. The vertical position can be easily calculated by dividing the data at offset 1(XP - XM) by the data at offset 2(YP - XM). The horizontal position can be easily calculated by dividing the data at offset 3(YP - YM) by the data at offset 4(XP - YM). 601 32099I–01/2012 AT32UC3L016/32/64 26.9 User Interface Table 26-3. Note: ADCIFB Register Memory Map Offset Register Name Access Reset 0x00 Control Register CR Write-only - 0x04 Mode Register MR Read/Write 0x00000000 0x08 ADC Configuration Register ACR Read/Write 0x00000000 0x0C Trigger Register TRGR Read/Write 0x00000000 0x10 Compare Value Register CVR Read/Write 0x00000000 0x14 Status Register SR Read-only 0x00000000 0x18 Interrupt Status Register ISR Read-only 0x00000000 0x1C Interrupt Clear Register ICR Write-only - 0x20 Interrupt Enable Register IER Write-only - 0x24 Interrupt Disable Register IDR Write-only - 0x28 Interrupt Mask Register IMR Read-only 0x00000000 0x2C Last Converted Data Register LCDR Read-only 0x00000000 0x30 Parameter Register PARAMETER Read-only -(1) 0x34 Version Register VERSION Read-only -(1) 0x40 Channel Enable Register CHER Write-only - 0x44 Channel Disable Register CHDR Write-only - 0x48 Channel Status Register CHSR Read-only 0x00000000 1. The reset values for these registers are device specific. Please refer to the Module Configuration section at the end of this chapter. 602 32099I–01/2012 AT32UC3L016/32/64 26.9.1 Control Register Register Name: CR Access Type: Write-only Offset: 0x00 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - DIS EN 7 6 5 4 3 2 1 0 - - - - - - START SWRST • DIS: ADCDIFB Disable Writing a zero to this bit has no effect. Writing a one to this bit disables the ADCIFB. Note: Disabling the ADCIFB effectively stops all clocks in the module so the user must make sure the ADCIFB is idle before disabling the ADCIFB. • EN: ADCIFB Enable Writing a zero to this bit has no effect. Writing a one to this bit enables the ADCIFB. Note: The ADCIFB must be enabled before use. • START: Start Conversion Writing a zero to this bit has no effect. Writing a one to this bit starts an Analog-to-Digital conversion. • SWRST: Software Reset Writing a zero to this bit has no effect. Writing a one to this bit resets the ADCIFB, simulating a hardware reset. 603 32099I–01/2012 AT32UC3L016/32/64 26.9.2 Name: Mode Register MR Access Type: Read/Write Offset: 0x04 Reset Value: 0x00000000 31 30 29 28 PENDBC 23 22 21 20 27 26 25 24 - - - - 19 18 17 16 TSPO 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - APOE ACE PENDET - - - TSAMOD • PENDBC: Pen Detect Debouncing Period Period = 2PENDBC*TCLK_ADC • TSPO: Touch Screen Pin Offset The Touch Screen Pin Offset field is used to indicate which AD pins are connected to the resistive touch screen film edges. Only an offset is specified and it is assumed that the resistive touch screen films are connected sequentially from the specified offset pin and up to and including offset + 3 (4 pins). • APOE: Analog Pin Output Enable 0: AD pins are not used to drive VDD in resistive touch screen sequence. 1: AD pins are used to drive VDD in resistive touch screen sequence. Note: If the selected I/O voltage configuration is incompatible with the Analog-to-Digital converter cell voltage specification, this bit must stay cleared to avoid damaging the ADC. In this case the ADP pins must be used to drive VDD instead, as described in Section 26.7.3. If the I/O and ADC voltages are compatible, the AD pins can be used directly by writing a one to this bit. In this case the ADP pins can be ignored. • ACE: Analog Compare Enable 0: The analog compare functionality is disabled. 1: The analog compare functionality is enabled. • PENDET: Pen Detect 0: The pen detect functionality is disabled. 1: The pen detect functionality is enabled. Note: Touch detection logic can only be enabled when the ADC sequencer is idle. For successful pen detection the user must make sure there is enough idle time between consecutive scans for the touch detection logic to settle. • TSAMOD: Touch Screen ADC Mode 0: Touch Screen Mode is disabled. 1: Touch Screen Mode is enabled. 604 32099I–01/2012 AT32UC3L016/32/64 26.9.3 Name: ADC Configuration Register ACR Access Type: Read/Write Offset: 0x08 Reset Value: 0x00000000 31 30 29 28 - - - - 23 22 21 20 - 27 26 25 24 18 17 16 10 9 8 3 2 1 0 - - - SLEEP SHTIM 19 STARTUP 15 14 - - 7 6 - - 13 12 11 PRESCAL 5 4 RES • SHTIM: Sample & Hold Time for ADC Channels T SAMPLE&HOLD = ( SHTIM + 3 ) ⋅ T CLK_ADC • STARTUP: Startup Time TARTUP = ( STARTUP + 1 ) ⋅ 8 ⋅ T CLK_AD • PRESCAL: Prescaler Rate Selection T CLK_ADC = ( PRESCAL + 1 ) ⋅ 2 ⋅ T CLK_ADCIFB • RES: Resolution Selection 0: 8-bit resolution. 1: 10-bit resolution. 2: 11-bit resolution, interpolated. 3: 12-bit resolution, interpolated. • SLEEP: ADC Sleep Mode 0: ADC Sleep Mode is disabled. 1: ADC Sleep Mode is enabled. 605 32099I–01/2012 AT32UC3L016/32/64 26.9.4 Name: Trigger Register TRGR Access Type: Read/Write Offset: 0x0C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 TRGPER[15:8] 23 22 21 20 TRGPER[7:0] 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - TRGMOD • TRGPER: Trigger Period Effective only if TRGMOD defines a Periodic Trigger. Defines the periodic trigger period, with the following equations: Trigger Period = TRGPER *TCLK_ADC • TRGMOD: Trigger Mode Table 26-4. Trigger Modes TRGMOD Selected Trigger Mode 0 0 0 No trigger, only software trigger can start conversions 0 0 1 External Trigger Rising Edge 0 1 0 External Trigger Falling Edge 0 1 1 External Trigger Any Edge 1 0 0 Pen Detect Trigger (shall be selected only if PENDET is set and TSAMOD = Touch Screen mode) 1 0 1 Periodic Trigger (TRGPER shall be initiated appropriately) 1 1 0 Continuous Mode 1 1 1 Peripheral Event Trigger 606 32099I–01/2012 AT32UC3L016/32/64 26.9.5 Name: Compare Value Register CVR Access Type: Read/Write Offset: 0x10 Reset Value: 0x00000000 31 30 29 28 - - - - 23 22 21 20 27 26 25 24 HV[11:8] 19 18 17 16 11 10 9 8 1 0 HV[7:0] 15 14 13 12 - - - - 7 6 5 4 LV[11:8] 3 2 LV[7:0] • HV: High Value Defines the high value used when comparing analog input. • LV: Low Value Defines the low value used when comparing analog input. 607 32099I–01/2012 AT32UC3L016/32/64 26.9.6 Name: Status Register SR Access Type: Read-only Offset: 0x14 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - EN 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - CELSE CGT CLT - - BUSY READY 7 6 5 4 3 2 1 0 - - NOCNT PENCNT - - OVRE DRDY • EN: Enable Status 0: The ADCIFB is disabled. 1: The ADCIFB is enabled. This bit is cleared when CR.DIS is written to one. This bit is set when CR.EN is written to one. • CELSE: Compare Else Status This bit is cleared when either CLT or CGT are detected or when analog compare is disabled. This bit is set when no CLT or CGT are detected on the last converted data and analog compare is enabled. • CGT: Compare Greater Than Status This bit is cleared when no compare greater than CVR.HV is detected on the last converted data or when analog compare is disabled. This bit is set when compare greater than CVR.HV is detected on the last converted data and analog compare is enabled. • CLT: Compare Lesser Than Status This bit is cleared when no compare lesser than CVR.LV is detected on the last converted data or when analog compare is disabled. This bit is set when compare lesser than CVR.LV is detected on the last converted data and analog compare is enabled. • BUSY: Busy Status This bit is cleared when the ADCIFB is ready to perform a conversion sequence. This bit is set when the ADCIFB is busy performing a convention sequence. • READY: Ready Status This bit is cleared when the ADCIFB is busy performing a conversion sequence This bit is set when the ADCIFB is ready to perform a conversion sequence. • NOCNT: No Contact Status This bit is cleared when no contact loss is detected or pen detect is disabled This bit is set when contact loss is detected and pen detect is enabled. • PENCNT: Pen Contact Status This bit is cleared when no contact is detected or pen detect is disabled. 608 32099I–01/2012 AT32UC3L016/32/64 This bit is set when pen contact is detected and pen detect is enabled. • OVRE: Overrun Error Status This bit is cleared when no Overrun Error has occurred since the start of a conversion sequence. This bit is set when one or more Overrun Error has occurred since the start of a conversion sequence. • DRDY: Data Ready Status 0: No data has been converted since the last reset. 1: One or more conversions have completed since the last reset and data is available in LCDR. This bit is cleared when CR.SWRST is written to one. This bit is set when one or more conversions have completed and data is available in LCDR. 609 32099I–01/2012 AT32UC3L016/32/64 26.9.7 Name: Interrupt Status Register ISR Access Type: Read-only Offset: 0x18 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - CELSE CGT CLT - - BUSY READY 7 6 5 4 3 2 1 0 - - NOCNT PENCNT - - OVRE DRDY • CELSE: Compare Else Status This bit is cleared when the corresponding bit in ICR is written to one. This bit is set when the corresponding bit in SR has a zero-to-one transition. • CGT: Compare Greater Than Status This bit is cleared when the corresponding bit in ICR is written to one. This bit is set when the corresponding bit in SR has a zero-to-one transition. • CLT: Compare Lesser Than Status This bit is cleared when the corresponding bit in ICR is written to one. This bit is set when the corresponding bit in SR has a zero-to-one transition. • BUSY: Busy Status This bit is cleared when the corresponding bit in ICR is written to one. This bit is set when the corresponding bit in SR has a zero-to-one transition. • READY: Ready Status This bit is cleared when the corresponding bit in ICR is written to one. This bit is set when the corresponding bit in SR has a zero-to-one transition. • NOCNT: No Contact Status This bit is cleared when the corresponding bit in ICR is written to one. This bit is set when the corresponding bit in SR has a zero-to-one transition. • PENCNT: Pen Contact Status This bit is cleared when the corresponding bit in ICR is written to one. This bit is set when the corresponding bit in SR has a zero-to-one transition. • OVRE: Overrun Error Status This bit is cleared when the corresponding bit in ICR is written to one. This bit is set when the corresponding bit in SR has a zero-to-one transition. • DRDY: Data Ready Status This bit is cleared when the corresponding bit in ICR is written to one. This bit is set when a conversion has completed and new data is available in LCDR. 610 32099I–01/2012 AT32UC3L016/32/64 26.9.8 Name: Interrupt Clear Register ICR Access Type: Write-only Offset: 0x1C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - CELSE CGT CLT - - BUSY READY 7 6 5 4 3 2 1 0 - - NOCNT PENCNT - - OVRE DRDY Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in ISR and the corresponding interrupt request. 611 32099I–01/2012 AT32UC3L016/32/64 26.9.9 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x20 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - CELSE CGT CLT - - BUSY READY 7 6 5 4 3 2 1 0 - - NOCNT PENCNT - - OVRE DRDY Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 612 32099I–01/2012 AT32UC3L016/32/64 26.9.10 Name: Interrupt Disable Register IDR Access Type: Write-only Offset: 0x24 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - CELSE CGT CLT - - BUSY READY 7 6 5 4 3 2 1 0 - - NOCNT PENCNT - - OVRE DRDY Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 613 32099I–01/2012 AT32UC3L016/32/64 26.9.11 Name: Interrupt Mask Register IMR Access Type: Read-only Offset: 0x28 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - CELSE CGT CLT - - BUSY READY 7 6 5 4 3 2 1 0 - - NOCNT PENCNT - - OVRE DRDY 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. A bit in this register is cleared by writing a one to the corresponding bit in Interrupt Disable Register (IDR). A bit in this register is set by writing a one to the corresponding bit in Interrupt Enable Register (IER). 614 32099I–01/2012 AT32UC3L016/32/64 26.9.12 Name: Last Converted Data Register LCDR Access Type: Read-only Offset: 0x2C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 11 10 9 8 1 0 LCCH 15 14 13 12 - - - - 7 6 5 4 LDATA[11:8] 3 2 LDATA[7:0] • LCCH: Last Converted Channel This field indicates what channel was last converted, i.e. what channel the LDATA represents. • LDATA: Last Data Converted The analog-to-digital conversion data is placed in this register at the end of a conversion on any analog channel and remains until a new conversion on any analog channel is completed. 615 32099I–01/2012 AT32UC3L016/32/64 26.9.13 Name: Parameter Register PARAMETER Access Type: Read-only Offset: 0x30 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 CH31 CH30 CH29 CH28 CH27 CH26 CH25 CH24 23 22 21 20 19 18 17 16 CH23 CH22 CH21 CH20 CH19 CH18 CH17 CH16 15 14 13 12 11 10 9 8 CH15 CH14 CH13 CH12 CH11 CH10 CH9 CH8 7 6 5 4 3 2 1 0 CH7 CH6 CH5 CH4 CH3 CH2 CH1 CH0 • CHn: Channel n Implemented 0: The corresponding channel is not implemented. 1: The corresponding channel is implemented. 616 32099I–01/2012 AT32UC3L016/32/64 26.9.14 Name: Version Register VERSION Access Type: Read-only Offset: 0x34 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant Number Reserved. No functionality associated. • VERSION: Version Number Version number of the Module. No functionality associated. 617 32099I–01/2012 AT32UC3L016/32/64 26.9.15 Name: Channel Enable Register CHER Access Type: Write-only Offset: 0x40 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 CH31 CH30 CH29 CH28 CH27 CH26 CH25 CH24 23 22 21 20 19 18 17 16 CH23 CH22 CH21 CH20 CH19 CH18 CH17 CH16 15 14 13 12 11 10 9 8 CH15 CH14 CH13 CH12 CH11 CH10 CH9 CH8 7 6 5 4 3 2 1 0 CH7 CH6 CH5 CH4 CH3 CH2 CH1 CH0 • CHn: Channel n Enable Writing a zero to a bit in this register has no effect Writing a one to a bit in this register enables the corresponding channel The number of available channels is device dependent. Please refer to the Module Configuration section at the end of this chapter for information regarding which channels are implemented. 618 32099I–01/2012 AT32UC3L016/32/64 26.9.16 Name: Channel Disable Register CHDR Access Type: Write-only Offset: 0x44 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 CH31 CH30 CH29 CH28 CH27 CH26 CH25 CH24 23 22 21 20 19 18 17 16 CH23 CH22 CH21 CH20 CH19 CH18 CH17 CH16 15 14 13 12 11 10 9 8 CH15 CH14 CH13 CH12 CH11 CH10 CH9 CH8 7 6 5 4 3 2 1 0 CH7 CH6 CH5 CH4 CH3 CH2 CH1 CH0 • CHn: Channel N Disable Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register disables the corresponding channel. Warning: If the corresponding channel is disabled during a conversion, or if it is disabled and then re-enabled during a conversion, its associated data and its corresponding DRDY and OVRE bits in SR are unpredictable. The number of available channels is device dependent. Please refer to the Module Configuration section at the end of this chapter for information regarding how many channels are implemented. 619 32099I–01/2012 AT32UC3L016/32/64 26.9.17 Name: Channel Status Register CHSR Access Type: Read-only Offset: 0x48 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 CH31 CH30 CH29 CH28 CH27 CH26 CH25 CH24 23 22 21 20 19 18 17 16 CH23 CH22 CH21 CH20 CH19 CH18 CH17 CH16 15 14 13 12 11 10 9 8 CH15 CH14 CH13 CH12 CH11 CH10 CH9 CH8 7 6 5 4 3 2 1 0 CH7 CH6 CH5 CH4 CH3 CH2 CH1 CH0 • CHn: Channel N Status 0: The corresponding channel is disabled. 1: The corresponding channel is enabled. A bit in this register is cleared by writing a one to the corresponding bit in Channel Disable Register (CHDR). A bit in this register is set by writing a one to the corresponding bit in Channel Enable Register (CHER). The number of available channels is device dependent. Please refer to the Module Configuration section at the end of this chapter for information regarding how many channels are implemented. 620 32099I–01/2012 AT32UC3L016/32/64 26.10 Module Configuration The specific configuration for each ADCIFB instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 26-5. Module Configuration Feature ADCIFB Number of ADC channels 9 (8 + 1 internal temperature sensor channel) Table 26-6. ADCIFB Clocks Clock Name Description CLK_ADCIFB Clock for the ADCIFB bus interface Table 26-7. Register Reset Values Register Reset Value VERSION 0x00000101 PARAMETER 0x000003FF Table 26-8. ADC Input Channels Channel Input CH0 AD0 CH1 AD1 CH2 AD2 CH4 AD4 CH5 AD5 CH6 AD6 CH7 AD7 CH8 AD8 CH9 Temperature sensor 621 32099I–01/2012 AT32UC3L016/32/64 27. Analog Comparator Interface (ACIFB) Rev: 2.0.2.2 27.1 Features • Controls an array of Analog Comparators • Low power option • • • • 27.2 – Single shot mode support Selectable settings for filter option – Filter length and hysteresis Window Mode – Detect inside/outside window – Detect above/below window Interrupt – On comparator result rising edge, falling edge, toggle – Inside window, outside window, toggle – When startup time has passed Can generate events to the peripheral event system Overview The Analog Comparator Interface (ACIFB) is able to control a number of Analog Comparators (AC) with identical behavior. An Analog Comparator compares two voltages and gives a compare output depending on this comparison. The ACIFB can be configured in normal mode using each comparator independently or in window mode using defined comparator pairs to observe a window. The number of channels implemented is device specific. Refer to the Module Configuration section at the end of this chapter for details. 622 32099I–01/2012 AT32UC3L016/32/64 27.3 Block Diagram Figure 27-1. ACIFB Block Diagram Analog Comparators ACP0 ACIFB TR.ACTEST0 CLK_ACIFB INP + INN ACOUT0 FILTER AC ACN0 GCLK INTERRUPT GENERATION CONF0.INSELN ……………... ACREFN ACPn Peripheral Bus INP IRQ PERIPHERAL EVENT GENERATION CTRL.ACTEST + ACOUTn INN EVENTS FILTER AC ACNn TRIGGER CONFn.INSELN TR.ACTESTn 27.4 I/O Lines Description There are two groups of analog comparators, A and B, as shown in Table 27-1. In normal mode, this grouping does not have any meaning. In window mode, two analog comparators, one from group A and the corresponding comparator from group B, are paired. Table 27-1. Analog Comparator Groups for Window Mode Group A Group B Pair Number AC0 AC1 0 AC2 AC3 1 AC4 AC5 2 AC6 AC7 3 Table 27-2. I/O Line Description Pin Name Pin Description Type ACAPn Positive reference pin for Analog Comparator A n Analog ACANn Negative reference pin for Analog Comparator A n Analog 623 32099I–01/2012 AT32UC3L016/32/64 Table 27-2. I/O Line Description Pin Name Pin Description Type ACBPn Positive reference pin for Analog Comparator B n Analog ACBNn Negative reference pin for Analog Comparator B n Analog ACREFN Reference Voltage for all comparators selectable for INN Analog The signal names corresponds to the groups A and B of analog comparators. For normal mode, the mapping from input signal names in the block diagram to the signal names is given in Table 27-3. Table 27-3. 27.5 Signal Name Mapping Pin Name Channel Number Normal Mode ACAP0/ACAN0 0 ACP0/ACN0 ACBP0/ACBN0 1 ACP1/ACN1 ACAP1/ACAN1 2 ACP2/ACN2 ACBP1/ACBN1 3 ACP3/ACN3 ACAP2/ACAN2 4 ACP4/ACN4 ACBP2/ACBN2 5 ACP5/ACN5 ACAP3/ACAN3 6 ACP6/ACN6 ACBP3/ACBN3 7 ACP7/ACN7 Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 27.5.1 I/O Lines The ACIFB pins are multiplexed with other peripherals. The user must first program the I/O Controller to give control of the pins to the ACIFB. 27.5.2 Power Management If the CPU enters a sleep mode that disables clocks used by the ACIFB, the ACIFB will stop functioning and resume operation after the system wakes up from sleep mode. 27.5.3 Clocks The clock for the ACIFB bus interface (CLK_ACIFB) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the ACIFB before disabling the clock, to avoid freezing the ACIFB in an undefined state. The ACIFB uses a GCLK as clock source for the Analog Comparators. The user must set up this GCLK at the right frequency. The CLK_ACIFB clock of the interface must be at least 4x the GCLK frequency used in the comparators. The GCLK is used both for measuring the startup time of a comparator, and to give a frequency for the comparisons done in Continuous Measurement Mode, see Section 27.6. Refer to the Electrical Characteristics chapter for GCLK frequency limitations. 624 32099I–01/2012 AT32UC3L016/32/64 27.5.4 Interrupts The ACIFB interrupt request line is connected to the interrupt controller. Using the ACIFB interrupt requires the interrupt controller to be programmed first. 27.5.5 Peripheral Events The ACIFB peripheral events are connected via the Peripheral Event System. Refer to the Peripheral Event System chapter for details. 27.5.6 Debug Operation When an external debugger forces the CPU into debug mode, the ACIFB continues normal operation. If the ACIFB is configured in a way that requires it to be periodically serviced by the CPU through interrupts or similar, improper operation or data loss may result during debugging. 27.6 Functional Description The ACIFB is enabled by writing a one to the Control Register Enable bit (CTRL.EN). Additionally, the comparators must be individually enabled by programming the MODE field in the AC Configuration Register (CONFn.MODE). The results from the individual comparators can either be used directly (normal mode), or the results from two comparators can be grouped to generate a comparison window (window mode). All comparators need not be in the same mode, some comparators may be in normal mode, while others are in window mode. There are restrictions on which AC channels that can be grouped together in a window pair, see Section 27.6.5. 27.6.1 Analog Comparator Operation Each AC channel can be in one of four different modes, determined by CONFn.MODE: • Off • Continuous Measurement Mode (CM) • User Triggered Single Measurement Mode (UT) • Event Triggered Single Measurement Mode (ET) After being enabled, a startup time defined in CTRL.SUT is required before the result of the comparison is ready. The GCLK is used for measuring the startup time of a comparator, During the startup time the AC output is not available. When the ACn Ready bit in the Status Register (SR.ACRDYn) is one, the output of ACn is ready. In window mode the result is available when both the comparator outputs are ready (SR.ACRDYn=1 and SR.ACRDYn+1=1). 27.6.1.1 Continuous Measurement Mode In CM, the Analog Comparator is continuously enabled and performing comparisons. This ensures that the result of the latest comparison is always available in the ACn Current Comparison Status bit in the Status Register (SR.ACCSn). Comparisons are done on every positive edge of GCLK. CM is enabled by writing CONFn.MODE to 1. After the startup time has passed, a comparison is done and SR is updated. Appropriate peripheral events and interrupts are also generated. New comparisons are performed continuously until the CONFn.MODE field is written to 0. 625 32099I–01/2012 AT32UC3L016/32/64 27.6.1.2 User Triggered Single Measurement Mode In the UT mode, the user starts a single comparison by writing a one to the User Start Single Comparison bit (CTRL.USTART). This mode is enabled by writing CONFn.MODE to 2. After the startup time has passed, a single comparison is done and SR is updated. Appropriate peripheral events and interrupts are also generated. No new comparisons will be performed. CTRL.USTART is cleared automatically by hardware when the single comparison has been done. 27.6.1.3 Event Triggered Single Measurement Mode This mode is enabled by writing CONFn.MODE to 3 and Peripheral Event Trigger Enable (CTRL.EVENTEN) to one. The ET mode is similar to the UT mode, the difference is that a peripheral event from another hardware module causes the hardware to automatically set the Peripheral Event Start Single Comparison bit (CTRL.ESTART). After the startup time has passed, a single comparison is done and SR is updated. Appropriate peripheral events and interrupts are also generated. No new comparisons will be performed. CTRL.ESTART is cleared automatically by hardware when the single comparison has been done. 27.6.1.4 Selecting Comparator Inputs Each Analog Comparator has one positive (INP) and one negative (INN) input. The positive input is fed from an external input pin (ACPn). The negative input can either be fed from an external input pin (ACNn) or from a reference voltage common to all ACs (ACREFN). The user selects the input source as follows: • In normal mode with the Negative Input Select and Positive Input Select fields (CONFn.INSELN and CONFn.INSELP). • In window mode with CONFn.INSELN, CONFn.INSELP and CONFn+1.INSELN, CONFn+1,INSELP. The user must configure CONFn.INSELN and CONFn+1.INSELP to the same source. 27.6.2 Interrupt Generation The interrupt request will be generated if the corresponding bit in the Interrupt Mask Register (IMR) is set. Bits in IMR are set by writing a one to the corresponding bit in the Interrupt Enable Register (IER), and cleared by writing a one to the corresponding bit in the Interrupt Disable Register (IDR). The interrupt request remains active until the corresponding bit in ISR is cleared by writing a one to the corresponding bit in the Interrupt Status Clear Register (ICR). 27.6.3 Peripheral Event Generation The ACIFB can be set up so that certain comparison results notify other parts of the device via the Peripheral Event system. Refer to Section 27.6.4.3 and Section 27.6.5.3 for information on which comparison results can generate events, and how to configure the ACIFB to achieve this. Zero or one event will be generated per comparison. 27.6.4 Normal Mode In normal mode all Analog Comparators are operating independently. 27.6.4.1 Normal Mode Output Each Analog Comparator generates one output ACOUT according to the input voltages on INP (AC positive input) and INN (AC negative input): 626 32099I–01/2012 AT32UC3L016/32/64 • ACOUT = 1 if VINP > VINN • ACOUT = 0 if VINP < VINN • ACOUT = 0 if the AC output is not available (SR.ACRDY = 0) The output can optionally be filtered, as described in Section 27.6.6. 27.6.4.2 Normal Mode Interrupt The AC channels can generate interrupts. The Interrupt Settings field in the Configuration Register (CONFn.IS) can be configured to select when the AC will generate an interrupt: • When VINP > VINN • When VINP < VINN • On toggle of the AC output (ACOUT) • When comparison has been done 27.6.4.3 Normal Mode Peripheral Events The ACIFB can generate peripheral events according to the configuration of CONFn.EVENN and CONFn.EVENP. • When VINP > VINN or • When VINP < VINN or • On toggle of the AC output (ACOUT) 27.6.5 Window Mode In window mode, two ACs (an even and the following odd build up a pair) are grouped. The negative input of ACn (even) and the positive input of ACn+1 (odd) has to be connected together externally to the device and are controlled by the Input Select fields in the AC Configuration Registers (CONFn.INSELN and CONFn+1.INSELP). The positive input of ACn (even) and the negative input of ACn+1 (odd) can still be configured independently by CONFn.INSELP and CONFn+1.INSELN, respectively. 627 32099I–01/2012 AT32UC3L016/32/64 Figure 27-2. Analog Comparator Interface in Window Mode ACAP0 + ACAN0 AC0 Filter - Interrupt Generator ACOUT0 ACWOUT COMMON Comparator pair 0 Window Module window event SR.WFCS0 - Filter SR.ACCS0 27.6.5.1 AC1 Peripheral Event Generator Window ACOUT1 + ACBP0 ACBN0 IRQ Window Mode Output When operating in window mode, each channel generates the same ACOUT outputs as in normal mode, see Section 27.6.4.1. Additionally, the ACIFB generates a window mode signal (acwout) according to the common input voltage to be compared: • ACWOUT = 1 if the common input voltage is inside the window, VACN(N+1) < Vcommon < VACP(N) • ACWOUT = 0 if the common input voltage is outside the window, Vcommon < VACN(N+1) or Vcommon > VACP(N) • ACWOUT = 0 if the window mode output is not available (SR.ACRDYn=0 or SR.ACRDYn+1=0) 27.6.5.2 Window Mode Interrupts When operating in window mode, each channel can generate the same interrupts as in normal mode, see Section 27.6.4.2. Additionally, when channels operate in window mode, programming Window Mode Interrupt Settings in the Window Mode Configuration Register (CONFWn.WIS) can cause interrupts to be generated when: • As soon as the common input voltage is inside the window. • As soon as the common input voltage is outside the window. • On toggle of the window compare output (ACWOUT). • When the comparison in both channels in the window pair is ready. 628 32099I–01/2012 AT32UC3L016/32/64 27.6.5.3 Window Mode Peripheral Events When operating in window mode, each channel can generate the same peripheral events as in normal mode, see Section 27.6.4.3. Additionally, when channels operate in window mode, programming Window Mode Event Selection Source (CONFWn.WEVSRC) can cause peripheral events to be generated when: • As soon as the common input voltage is inside the window. • As soon as the common input voltage is outside the window. • On toggle of the window compare output (ACWOUT) • Whenever a comparison is ready and the common input voltage is inside the window. • Whenever a comparison is ready and the common input voltage is outside the window. • When the comparison in both channels in the window pair is ready. 27.6.6 Filtering The output of the comparator can be filtered to reduce noise. The filter length is determined by the Filter Length field in the CONFn register (CONFn.FLEN). The filter samples the Analog Comparator output at the GCLK frequency for 2CONFn.FLEN samples. A separate counter (CNT) counts the number of cycles the AC output was one. This filter is deactivated if CONFn.FLEN equals 0. If the filter is enabled, the Hysteresis Value field HYS in the CONFn register (CONFn.HYS) can be used to define a hysteresis value. The hysteresis value should be chosen so that: 2 FLEN ---------------- > HYS 2 The filter function is defined by: 2 FLEN CNT ≥ ⎛ ---------------- + HYS⎞ ⇒ comp = 1 ⎝ 2 ⎠ FLEN FLEN 2 2 ⎛ --------------- + HYS⎞ > CNT ≥ ⎛ ---------------- – H YS⎞ ⇒ comp unchanged ⎝ 2 ⎠ ⎝ 2 ⎠ 2 FLEN CNT < ⎛ ---------------- – H YS⎞ ⇒ comp = 0 ⎝ 2 ⎠ The filtering algorithm is explained in Figure 27-3. 2FLEN measurements are sampled. If the number of measurements that are zero is less than (2FLEN/2 - HYS), the filtered result is zero. If the number of measurements that are one is more than (2FLEN/2 + HYS), the filtered result is one. Otherwise, the result is unchanged. 629 32099I–01/2012 AT32UC3L016/32/64 Figure 27-3. The Filtering Algorithm 2 FLEN 2 0 ”Result=0" Result = UNCHANGED HYS 27.7 ”Result=1" 2 FLEN HYS Peripheral Event Triggers Peripheral events from other modules can trigger comparisons in the ACIFB. All channels that are set up in Event Triggered Single Measurement Mode will be started simultaneously when a peripheral event is received. Channels that are operating in Continuous Measurement Mode or User Triggered Single Measurement Mode will be unaffected by the received event. The software can still operate these channels independently of channels in Event Triggered Single Measurement Mode. A peripheral event will trigger one or more comparisons, in normal or window mode. 27.8 AC Test mode By writing the Analog Comparator Test Mode (CR.ACTEST) bit to one, the outputs from the ACs are overridden by the value in the Test Register (TR), see Figure 27-1. This is useful for software development. 630 32099I–01/2012 AT32UC3L016/32/64 27.9 User Interface Table 27-4. ACIFB Register Memory Map Offset Register Register Name Access Reset 0x00 Control Register CTRL Read/Write 0x00000000 0x04 Status Register SR Read-only 0x00000000 0x10 Interrupt Enable Register IER Write-only 0x00000000 0x14 Interrupt Disable Register IDR Write-only 0x00000000 0x18 Interrupt Mask Register IMR Read-only 0x00000000 0x1C Interrupt Status Register ISR Read-only 0x00000000 0x20 Interrupt Status Clear Register ICR Write-only 0x00000000 0x24 Test Register TR Read/Write 0x00000000 0x30 Parameter Register PARAMETER Read-only -(1) 0x34 Version Register VERSION Read-only -(1) 0x80 Window0 Configuration Register CONFW0 Read/Write 0x00000000 0x84 Window1 Configuration Register CONFW1 Read/Write 0x00000000 0x88 Window2 Configuration Register CONFW2 Read/Write 0x00000000 0x8C Window3 Configuration Register CONFW3 Read/Write 0x00000000 0xD0 AC0 Configuration Register CONF0 Read/Write 0x00000000 0xD4 AC1 Configuration Register CONF1 Read/Write 0x00000000 0xD8 AC2 Configuration Register CONF2 Read/Write 0x00000000 0xDC AC3 Configuration Register CONF3 Read/Write 0x00000000 0xE0 AC4 Configuration Register CONF4 Read/Write 0x00000000 0xE4 AC5 Configuration Register CONF5 Read/Write 0x00000000 0xE8 AC6 Configuration Register CONF6 Read/Write 0x00000000 0xEC AC7 Configuration Register CONF7 Read/Write 0x00000000 Note: 1. The reset values for these registers are device specific. Please refer to the Module Configuration section at the end of this chapter. 631 32099I–01/2012 AT32UC3L016/32/64 27.9.1 Name: Control Register CTRL Access Type: Read/Write Offset: 0x00 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - 15 14 13 12 11 10 9 8 SUT[9:8] SUT[7:0] 7 6 5 4 3 2 1 0 ACTEST - ESTART USTART - - -EVENTEN EN • SUT: Startup Time SUT Analog Comparator startup time = ----------------- . F GCLK Each time an AC is enabled, the AC comparison will be enabled after the startup time of the AC. • ACTEST: Analog Comparator Test Mode 0: The Analog Comparator outputs feeds the channel logic in ACIFB. 1: The Analog Comparator outputs are bypassed with the AC Test Register. • ESTART: Peripheral Event Start Single Comparison Writing a zero to this bit has no effect. Writing a one to this bit starts a comparison and can be used for test purposes. This bit is cleared when comparison is done. This bit is set when an enabled peripheral event is received. • USTART: User Start Single Comparison Writing a zero to this bit has no effect. Writing a one to this bit starts a Single Measurement Mode comparison. This bit is cleared when comparison is done. • EVENTEN: Peripheral Event Trigger Enable 0: A peripheral event will not trigger a comparison. 1: Enable comparison triggered by a peripheral event. • EN: ACIFB Enable 0: The ACIFB is disabled. 1: The ACIFB is enabled. 632 32099I–01/2012 AT32UC3L016/32/64 27.9.2 Name: Status Register SR Access Type: Read-only Offset: 0x04 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - WFCS3 WFCS2 WFCS1 WFCS0 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 ACRDY7 ACCS7 ACRDY6 ACCS6 ACRDY5 ACCS5 ACRDY4 ACCS4 7 6 5 4 3 2 1 0 ACRDY3 ACCS3 ACRDY2 ACCS2 ACRDY1 ACCS1 ACRDY0 ACCS0 • WFCSn: Window Mode Current Status This bit is cleared when the common input voltage is outside the window. This bit is set when the common input voltage is inside the window. • ACRDYn: ACn Ready This bit is cleared when the AC output (ACOUT) is not ready. This bit is set when the AC output (ACOUT) is ready, AC is enabled and its startup time is over. • ACCSn: ACn Current Comparison Status This bit is cleared when VINP is currently lower than VINN This bit is set when VINP is currently greater than VINN. 633 32099I–01/2012 AT32UC3L016/32/64 27.9.3 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x10 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - WFINT3 WFINT2 WFINT1 WFINT0 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 SUTINT7 ACINT7 SUTINT6 ACINT6 SUTINT5 ACINT5 SUTINT4 ACINT4 7 6 5 4 3 2 1 0 SUTINT3 ACINT3 SUTINT2 ACINT2 SUTINT1 ACINT1 SUTINT0 ACINT0 Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 634 32099I–01/2012 AT32UC3L016/32/64 27.9.4 Name: Interrupt Disable Register IDR Access Type: Write-only Offset: 0x14 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - WFINT3 WFINT2 WFINT1 WFINT0 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 SUTINT7 ACINT7 SUTINT6 ACINT6 SUTINT5 ACINT5 SUTINT4 ACINT4 7 6 5 4 3 2 1 0 SUTINT3 ACINT3 SUTINT2 ACINT2 SUTINT1 ACINT1 SUTINT0 ACINT0 Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 635 32099I–01/2012 AT32UC3L016/32/64 27.9.5 Name: Interrupt Mask Register IMR Access Type: Read-only Offset: 0x18 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - WFINT3 WFINT2 WFINT1 WFINT0 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 SUTINT7 ACINT7 SUTINT6 ACINT6 SUTINT5 ACINT5 SUTINT4 ACINT4 7 6 5 4 3 2 1 0 SUTINT3 ACINT3 SUTINT2 ACINT2 SUTINT1 ACINT1 SUTINT0 ACINT0 • WFINTn: Window Mode Interrupt Mask 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. This bit is cleared when the corresponding bit in IDR is written to one. This bit is set when the corresponding bit in IER is written to one. • SUTINTn: ACn Startup Time Interrupt Mask 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. This bit is cleared when the corresponding bit in IDR is written to one. This bit is set when the corresponding bit in IER is written to one. • ACINTn: ACn Interrupt Mask 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. This bit is cleared when the corresponding bit in IDR is written to one. This bit is set when the corresponding bit in IER is written to one. 636 32099I–01/2012 AT32UC3L016/32/64 27.9.6 Name: Interrupt Status Register ISR Access Type: Read-only Offset: 0x1C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - WFINT3 WFINT2 WFINT1 WFINT0 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 SUTINT7 ACINT7 SUTINT6 ACINT6 SUTINT5 ACINT5 SUTINT4 ACINT4 7 6 5 4 3 2 1 0 SUTINT3 ACINT3 SUTINT2 ACINT2 SUTINT1 ACINT1 SUTINT0 ACINT0 • WFINTn: Window Mode Interrupt Status 0: No Window Mode Interrupt is pending. 1: Window Mode Interrupt is pending. This bit is cleared when the corresponding bit in ICR is written to one. This bit is set when the corresponding channel pair operating in window mode generated an interrupt. • SUTINTn: ACn Startup Time Interrupt Status 0: No Startup Time Interrupt is pending. 1: Startup Time Interrupt is pending. This bit is cleared when the corresponding bit in ICR is written to one. This bit is set when the startup time of the corresponding AC has passed. • ACINTn: ACn Interrupt Status 0: No Normal Mode Interrupt is pending. 1: Normal Mode Interrupt is pending. This bit is cleared when the corresponding bit in ICR is written to one. This bit is set when the corresponding channel generated an interrupt. 637 32099I–01/2012 AT32UC3L016/32/64 27.9.7 Name: Interrupt Status Clear Register ICR Access Type: Write-only Offset: 0x20 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - WFINT3 WFINT2 WFINT1 WFINT0 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 SUTINT7 ACINT7 SUTINT6 ACINT6 SUTINT5 ACINT5 SUTINT4 ACINT4 7 6 5 4 3 2 1 0 SUTINT3 ACINT3 SUTINT2 ACINT2 SUTINT1 ACINT1 SUTINT0 ACINT0 Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in ISR and the corresponding interrupt request. 638 32099I–01/2012 AT32UC3L016/32/64 27.9.8 Name: Test Register TR Access Type: Read/Write Offset: 0x24 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 ACTEST7 ACTEST6 ACTEST5 ACTEST4 ACTEST3 ACTEST2 ACTEST1 ACTEST0 • ACTESTn: AC Output Override Value If CTRL.ACTEST is set, the ACn output is overridden with the value of ACTESTn. 639 32099I–01/2012 AT32UC3L016/32/64 27.9.9 Name: Parameter Register PARAMETER Access Type: Read-only Offset: 0x30 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - WIMPL3 WIMPL2 WIMPL1 WIMPL0 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 ACIMPL7 ACIMPL6 ACIMPL5 ACIMPL4 ACIMPL3 ACIMPL2 ACIMPL1 ACIMPL0 • WIMPLn: Window Pair n Implemented 0: Window Pair not implemented. 1: Window Pair implemented. • ACIMPLn: Analog Comparator n Implemented 0: Analog Comparator not implemented. 1: Analog Comparator implemented. 640 32099I–01/2012 AT32UC3L016/32/64 27.9.10 Name: Version Register VERSION Access Type: Read-only Offset: 0x34 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant Number Reserved. No functionality associated. • VERSION: Version Number Version number of the module. No functionality associated. 641 32099I–01/2012 AT32UC3L016/32/64 27.9.11 Name: Window Configuration Register CONFWn Access Type: Read/Write Offset: 0x80,0x84,0x88,0x8C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - WFEN 15 14 13 12 11 10 9 8 - - - - WEVEN 7 6 5 4 3 2 - - - - - - WEVSRC 0 1 WIS • WFEN: Window Mode Enable 0: The window mode is disabled. 1: The window mode is enabled. • WEVEN: Window Event Enable 0: Event from awout is disabled. 1: Event from awout is enabled. • WEVSRC: Event Source Selection for Window Mode 000: Event on acwout rising edge. 001: Event on acwout falling edge. 010: Event on awout rising or falling edge. 011: Inside window. 100: Outside window. 101: Measure done. 110-111: Reserved. • WIS: Window Mode Interrupt Settings 00: Window interrupt as soon as the input voltage is inside the window. 01: Window interrupt as soon as the input voltage is outside the window. 10: Window interrupt on toggle of window compare output. 11: Window interrupt when evaluation of input voltage is done. 642 32099I–01/2012 AT32UC3L016/32/64 27.9.12 Name: AC Configuration Register CONFn Access Type: Read/Write Offset: 0xD0,0xD4,0xD8,0xDC,0xE0,0xE4,0xE8,0xEC Reset Value: 0x00000000 31 30 - 29 28 27 26 FLEN 25 24 HYS 23 22 21 20 19 18 17 16 - - - - - - EVENP EVENN 15 14 13 12 11 10 9 8 - - - - 7 6 5 4 - - MODE INSELP INSELN 3 2 - - 0 1 IS • FLEN: Filter Length 000: Filter off. n: Number of samples to be averaged =2n. • HYS: Hysteresis Value 0000: No hysteresis. 1111: Max hysteresis. • EVENN: Event Enable Negative 0: Do not output event when ACOUT is zero. 1: Output event when ACOUT is zero. • EVENP: Event Enable Positive 0: Do not output event when ACOUT is one. 1: Output event when ACOUT is one. • INSELP: Positive Input Select 00: ACPn pin selected. 01: Reserved. 10: Reserved. 11: Reserved. • INSELN: Negative Input Select 00: ACNn pin selected. 01: ACREFN pin selected. 10: Reserved. 11: Reserved. • MODE: Mode 00: Off. 01: Continuous Measurement Mode. 10: User Triggered Single Measurement Mode. 11: Event Triggered Single Measurement Mode. 643 32099I–01/2012 AT32UC3L016/32/64 • IS: Interrupt Settings 00: Comparator interrupt when as VINP > VINN. 01: Comparator interrupt when as VINP < VINN. 10: Comparator interrupt on toggle of Analog Comparator output. 11: Comparator interrupt when comparison of VINP and VINN is done. 644 32099I–01/2012 AT32UC3L016/32/64 27.10 Module Configuration The specific configuration for each ACIFB instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Refer to the Power Manager chapter for details. Table 27-5. ACIFB Configuration Feature ACIFB Number of channels 8 Table 27-6. ACIFB Clocks Clock Name Description CLK_ACIFB Clock for the ACIFB bus interface GCLK The generic clock used for the ACIFB is GCLK4 Table 27-7. Register Reset Values Register Reset Value VERSION 0x00000202 PARAMETER 0x000F00FF 645 32099I–01/2012 AT32UC3L016/32/64 28. Capacitive Touch Module (CAT) Rev: 2.0.0.1 28.1 Features • • • • • • 28.2 QTouch® method allows N touch sensors to be implemented using 2N physical pins QMatrix method allows X by Y matrix of sensors to be implemented using (X+2Y) physical pins One autonomous QTouch sensor operates without CPU intervention Interfaced with peripheral DMA controller to reduce processor overhead External synchronization to reduce 50 or 60 Hz mains interference Spread spectrum sensor drive capability Overview The Capacitive Touch Module (CAT) senses touch on external capacitive touch sensors. Capacitive touch sensors use no external mechanical components, and therefore demand less maintenance in the user application. The module implements the QTouch method of capturing signals from capacitive touch sensors. The QTouch method is generally suitable for small numbers of sensors since it requires 2 physical pins per sensor. The module also implements the QMatrix method, which is more appropriate for large numbers of sensors since it allows an X by Y matrix of sensors to be implemented using only (X+2Y) physical pins. The module allows methods to function together, so N touch sensors and an X by Y matrix of sensors can be implemented using (2N+X+2Y) physical pins. In addition, the module allows sensors using the QTouch method to be divided into two groups. Each QTouch group can be configured with different properties. This eases the implementation of multiple kinds of controls such as push buttons, wheels, and sliders. The module also implements one autonomous QTouch sensor that is capable of detecting touch or proximity without CPU intervention. This allows proximity or activation detection in low-power sleep modes. 646 32099I–01/2012 AT32UC3L016/32/64 28.3 Block Diagram Figure 28-1. CAT Block Diagram Capacitive Touch Module (CAT) CLK_CAT Peripheral Bus Interface Registers Counters Finite State Machine Capacitor Charge and Discharge Sequence Generator GCLK_CAT CSAn CSBn I/O Controller Pins SMP VDIVEN NOTE: Italicized signals and blocks are used only for QMatrix operation DIS Discharge Current Sources Yn SYNC Peripheral Event System Analog Comparator Interface Analog Comparators 647 32099I–01/2012 AT32UC3L016/32/64 28.4 I/O Lines Description Table 28-1. 28.5 I/O Lines Description Name Description Type CSAn Capacitive sense A line n I/O CSBn Capacitive sense B line n I/O DIS Discharge current control (only used for QMatrix) Analog SMP SMP line (only used for QMatrix) Output SYNC Synchronize signal VDIVEN Voltage divider enable (only used for QMatrix) Input Output Product Dependencies In order to use the CAT module, other parts of the system must be configured correctly, as described below. 28.5.1 I/O Lines The CAT pins may be multiplexed with other peripherals. The user must first program the I/O Controller to give control of the pins to the CAT module. In QMatrix mode, the Y lines must be driven by the CAT and analog comparators sense the voltage on the Y lines. Thus, the CAT (not the Analog Comparator Interface) must be the selected function for the Y lines in the I/O Controller. By writing ones and zeros to bits in the Pin Mode Registers (PINMODEx), most of the CAT pins can be individually selected to implement the QTouch method or the QMatrix method. Each pin has a different name and function depending on whether it is implementing the QTouch method or the QMatrix method. The following table shows the pin names for each method and the bits in the PINMODEx registers which control the selection of the QTouch or QMatrix method. Table 28-2. Pin Selection Guide CAT Module Pin Name QTouch Method Pin Name QMatrix Method Pin Name Selection Bit in PINMODEx Register CSA0 SNS0 X0 SP0 CSB0 SNSK0 X1 SP0 CSA1 SNS1 Y0 SP1 CSB1 SNSK1 YK0 SP1 CSA2 SNS2 X2 SP2 CSB2 SNSK2 X3 SP2 CSA3 SNS3 Y1 SP3 CSB3 SNSK3 YK1 SP3 CSA4 SNS4 X4 SP4 CSB4 SNSK4 X5 SP4 CSA5 SNS5 Y2 SP5 648 32099I–01/2012 AT32UC3L016/32/64 Table 28-2. 28.5.2 Pin Selection Guide CAT Module Pin Name QTouch Method Pin Name QMatrix Method Pin Name Selection Bit in PINMODEx Register CSB5 SNSK5 YK2 SP5 CSA6 SNS6 X6 SP6 CSB6 SNSK6 X7 SP6 CSA7 SNS7 Y3 SP7 CSB7 SNSK7 YK3 SP7 CSA8 SNS8 X8 SP8 CSB8 SNSK8 X9 SP8 CSA9 SNS9 Y4 SP9 CSB9 SNSK9 YK4 SP9 CSA10 SNS10 X10 SP10 CSB10 SNSK10 X11 SP10 CSA11 SNS11 Y5 SP11 CSB11 SNSK11 YK5 SP11 CSA12 SNS12 X12 SP12 CSB12 SNSK12 X13 SP12 CSA13 SNS13 Y6 SP13 CSB13 SNSK13 YK6 SP13 CSA14 SNS14 X14 SP14 CSB14 SNSK14 X15 SP14 CSA15 SNS15 Y7 SP15 CSB15 SNSK15 YK7 SP15 CSA16 SNS16 X16 SP16 CSB16 SNSK16 X17 SP16 Clocks The clock for the CAT module, CLK_CAT, is generated by the Power Manager (PM). This clock is turned on by default, and can be enabled and disabled in the PM. The user must ensure that CLK_CAT is enabled before using the CAT module. QMatrix operations also require the CAT generic clock, GCLK_CAT. This generic clock is generated by the System Control Interface (SCIF), and is shared between the CAT and the Analog Comparator Interface. The user must ensure that the GCLK_CAT is enabled in the SCIF before using QMatrix functionality in the CAT module. For proper QMatrix operation, the frequency of GCLK_CAT must be less than half the frequency of CLK_CAT. If only QTouch functionality is used, then GCLK_CAT is unnecessary. 28.5.3 Interrupts The CAT interrupt request line is connected to the interrupt controller. Using CAT interrupts requires the interrupt controller to be programmed first. 649 32099I–01/2012 AT32UC3L016/32/64 28.5.4 Peripheral Events The CAT module receives nine peripheral events, one from the AST and the remaining eight from on-chip analog comparators. These peripheral events are connected via the Peripheral Event System. Refer to the Peripheral Event System chapter for additional details. 28.5.5 Peripheral Direct Memory Access The CAT module provides handshake capability for a Peripheral DMA Controller. One handshake controls transfers from the Acquired Count Register (ACOUNT) to memory. A second handshake requests burst lengths for each (X,Y) pair to the Matrix Burst Length Register (MBLEN) when using the QMatrix acquisition method. The Peripheral DMA Controller must be configured properly and enabled in order to perform direct memory access transfers to/from the CAT module. 28.5.6 Analog Comparators When the CAT module is performing QMatrix acquisition, it requires that on-chip analog comparators be used as part of the process. These analog comparators are not controlled directly by the CAT module, but by a separate Analog Comparator (AC) Interface. This interface must be configured properly and enabled before the CAT module is used. This includes configuring the generic clock input for the analog comparators to the proper sampling frequency. 28.5.7 28.6 28.6.1 Debug Operation When an external debugger forces the CPU into debug mode, the CAT continues normal operation. If the CAT is configured in a way that requires it to be periodically serviced by the CPU through interrupts or similar, improper operation or data loss may result during debugging. Functional Description Acquisition Types The CAT module can perform three types of QTouch acquisition from capacitive touch sensors: autonomous QTouch (one sensor only), QTouch group A, and QTouch group B. The CAT module can also perform QMatrix acquisition. Each type of acquisition has an associated set of pin selection and configuration registers that allow a large degree of flexibility. The following schematic diagrams show typical hardware connections for QTouch and QMatrix sensors, respectively: Figure 28-2. CAT Touch Connections SNSKn QTouch Sensor 32-bit AVR Microcontroller Cs (Sense Capacitor) SNSn 650 32099I–01/2012 AT32UC3L016/32/64 Figure 28-3. CAT Matrix Connections X2 X3 X6 QMatrix Sensor Array X7 YK0 Y0 Cs0 (Sense Capacitor) YK1 Y1 Cs1 (Sense Capacitor) 32-bit AVR Microcontroller Rsmp1 Rsmp0 SMP DIS Rdis VDIVEN Ra ACREFN Rb NOTE: If the CAT internal current sources will be enabled, the SMP signal and Rsmp resistors should NOT be included in the design. If the CAT internal current sources will NOT be enabled, the DIS signal and Rdis resistor should NOT be included in the design. In order to use the autonomous QTouch detection capability, the user must first set up the Autonomous Touch Pin Select Register (ATPINS) and Autonomous Touch Configuration Registers (ATCFG0 through 3) with appropriate values. The module can then be enabled using the Control Register (CTRL). After the module is enabled, the module will acquire data from the autonomous QTouch sensor and use it to determine whether the sensor is activated. The active/inactive status of the autonomous QTouch sensor is reported in the Status Register (SR), and it is also possible to configure the CAT to generate an interrupt whenever the status changes. The module will continue acquiring autonomous QTouch sensor data and updating autonomous QTouch status until the module is disabled or reset. In order to use the QMatrix, QTouch group A, or QTouch group B acquisition capabilities, it is first necessary to set up the appropriate pin mode registers (PINMODE0 and PINMODE1) and configuration registers (MGCFG0, MGCFG1, TGACFG0, TGACFG1, TGBCFG0, and TGBCFG1). The module must then be enabled using the CTRL register. In order to initiate acquisition, it is necessary to perform a write to the Acquisition Initiation and Selection Register (AISR). The specific value written to AISR determines which type of acquisition will be performed: QMatrix, QTouch group A, or QTouch group B. The CPU can initiate acquisition by writing to the AISR. While QMatrix, QTouch group A, or QTouch group B acquisition is in progress, the module collects count values from the sensors and buffers them. Availability of acquired count data is indicated by the Acquisition Ready (ACREADY) bit in the Status Register (SR). The CPU or the Peripheral DMA Controller can then read the acquired counts from the ACOUNT register. 651 32099I–01/2012 AT32UC3L016/32/64 Because the CAT module is configured with Peripheral DMA Controller capability that can transfer data from memory to MBLEN and from ACOUNT to memory, the Peripheral DMA Controller can perform long acquisition sequences and store results in memory without CPU intervention. 28.6.2 Prescaler and Charge Length Each QTouch acquisition type (autonomous QTouch, QTouch group A, and QTouch group B) has its own prescaler. Each QTouch prescaler divides down the CLK_CAT clock to an appropriate sampling frequency for its particular acquisition type. Typical frequencies are 1 MHz for QTouch acquisition and 4MHz for QMatrix burst timing control. Each QTouch prescaler is controlled by the DIV field in the appropriate Configuration Register 0 (ATCFG0, TGACFG0, or TGBCFG0). The QMatrix burst timing prescaler is controlled by the DIV field in MGCFG0. Each prescaler uses the following formula to generate the sampling clock: Sampling clock = CLK_CAT / (2(DIV+1)) The capacitive sensor charge length, discharge length, and settle length can be determined for each acquisition type using the CHLEN, DILEN, and SELEN fields in Configuration Registers 0 and 1. The lengths are specified in terms of prescaler clocks. In addition, the QMatrix Cx discharge length can be determined using the CXDILEN field in MGCFG2. For QMatrix acquisition, the duration of CHLEN should not be set to the same value as the period of any periodic signal on any other pin. If the duration of CHLEN is the same as the period of a signal on another pin, it is likely that the other signal will significantly affect measurements due to stray capacitive coupling. For example, if a 1 MHz signal is generated on another pin of the device, then CHLEN should not be 1 microsecond. For the QMatrix method, burst and capture lengths are set for each (X,Y) pair by writing the desired length values to the MBLEN register. The write must be done before each X line can start its acquisition and is indicated by the status bit MBLREQ in the Status Register (SR). A DMA handshake interface is also connected to this status bit to reduce CPU overhead during QMatrix acquisitions. Four burst lengths (BURST0..3) can be written at one time into the MBLEN register. If the current configuration uses Y lines larger than Y3 the register has to be written a second time. The first write to MBLEN specifies the burst length for Y lines 0 to 3 in the BURST0 to BURST3 fields, respectively. The second write specifies the burst length for Y lines 4 to 7 in fields BURST0 to BURST3, respectively, and so on. The Y and YK pins remain clamped to ground apart from the specified number of burst pulses, when charge is transferred and captured into the sampling capacitor. 28.6.3 Capacitive Count Acquisition For the QMatrix, QTouch group A, and QTouch group B types of acquisition, the module acquires count values from the sensors, buffers them, and makes them available for reading in the ACOUNT register. Further processing of the count values must be performed by the CPU. When the module performs QMatrix acquisition using multiple Y lines, it starts the capture for each Y line at the appropriate time in the burst sequence so that all captures finish simultaneously. For example, suppose that an acquisition is performed on Y0 and Y1 with BURST0=53 and BURST1=60. The module will first toggle the X line 7 times while capturing on Y1 while Y0 and YK0 are clamped to ground. The module will then toggle the X line 53 times while capturing on both Y1 and Y0. 652 32099I–01/2012 AT32UC3L016/32/64 28.6.4 Autonomous QTouch For autonomous QTouch, a complete detection algorithm is implemented within the CAT module. The additional parameters needed to control the autonomous QTouch detection algorithm must be specified by the user in the ATCFG2 and ATCFG3 registers. Autonomous QTouch sensitivity and out-of-touch sensitivity can be adjusted with the SENSE and OUTSENS fields, respectively, in ATCFG2. Each field accepts values from one to 255 where 255 is the least sensitive setting. The value in the OUTSENS field should be smaller than the value in the SENSE field. To avoid false positives a detect integration filtering technique can be used. The number of successive detects required is specified in the FILTER field of the ATCFG2 register. To compensate for changes in capacitance the CAT can recalibrate the autonomous QTouch sensor periodically. The timing of this calibration is done with the NDRIFT and PDRIFT fields in the Configuration register, ATCFG3. It is recommended that the PDRIFT value is smaller than the NDRIFT value. The autonomous QTouch sensor will also recalibrate if the count value goes too far positive beyond a threshold. This positive recalibration threshold is specified by the PTHR field in the ATCFG3 register. The following block diagram shows the sequence of acquisition and processing operations used by the CAT module. The AISR written bit is internal and not visible in the user interface. 28.6.5 Peripheral Events The peripheral event from the AST is used to trigger one iteration of autonomous touch detection when the device is in a sleep mode that has disabled CLK_CAT. When CLK_CAT is disabled and the peripheral event from the AST becomes active, a request will automatically be made to enable CLK_CAT. When CLK_CAT is enabled, the CAT will perform one iteration of the autonomous touch detection algorithm. After this, the CLK_CAT will be disabled, and the CAT module will remain in a frozen state until the next AST peripheral event. The eight peripheral events from the analog comparators are automatically used by the CAT when performing QMatrix acquisition. The CAT will automatically use the negative peripheral events from the AC Interface on every Y pin in QMatrix mode. When QMatrix acquisition is used the analog comparator corresponding to the selected Y pins must be enabled and converting continuously, using the Y pin as the positive reference and the ACREFN as negative reference. 653 32099I–01/2012 AT32UC3L016/32/64 Figure 28-4. CAT Acquisition and Processing Sequence Idle AISR written flag set? No Yes Autonomous touch enabled (ATEN)? No 28.6.6 Acquire counts Yes Acquire autonomous touch count Wait for all acquired counts to be transferred Update autonomous touch detection algorithm Clear AISR written flag Spread Spectrum Sensor Drive To reduce electromagnetic compatibility issues, the capacitive sensors can be driven with a spread spectrum signal. To enable spread spectrum drive for a specific acquisition type, the user must write a one to the SPREAD bit in the appropriate Configuration Register 1 (MGCFG1, ATCFG1, TGACFG1, or TGBCFG1). During spread spectrum operation, the length of each pulse within a burst is varied in a deterministic pattern, so that the exact same burst pattern is used for a specific burst length. The maximum spread is determined by the MAXDEV field in the Spread Spectrum Configuration Register (SSCFG) register. The prescaler divisor is varied in a sawtooth pattern from (2(DIV+1))-MAXDEV to (2(DIV+1))+MAXDEV and then back to (2(DIV+1))-MAXDEV. For example, if DIV is 2 and MAXDEV is 3, the prescaler divisor will have the following sequence: 6, 7, 8, 654 32099I–01/2012 AT32UC3L016/32/64 9, 3, 4, 5, 6, 7, 8, 9, 3, 4, etc. MAXDEV must not exceed the value of (2(DIV+1)), or undefined behavior will occur. 28.6.7 Synchronization To prevent interference from the 50 or 60Hz mains line the CAT can trigger acquisition on the SYNC signal. The SYNC signal should be derived from the mains line. The acquisition will trigger on a falling edge of this signal. To enable synchronization for a specific acquisition type, the user must write a one to the SYNC bit in the appropriate Configuration Register 1 (MGCFG1, ATCFG1, TGACFG1, or TGBCFG1). For QMatrix acquisition, all X lines must be sampled at a specific phase of the noise signal for the synchronization to be effective. This can be accomplished by the synchronization timer, which is enabled by writing a non-zero value to the SYNCTIM field in the MGCFG2 register. This ensures that the start of the acquisition of each X line is spaced at regular intervals, defined by the SYNCTIM field. 28.6.8 Resistive Drive By default, the CAT pins are driven with normal I/O drive properties. Some of the CSA and CSB pins can optionally drive with a 1k output resistance for improved EMC. The pins that have this capability are listed in the Module Configuration section. 28.6.9 Discharge Current Sources The device integrates discharge current sources, which can be used to discharge the sampling capacitors during the QMatrix measurement phase. The discharge current sources are enabled by writing the GLEN bit in the Discharge Current Source (DICS) register to one. This enables an internal reference voltage, which can be either the internal 1.1V band gap voltage or VDDIO/3, as selected by the INTVREFSEL bit in the DICS register. If the DICS.INTREFSEL bit is one, the reference voltage is applied across an internal reference resistor, Rint. Otherwise, the voltage is applied to the DIS pin, and an external reference resistor must be connected between DIS and ground. The nominal discharge current is given by the following formula, where Vref is the reference voltage, R ref is the value of the reference resistor, trim is the value written to the DICS.TRIM field, and k is a constant of proportionality: I = (Vref/Rref)*(1+(k*trim)) The values for the internal reference resistor, Rint, and the constant, k, may be found in the Electrical Characteristics section. The nominal discharge current may be programmed between 2 and 20µA. The reference current can be fine-tuned by adjusting the trim value in the DICS.TRIM field. The reference current is mirrored to each Y-pin if the corresponding bit is written to one in the DICS.SOURCES field. 28.6.10 Voltage Divider Enable (VDIVEN) Capability In many QMatrix applications, the sense capacitors will be charged to 50 mV or more and the negative reference pin (ACREFN) of the analog comparators can be tied directly to ground. In that case, the relatively small input offset voltage of the comparators will not cause acquisition problems. However, in certain specialized QMatrix applications such as interpolated touch screens, it may be desirable for the sense capacitors to be charged to less than 25 mV. When such small voltages are used on the sense capacitors, the input offset voltage of the comparators becomes an issue and can cause QMatrix acquisition problems. 655 32099I–01/2012 AT32UC3L016/32/64 Problems with QMatrix acquisition of small sense capapcitor voltages can be solved by connecting the negative reference pin (ACREFN) to a voltage divider that produces a small positive voltage (20 mV, typically) to cancel any negative input offset voltage. With a 3.3V supply, recommended values for the voltage divider are Ra (resistor from positive supply to ACREFN) of 8200 ohm and Rb (resistor from ACREFN to ground) of 50 ohm. These recommended values will produce 20 mV on the ACREFN pin, which should generally be enough to compensate for the worst-case negative input offset of the analog comparators. Unfortunately, such a voltage divider constantly draws a small current from the power supply, reducing battery life in portable applications. In order to prevent this constant power drain, the CAT module provides a voltage divider enable pin (VDIVEN) that can be used for driving the voltage divider. The VDIVEN pin provides power to the voltage divider only when the comparators are actually performing QMatrix comparisions. When the comparators are inactive, the VDIVEN output is zero. This minimizes the power consumed by the voltage divider. 656 32099I–01/2012 AT32UC3L016/32/64 28.7 User Interface Table 28-3. CAT Register Memory Map Offset Register Register Name Access Reset 0x00 Control Register CTRL Read/Write 0x00000000 0x04 Autonomous Touch Pin Selection Register ATPINS Read/Write 0x00000000 0x08 Pin Mode Register 0 PINMODE0 Read/Write 0x00000000 0x0C Pin Mode Register 1 PINMODE1 Read/Write 0x00000000 0x10 Autonomous Touch Configuration Register 0 ATCFG0 Read/Write 0x00000000 0x14 Autonomous Touch Configuration Register 1 ATCFG1 Read/Write 0x00000000 0x18 Autonomous Touch Configuration Register 2 ATCFG2 Read/Write 0x00000000 0x1C Autonomous Touch Configuration Register 3 ATCFG3 Read/Write 0x00000000 0x20 Touch Group A Configuration Register 0 TGACFG0 Read/Write 0x00000000 0x24 Touch Group A Configuration Register 1 TGACFG1 Read/Write 0x00000000 0x28 Touch Group B Configuration Register 0 TGBCFG0 Read/Write 0x00000000 0x2C Touch Group B Configuration Register 1 TGBCFG1 Read/Write 0x00000000 0x30 Matrix Group Configuration Register 0 MGCFG0 Read/Write 0x00000000 0x34 Matrix Group Configuration Register 1 MGCFG1 Read/Write 0x00000000 0x38 Matrix Group Configuration Register 2 MGCFG2 Read/Write 0x00000000 0x3C Status Register SR Read-only 0x00000000 0x40 Status Clear Register SCR Write-only - 0x44 Interrupt Enable Register IER Write-only - 0x48 Interrupt Disable Register IDR Write-only - 0x4C Interrupt Mask Register IMR Read-only 0x00000000 0x50 Acquisition Initiation and Selection Register AISR Read/Write 0x00000000 0x54 Acquired Count Register ACOUNT Read-only 0x00000000 0x58 Matrix Burst Length Register MBLEN Write-only - 0x5C Discharge Current Source Register DICS Read/Write 0x00000000 0x60 Spread Spectrum Configuration Register SSCFG Read/Write 0x00000000 0x64 CSA Resistor Control Register CSARES Read/Write 0x00000000 0x68 CSB Resistor Control Register CSBRES Read/Write 0x00000000 0x6C Autonomous Touch Base Count Register ATBASE Read-only 0x00000000 0x70 Autonomous Touch Current Count Register ATCURR Read-only 0x00000000 0x80 Analog Comparator Shift Offset Register 0 ACSHI0 Read/Write 0x00000000 0x84 Analog Comparator Shift Offset Register 1 ACSHI1 Read/Write 0x00000000 0x88 Analog Comparator Shift Offset Register 2 ACSHI2 Read/Write 0x00000000 0x8C Analog Comparator Shift Offset Register 3 ACSHI3 Read/Write 0x00000000 0x90 Analog Comparator Shift Offset Register 4 ACSHI4 Read/Write 0x00000000 0x94 Analog Comparator Shift Offset Register 5 ACSHI5 Read/Write 0x00000000 657 32099I–01/2012 AT32UC3L016/32/64 Table 28-3. CAT Register Memory Map Offset Register Register Name Access Reset 0x98 Analog Comparator Shift Offset Register 6 ACSHI6 Read/Write 0x00000000 0x9C Analog Comparator Shift Offset Register 7 ACSHI7 Read/Write 0x00000000 0xF8 Parameter Register PARAMETER Read-only 0xFC Version Register VERSION Read-only 658 32099I–01/2012 AT32UC3L016/32/64 28.7.1 Name: Control Register CTRL Access Type: Read/Write Offset: 0x00 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 SWRST - - - - - - EN • SWRST: Software reset Writing a zero to this bit has no effect. Writing a one to this bit resets the module. The module will be disabled after the reset. This bit always reads as zero. • EN: Module enable 0: Module is disabled. 1: Module is enabled. 659 32099I–01/2012 AT32UC3L016/32/64 28.7.2 Name: Autonomous Touch Pin Selection Register ATPINS Access Type: Read/Write Offset: 0x04 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - ATEN 7 6 5 4 3 2 1 0 - - - ATSP • ATEN: Autonomous Touch Enable 0: Autonomous QTouch acquisition and detection is disabled. 1: Autonomous QTouch acquisition and detection is enabled using the sense pair specified in ATSP. • ATSP: Autonomous Touch Sense Pair Selects the sense pair that will be used by the autonomous QTouch sensor. A value of n will select sense pair n (CSAn and CSBn pins). 660 32099I–01/2012 AT32UC3L016/32/64 28.7.3 Name: Pin Mode Registers 0 and 1 PINMODE0 and PINMODE1 Access Type: Read/Write Offset: 0x08, 0x0C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 23 22 21 20 - 15 14 13 SP[16] 12 11 10 9 8 3 2 1 0 SP[15:8] 7 6 5 4 SP[7:0] • SP: Sense Pair Mode Selection Each SP[n] bit determines the operation mode of sense pair n (CSAn and CSBn pins). The (PINMODE1.SP[n] PINMODE0.SP[n]) bits have the following definitions: 00: Sense pair n disabled. 01: Sense pair n is assigned to QTouch Group A. 10: Sense pair n is assigned to QTouch Group B. 11: Sense pair n is assigned to the QMatrix Group. 661 32099I–01/2012 AT32UC3L016/32/64 28.7.4 Name: Autonomous Touch Configuration Register 0 ATCFG0 Access Type: Read/Write Offset: 0x10 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 DIV[15:8] 23 22 21 20 DIV[7:0] 15 14 13 12 CHLEN 7 6 5 4 SELEN • DIV: Clock Divider The prescaler is used to ensure that the CLK_CAT clock is divided to around 1 MHz to produce the sampling clock.The prescaler uses the following formula to generate the sampling clock: Sampling clock = CLK_CAT / (2(DIV+1)) • CHLEN: Charge Length For the autonomous QTouch sensor, specifies how many sample clock cycles should be used for transferring charge to the sense capacitor. • SELEN: Settle Length For the autonomous QTouch sensor, specifies how many sample clock cycles should be used for settling after charge transfer. 662 32099I–01/2012 AT32UC3L016/32/64 28.7.5 Name: Autonomous Touch Configuration Register 1 ATCFG1 Access Type: Read/Write Offset: 0x14 Reset Value: 0x00000000 31 30 29 23 28 27 DISHIFT 22 21 26 - 20 25 24 SYNC SPREAD 19 18 17 16 11 10 9 8 3 2 1 0 DILEN 15 14 13 12 MAX[15:8] 7 6 5 4 MAX[7:0] • DISHIFT: Discharge Shift For the autonomous QTouch sensor, specifies how many bits the DILEN field should be shifted before using it to determine the discharge time. • SYNC: Sync Pin For the autonomous QTouch sensor, specifies that acquisition shall begin when a falling edge is received on the SYNC line. • SPREAD: Spread Spectrum Sensor Drive For the autonomous QTouch sensor, specifies that spread spectrum sensor drive shall be used. • DILEN: Discharge Length For the autonomous QTouch sensor, specifies how many sample clock cycles the CAT should use to discharge the capacitors before charging them. • MAX: Maximum Count For the autonomous QTouch sensor, specifies how many counts the maximum acquisition should be. 663 32099I–01/2012 AT32UC3L016/32/64 28.7.6 Name: Autonomous Touch Configuration Register 2 ATCFG2 Access Type: Read/Write Offset: 0x18 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 23 22 21 20 15 14 FILTER 13 12 11 10 9 8 3 2 1 0 OUTSENS 7 6 5 4 SENSE • FILTER: Autonomous Touch Filter Setting For the autonomous QTouch sensor, specifies how many positive detects in a row the CAT needs to have on the autonomous QTouch sensor before reporting it as a touch. A FILTER value of 0 is not allowed and will result in undefined behavior. • OUTSENS: Out-of-Touch Sensitivity For the autonomous QTouch sensor, specifies how sensitive the out-of-touch detector should be. • SENSE: Sensitivity For the autonomous QTouch sensor, specifies how sensitive the touch detector should be. 664 32099I–01/2012 AT32UC3L016/32/64 28.7.7 Name: Autonomous Touch Configuration Register 3 ATCFG3 Access Type: Read/Write Offset: 0x1C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 23 22 21 20 PTHR 15 14 13 12 PDRIFT 7 6 5 4 NDRIFT • PTHR: Positive Recalibration Threshold For the autonomous QTouch sensor, specifies how far a sensor’s signal must move in a positive direction from the reference in order to cause a recalibration. • PDRIFT: Positive Drift Compensation For the autonomous QTouch sensor, specifies how often a positive drift compensation should be performed. When this field is zero, positive drift compensation will never be performed. When this field is non-zero, the positive drift compensation time interval is given by the following formula: Tpdrift = PDRIFT * 65536 * (sample clock period) • NDRIFT: Negative Drift Compensation For the autonomous QTouch sensor, specifies how often a negative drift compensation should be performed. When this field is zero, negative drift compensation will never be performed. When this field is non-zero, the negative drift compensation time interval is given by the following formula: Tndrift = NDRIFT * 65536 * (sample clock period) With the typical sample clock frequency of 1 MHz, PDRIFT and NDRIFT can be set from 0.066 seconds to 16.7 seconds with 0.066 second resolution. 665 32099I–01/2012 AT32UC3L016/32/64 28.7.8 Name: Touch Group x Configuration Register 0 TGxCFG0 Access Type: Read/Write Offset: 0x20, 0x28 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 DIV[15:8] 23 22 21 20 DIV[7:0] 15 14 13 12 CHLEN 7 6 5 4 SELEN • DIV: Clock Divider The prescaler is used to ensure that the CLK_CAT clock is divided to around 1 MHz to produce the sampling clock.The prescaler uses the following formula to generate the sampling clock: Sampling clock = CLK_CAT / (2(DIV+1)) • CHLEN: Charge Length For the QTouch method, specifies how many sample clock cycles should be used for transferring charge to the sense capacitor. • SELEN: Settle Length For the QTouch method, specifies how many sample clock cycles should be used for settling after charge transfer. 666 32099I–01/2012 AT32UC3L016/32/64 28.7.9 Name: Touch Group x Configuration Register 1 TGxCFG1 Access Type: Read/Write Offset: 0x24, 0x2C Reset Value: 0x00000000 31 30 - - 23 22 29 28 DISHIFT 21 20 27 26 25 24 - - SYNC SPREAD 19 18 17 16 11 10 9 8 3 2 1 0 DILEN 15 14 13 12 MAX[15:8] 7 6 5 4 MAX[7:0] • DISHIFT: Discharge Shift For the sensors in QTouch group x, specifies how many bits the DILEN field should be shifted before using it to determine the discharge time. • SYNC: Sync Pin For sensors in QTouch group x, specifies that acquisition shall begin when a falling edge is received on the SYNC line. • SPREAD: Spread Spectrum Sensor Drive For sensors in QTouch group x, specifies that spread spectrum sensor drive shall be used. • DILEN: Discharge Length For sensors in QTouch group x, specifies how many clock cycles the CAT should use to discharge the capacitors before charging them. • MAX: Touch Maximum Count For sensors in QTouch group x, specifies how many counts the maximum acquisition should be. 667 32099I–01/2012 AT32UC3L016/32/64 28.7.10 Name: Matrix Group Configuration Register 0 MGCFG0 Access Type: Read/Write Offset: 0x30 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 DIV[15:8] 23 22 21 20 DIV[7:0] 15 14 13 12 CHLEN 7 6 5 4 SELEN • DIV: Clock Divider The prescaler is used to ensure that the CLK_CAT clock is divided to around 4 MHz to produce the burst timing clock.The prescaler uses the following formula to generate the burst timing clock: Burst timing clock = CLK_CAT / (2(DIV+1)) • CHLEN: Charge Length For QMatrix sensors, specifies how many burst prescaler clock cycles should be used for transferring charge to the sense capacitor. • SELEN: Settle Length For QMatrix sensors, specifies how many burst prescaler clock cycles should be used for settling after charge transfer. 668 32099I–01/2012 AT32UC3L016/32/64 28.7.11 Name: Matrix Group Configuration Register 1 MGCFG1 Access Type: Read/Write Offset: 0x34 Reset Value: 0x00000000 31 30 29 23 28 27 DISHIFT 22 21 26 - 20 25 24 SYNC SPREAD 19 18 17 16 11 10 9 8 3 2 1 0 DILEN 15 14 13 12 MAX[15:8] 7 6 5 4 MAX[7:0] • DISHIFT: Discharge Shift For QMatrix sensors, specifies how many bits the DILEN field should be shifted before using it to determine the discharge time. • SYNC: Sync Pin For QMatrix sensors, specifies that acquisition shall begin when a falling edge is received on the SYNC line. • SPREAD: Spread Spectrum Sensor Drive For QMatrix sensors, specifies that spread spectrum sensor drive shall be used. • DILEN: Discharge Length For QMatrix sensors, specifies how many burst prescaler clock cycles the CAT should use to discharge the capacitors at the beginning of a burst sequence. • MAX: Maximum Count For QMatrix sensors, specifies how many counts the maximum acquisition should be. 669 32099I–01/2012 AT32UC3L016/32/64 28.7.12 Name: Matrix Group Configuration Register 2 MGCFG2 Access Type: Read/Write Offset: 0x38 Reset Value: 0x00000000 31 30 29 ACCTRL 23 28 27 26 CONSEN 22 21 25 24 20 19 18 17 16 11 10 9 8 CXDILEN 15 14 13 12 7 6 SYNCTIM[11:8] 5 4 3 2 1 0 SYNCTIM[7:0] • ACCTRL: Analog Comparator Control When written to one, allows the CAT to disable the analog comparators when they are not needed. When written to zero, the analog comparators are always enabled. • CONSEN: Consensus Filter Length For QMatrix sensors, specifies that discharge will be terminated when CONSEN out of the most recent 5 comparator samples are positive. For example, a value of 3 in the CONSEN field will terminate discharge when 3 out of the most recent 5 comparator samples are positive. When CONSEN has the default value of 0, discharge will be terminated immediately when the comparator output goes positive. • CXDILEN: Cx Capacitor Discharge Length For QMatrix sensors, specifies how many burst prescaler clock cycles the CAT should use to discharge the Cx capacitor at the end of each burst cycle. • SYNCTIM: Sync Time Interval When non-zero, determines the number of prescaled clock cycles between the start of the acquisition on each X line for QMatrix acquisition. 670 32099I–01/2012 AT32UC3L016/32/64 28.7.13 Name: Status Register SR Access Type: Read-only Offset: 0x3C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - ACQDONE ACREADY 7 6 5 4 3 2 1 0 - - - MBLREQ ATSTATE ATSC ATCAL ENABLED • ACQDONE: Acquisition Done 0: Acquisition is not done (still in progress). 1: Acquisition is complete. • ACREADY: Acquired Count Data is Ready 0: Acquired count data is not available in the ACOUNT register. 1: Acquired count data is available in the ACOUNT register. • MBLREQ: Matrix Burst Length Required 0: The QMatrix acquisition does not require any burst lengths. 1: The QMatrix acquisition requires burst lengths for the current X line. • ATSTATE: Autonomous Touch Sensor State 0: The autonomous QTouch sensor is not active. 1: The autonomous QTouch sensor is active. • ATSC: Autonomous Touch Sensor Status Interrupt 0: No status change in the autonomous QTouch sensor. 1: Status change in the autonomous QTouch sensor. • ATCAL: Autonomous Touch Calibration Ongoing 0: The autonomous QTouch sensor is not calibrating. 1: The autonomous QTouch sensor is calibrating. • ENABLED: Module Enabled 0: The module is disabled. 1: The module is enabled. 671 32099I–01/2012 AT32UC3L016/32/64 28.7.14 Name: Status Clear Register SCR Access Type: Write-only Offset: 0x40 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - ACQDONE ACREADY 7 6 5 4 3 2 1 0 - - - - - ATSC ATCAL - Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in SR and the corresponding interrupt request. 672 32099I–01/2012 AT32UC3L016/32/64 28.7.15 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x44 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - ACQDONE ACREADY 7 6 5 4 3 2 1 0 - - - - - ATSC ATCAL - Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 673 32099I–01/2012 AT32UC3L016/32/64 28.7.16 Name: Interrupt Disable Register IDR Access Type: Write-only Offset: 0x48 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - ACQDONE ACREADY 7 6 5 4 3 2 1 0 - - - - - ATSC ATCAL - Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 674 32099I–01/2012 AT32UC3L016/32/64 28.7.17 Name: Interrupt Mask Register IMR Access Type: Read-only Offset: 0x4C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - ACQDONE ACREADY 7 6 5 4 3 2 1 0 - - - - - ATSC ATCAL - 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. A bit in this register is cleared when the corresponding bit in IDR is written to one. A bit in this register is set when the corresponding bit in IER is written to one. 675 32099I–01/2012 AT32UC3L016/32/64 28.7.18 Name: Acquisition Initiation and Selection Register AISR Access Type: Read/Write Offset: 0x50 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 23 22 21 20 - 15 14 13 12 - 7 6 5 4 - ACQSEL • ACQSEL: Acquisition Type Selection A write to this register initiates an acquisition of the following type: 00: QTouch Group A. 01: QTouch Group B. 10: QMatrix Group. 11: Undefined behavior. A read of this register will return the value that was previously written. 676 32099I–01/2012 AT32UC3L016/32/64 28.7.19 Name: Acquired Count Register ACOUNT Access Type: Read-only Offset: 0x54 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 Y 23 22 21 20 SPORX 15 14 13 12 COUNT[15:8] 7 6 5 4 COUNT[7:0] • Y: Y index The Y index (for QMatrix method) associated with this count value. • SPORX: Sensor pair or X index The sensor pair index (for QTouch method) or X index (for QMatrix method) associated with this count value. • COUNT: Count value The signal (number of counts) acquired on the channel specified in the SPORX and Y fields. When multiple acquired count values are read from a QTouch acquisition, the Y field will always be 0 and the SPORX value will increase monotonically. For example, suppose a QTouch acquisition is performed using sensor pairs SP1, SP4, and SP9. The first count read will have SPORX=1, the second read will have SPORX=4, and the third read will have SPORX=9. When multiple acquired count values are read from a QMatrix acquisition, the SPORX value will stay the same while Y increases monotonically through all Y values in the group. Then SPORX will increase to the next X value in the group. For example, a QMatrix acquisition with X=2,3 and Y=4,7 would provide count values in the following order: X=2 and Y=4, then X=2 and Y=7, then X=3 and Y=4, and finally X=3 and Y=7. 677 32099I–01/2012 AT32UC3L016/32/64 28.7.20 Name: Matrix Burst Length Register MBLEN Access Type: Write-only Offset: 0x58 Reset Value: - 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 BURST0 23 22 21 20 BURST1 15 14 13 12 BURST2 7 6 5 4 BURST3 • BURSTx: Burst Length x For QMatrix sensors, specifies how many times the switching sequence should be repeated before acquisition begins for each channel. Each count in the BURSTx field specifies 1 repeat of the switching sequence, so the actual burst length will be BURST. Before doing a QMatrix acquisition on one X line this register has to be written with the burst values for the current XY pairs. For each X line this register needs to be programmed with all the Y values. If Y values larger than 3 are used the register has to be written several times in order to specify all burst lengths. The Status Register bit MBLREQ is set to 1 when the CAT is waiting for values to be written into this register. 678 32099I–01/2012 AT32UC3L016/32/64 28.7.21 Name: Discharge Current Source Register DICS Access Type: Read/Write Offset: 0x5C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 FSOURCES[7:0] 23 22 21 20 19 18 17 16 GLEN - - - - - INTVREFSEL INTREFSEL 15 14 13 12 11 10 9 8 - - - 7 6 5 1 0 TRIM 4 3 2 SOURCES[7:0] • FSOURCES: Force Discharge Current Sources When FSOURCES[n] is 0, the corresponding discharge current source behavior depends on SOURCES[n]. When FSOURCES[n] is 1, the corresponding discharge current source is forced to be enabled continuously. This is useful for testing or debugging but should not be done during normal acquisition. • GLEN: Global Enable 0: The current source module is globally disabled. 1: The current source module is globally enabled. • INTVREFSEL: Internal Voltage Reference Select 0: The voltage for the reference resistor is generated from the internal band gap circuit. 1: The voltage for the reference resistor is VDDIO/3. • INTREFSEL: Internal Reference Select 0: The reference current flows through an external resistor on the DIS pin. 1: The reference current flows through the internal reference resistor. • TRIM: Reference Current Trimming This field is used to trim the discharge current. 0x00 corresponds to the minimum current value, and 0x1F corresponds to the maximum current value. • SOURCES: Enable Discharge Current Sources When SOURCES[n] is 0, the corresponding discharge current source is disabled. When SOURCES[n] is 1, the corresponding discharge current source is enabled at appropriate times during acquisition. 679 32099I–01/2012 AT32UC3L016/32/64 28.7.22 Name: Spread Spectrum Configuration Register SSCFG Access Type: Read/Write Offset: 0x60 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 MAXDEV • MAXDEV: Maximum Deviation When spread spectrum burst is enabled, MAXDEV indicates the maximum number of prescaled clock cycles the burst pulse will be extended or shortened. 680 32099I–01/2012 AT32UC3L016/32/64 28.7.23 Name: CSA Resistor Control Register CSARES Access Type: Read/Write Offset: 0x64 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 23 22 21 20 - 15 14 13 RES[16] 12 11 10 9 8 3 2 1 0 RES[15:8] 7 6 5 4 RES[7:0] • RES: Resistive Drive Enable When RES[n] is 0, CSA[n] has the same drive properties as normal I/O pads. When RES[n] is 1, CSA[n] has a nominal output resistance of 1kOhm during the burst phase. 681 32099I–01/2012 AT32UC3L016/32/64 28.7.24 Name: CSB Resistor Control Register CSBRES Access Type: Read/Write Offset: 0x68 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 23 22 21 20 - 15 14 13 RES[16] 12 11 10 9 8 3 2 1 0 RES[15:8] 7 6 5 4 RES[7:0] • RES: Resistive Drive Enable When RES[n] is 0, CSB[n] has the same drive properties as normal I/O pads. When RES[n] is 1, CSB[n] has a nominal output resistance of 1kOhm during the burst phase. 682 32099I–01/2012 AT32UC3L016/32/64 28.7.25 Name: Autonomous Touch Base Count Register ATBASE Access Type: Read-only Offset: 0x6C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 23 22 21 20 - 15 14 13 12 COUNT[15:8] 7 6 5 4 COUNT[7:0] • COUNT: Count value The base count currently stored by the autonomous touch sensor. This is useful for autonomous touch debugging purposes. 683 32099I–01/2012 AT32UC3L016/32/64 28.7.26 Name: Autonomous Touch Current Count Register ATCURR Access Type: Read-only Offset: 0x70 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 23 22 21 20 - 15 14 13 12 COUNT[15:8] 7 6 5 4 COUNT[7:0] • COUNT: Count value The current count acquired by the autonomous touch sensor. This is useful for autonomous touch debugging purposes. 684 32099I–01/2012 AT32UC3L016/32/64 28.7.27 Name: Analog Comparator Shift Offset Register x ACSHIx Access Type: Read/Write Offset: 0x80, 0x84, 0x88, 0x8C, 0x90, 0x94, 0x98, and 0x9C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 1 0 23 22 21 20 - 15 14 13 12 7 6 SHIVAL[11:8] 5 4 3 2 SHIVAL[7:0] • SHIVAL: Shift Offset Value Specifies the amount to shift the count value from each comparator. This allows the offset of each comparator to be compensated. 685 32099I–01/2012 AT32UC3L016/32/64 28.7.28 Name: Parameter Register PARAMETER Access Type: Read-only Offset: 0xF8 Reset Value: - 31 30 29 28 27 26 25 24 19 18 17 16 11 10 9 8 3 2 1 0 23 22 21 20 SP[23:16] 15 14 13 12 SP[15:8] 7 6 5 4 SP[7:0] • SP[n]: Sensor pair implemented 0: The corresponding sensor pair is not implemented 1: The corresponding sensor pair is implemented. 686 32099I–01/2012 AT32UC3L016/32/64 28.7.29 Name: Version Register VERSION Access Type: Read-only Offset: 0xFC Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant number Reserved. No functionality associated. • VERSION: Version number Version number of the module. No functionality associated. 687 32099I–01/2012 AT32UC3L016/32/64 28.8 Module Configuration The specific configuration the CAT module is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 28-4. Feature CAT Number of touch sensors/Size of matrix Allows up to 17 touch sensors, or up to 16 by 8 matrix sensors to be interfaced. Table 28-5. CAT Clocks Clock Name Description CLK_CAT Clock for the CAT bus interface GCLK The generic clock used for the CAT is GCLK4 Table 28-6. 28.8.1 CAT Configuration Register Reset Values Register Reset Value VERSION 0x00000200 PARAMETER 0x0001FFFF Resistive Drive By default, the CAT pins are driven with normal I/O drive properties. Some of the CSA and CSB pins can optionally drive with a 1k output resistance for improved EMC. To enable resistive drive on a pin, the user must write a one to the corresponding bit in the CSA Resistor Control Register (CSARES) or CSB Resistor Control Register (CSBRES) register. 688 32099I–01/2012 AT32UC3L016/32/64 29. Glue Logic Controller (GLOC) Rev: 1.0.0.0 29.1 Features • • • • 29.2 Glue logic for general purpose PCB design Programmable lookup table Up to four inputs supported per lookup table Optional filtering of output Overview The Glue Logic Controller (GLOC) contains programmable logic which can be connected to the device pins. This allows the user to eliminate logic gates for simple glue logic functions on the PCB. The GLOC consists of a number of lookup table (LUT) units. Each LUT can generate an output as a user programmable logic expression with four inputs. Inputs can be individually masked. The output can be combinatorially generated from the inputs, or filtered to remove spikes. 29.3 Block Diagram TRUTH PERIPHERAL BUS Figure 29-1. GLOC Block Diagram OUT[0] ... OUT[n] FILTER FILTEN GCLK AEN CLK_GLOC IN[3:0] … IN[(4n+3):4n] 689 32099I–01/2012 AT32UC3L016/32/64 29.4 I/O Lines Description Table 29-1. I/O Lines Description Pin Name Pin Description Type IN0-INm Inputs to lookup tables Input OUT0-OUTn Output from lookup tables Output Each LUT have 4 inputs and one output. The inputs and outputs for the LUTs are mapped sequentially to the inputs and outputs. This means that LUT0 is connected to IN0 to IN3 and OUT0. LUT1 is connected to IN4 to IN7 and OUT1. In general, LUTn is connected to IN[4n] to IN[4n+3] and OUTn. 29.5 Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 29.5.1 I/O Lines The pins used for interfacing the GLOC may be multiplexed with I/O Controller lines. The programmer must first program the I/O Controller to assign the desired GLOC pins to their peripheral function. If I/O lines of the GLOC are not used by the application, they can be used for other purposes by the I/O Controller. It is only required to enable the GLOC inputs and outputs actually in use. Pullups for pins configured to be used by the GLOC will be disabled. 29.5.2 Clocks The clock for the GLOC bus interface (CLK_GLOC) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the GLOC before disabling the clock, to avoid freezing the module in an undefined state. Additionally, the GLOC depends on a dedicated Generic Clock (GCLK). The GCLK can be set to a wide range of frequencies and clock sources, and must be enabled by the System Control Interface (SCIF) before the GLOC filter can be used. 29.5.3 Debug Operation When an external debugger forces the CPU into debug mode, the GLOC continues normal operation. 29.6 29.6.1 Functional Description Enabling the Lookup Table Inputs Since the inputs to each lookup table (LUT) unit can be multiplexed with other peripherals, each input must be explicitly enabled by writing a one to the corresponding enable bit (AEN) in the corresponding Control Register (CR). If no inputs are enabled, the output OUTn will be the least significant bit in the TRUTHn register. 690 32099I–01/2012 AT32UC3L016/32/64 29.6.2 Configuring the Lookup Table The lookup table in each LUT unit can generate any logic expression OUT as a function of up to four inputs, IN[3:0]. The truth table for the expression is written to the TRUTH register for the LUT. Table 29-2 shows the truth table for LUT0. The truth table for LUTn is written to TRUTHn, and the corresponding input and outputs will be IN[4n] to IN[4n+3] and OUTn. Table 29-2. 29.6.3 Truth Table for the Lookup Table in LUT0 IN[3] IN[2] IN[1] IN[0] OUT[0] 0 0 0 0 TRUTH0[0] 0 0 0 1 TRUTH0[1] 0 0 1 0 TRUTH0[2] 0 0 1 1 TRUTH0[3] 0 1 0 0 TRUTH0[4] 0 1 0 1 TRUTH0[5] 0 1 1 0 TRUTH0[6] 0 1 1 1 TRUTH0[7] 1 0 0 0 TRUTH0[8] 1 0 0 1 TRUTH0[9] 1 0 1 0 TRUTH0[10] 1 0 1 1 TRUTH0[11] 1 1 0 0 TRUTH0[12] 1 1 0 1 TRUTH0[13] 1 1 1 0 TRUTH0[14] 1 1 1 1 TRUTH0[15] Output Filter By default, the output OUTn is a combinatorial function of the inputs IN[4n] to IN[4n+3]. This may cause some short glitches to occur when the inputs change value. It is also possible to clock the output through a filter to remove glitches. This requires that the corresponding generic clock (GCLK) has been enabled before use. The filter can then be enabled by writing a one to the Filter Enable (FILTEN) bit in CRn. The OUTn output will be delayed by three to four GCLK cycles when the filter is enabled. 691 32099I–01/2012 AT32UC3L016/32/64 29.7 User Interface Table 29-3. GLOC Register Memory Map Offset Register Register Name Access Reset 0x00+n*0x08 Control Register n CRn Read/Write 0x00000000 0x04+n*0x08 Truth Table Register n TRUTHn Read/Write 0x00000000 0x38 Parameter Register PARAMETER Read-only - (1) 0x3C Version Register VERSION Read-only - (1) Note: 1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter. 692 32099I–01/2012 AT32UC3L016/32/64 29.7.1 Name: Control Register n CRn Access Type: Read/Write Offset: 0x00+n*0x08 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 FILTEN - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - AEN • FILTEN: Filter Enable 1: The output is glitch filtered 0: The output is not glitch filtered • AEN: Enable IN Inputs Input IN[n] is enabled when AEN[n] is one. Input IN[n] is disabled when AEN[n] is zero, and will not affect the OUT value. 693 32099I–01/2012 AT32UC3L016/32/64 29.7.2 Name: Truth Table Register n TRUTHn Access Type: Read/Write Offset: 0x04+n*0x08 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 TRUTH[15:8] 7 6 5 4 TRUTH[7:0] • TRUTH: Truth Table Value This value defines the output OUT as a function of inputs IN: OUT = TRUTH[IN] 694 32099I–01/2012 AT32UC3L016/32/64 29.7.3 Name: Parameter Register PARAMETER Access Type: Read-only Offset: 0x38 Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 LUTS • LUTS: Lookup Table Units Implemented This field contains the number of lookup table units implemented in this device. 695 32099I–01/2012 AT32UC3L016/32/64 29.7.4 Name: VERSION Register VERSION Access Type: Read-only Offset: 0x3C Reset Value: - 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - 15 14 13 12 9 8 - - - - 7 6 5 4 VARIANT 11 10 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VARIANT: Variant Number Reserved. No functionality associated. • VERSION: Version Number Version number of the module. No functionality associated. 696 32099I–01/2012 AT32UC3L016/32/64 29.8 Module Configuration The specific configuration for each GLOC instance is listed in the following tables.The GLOC bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 29-4. GLOC Configuration Feature GLOC Number of LUT units 2 Table 29-5. GLOC Clocks Clock Name Description CLK_GLOC Clock for the GLOC bus interface GCLK The generic clock used for the GLOC is GCLK5 Table 29-6. Register Reset Values Register Reset Value VERSION 0x00000100 PARAMETER 0x00000002 697 32099I–01/2012 AT32UC3L016/32/64 30. aWire UART (AW) Rev: 2.1.0.0 30.1 Features • Asynchronous receiver or transmitter when the aWire system is not used for debugging. • One- or two-pin operation supported. 30.2 Overview If the AW is not used for debugging, the aWire UART can be used by the user to send or receive data with one start bit, eight data bits, no parity bits, and one stop bit. This can be controlled through the aWire UART user interface. This chapter only describes the aWire UART user interface. For a description of the aWire Debug Interface, please see the Programming and Debugging chapter. 30.3 Block Diagram Figure 30-1. aWire Debug Interface Block Diagram PB aWire Debug Interface Flash Controller CHIP_ERASE command AW User Interface CPU HALT command RESET command Power Manager External reset AW_ENABLE AW CONTROL Reset filter RESET_N Baudrate Detector SAB interface UART RW SZ ADDR DATA CRC SAB 698 32099I–01/2012 AT32UC3L016/32/64 30.4 I/O Lines Description Table 30-1. I/O Lines Description Name Description Type DATA aWire data multiplexed with the RESET_N pin. Input/Output 30.5 Product Dependencies In order to use this module, other parts of the system must be configured correctly, as described below. 30.5.1 I/O Lines The pin used by AW is multiplexed with the RESET_N pin. The reset functionality is the default function of this pin. To enable the aWire functionality on the RESET_N pin the user must enable the aWire UART user interface. 30.5.2 Power Management If the CPU enters a sleep mode that disables clocks used by the aWire UART user interface, the aWire UART user interface will stop functioning and resume operation after the system wakes up from sleep mode. 30.5.3 Clocks The aWire UART uses the internal 120 MHz RC oscillator (RC120M) as clock source for its operation. When using the aWire UART user interface RC120M must enabled using the Clock Request Register (see Section 30.6.1). The clock for the aWire UART user interface (CLK_AW) is generated by the Power Manager. This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the aWire UART user interface before disabling the clock, to avoid freezing the aWire UART user interface in an undefined state. 30.5.4 Interrupts The aWire UART user interface interrupt request line is connected to the interrupt controller. Using the aWire UART user interface interrupt requires the interrupt controller to be programmed first. 30.5.5 Debug Operation If the AW is used for debugging the aWire UART user interface will not be usable. When an external debugger forces the CPU into debug mode, the aWire UART user interface continues normal operation. If the aWire UART user interface is configured in a way that requires it to be periodically serviced by the CPU through interrupts or similar, improper operation or data loss may result during debugging. 30.5.6 30.6 External Components The AW needs an external pullup on the RESET_N pin to ensure that the pin is pulled up when the bus is not driven. Functional Description The aWire UART user interface can be used as a spare Asynchronous Receiver or Transmitter when AW is not used for debugging. 699 32099I–01/2012 AT32UC3L016/32/64 30.6.1 How to Initialize The Module To initialize the aWire UART user interface the user must first enable the clock by writing a one to the Clock Enable bit in the Clock Request Register (CLKR.CLKEN) and wait for the Clock Enable bit in the Status Register (SR.CENABLED) to be set. After doing this either receive, transmit or receive with resync must be selected by writing the corresponding value into the Mode field of the Control (CTRL.MODE) Register. Due to the RC120M being asynchronous with the system clock values must be allowed to propagate in the system. During this time the aWire master will set the Busy bit in the Status Register (SR.BUSY). After the SR.BUSY bit is cleared the Baud Rate field in the Baud Rate Register (BRR.BR) can be written with the wanted baudrate ( f br ) according to the following formula ( f aw is the RC120M clock frequency): 8f aw f br = ----------BR After this operation the user must wait until the SR.BUSY is cleared. The interface is now ready to be used. 30.6.2 Basic Asynchronous Receiver Operation The aWire UART user interface must be initialized according to the sequence above, but the CTRL.MODE field must be written to one (Receive mode). When a data byte arrives the aWire UART user interface will indicate this by setting the Data Ready Interrupt bit in the Status Register (SR.DREADYINT). The user must read the Data in the Receive Holding Register (RHR.RXDATA) and clear the Interrupt bit by writing a one to the Data Ready Interrupt Clear bit in the Status Clear Register (SCR.DREADYINT). The interface is now ready to receive another byte. 30.6.3 Basic Asynchronous Transmitter Operation The aWire UART user interface must be initialized according to the sequence above, but the CTRL.MODE field must be written to two (Transmit mode). To transmit a data byte the user must write the data to the Transmit Holding Register (THE.TXDATA). Before the next byte can be written the SR.BUSY must be cleared. 30.6.4 Basic Asynchronous Receiver with Resynchronization By writing three into CTRL.MODE the aWire UART user interface will assume that the first byte it receives is a sync byte (0x55) and set BRR.BR according to this. All subsequent transfers will assume this baudrate, unless BRR.BR is rewritten by the user. To make the aWire UART user interface accept a new sync resynchronization the aWire UART user interface must be disabled by writing zero to CTRL.MODE and then reenable the interface. 30.6.5 Overrun In Receive mode an overrun can occur if the user has not read the previous received data from the RHR.RXDATA when the newest data should be placed there. Such a condition is flagged by setting the Overrun bit in the Status Register (SR.OVERRUN). If SR.OVERRUN is set the newest data received is placed in RHR.RXDATA and the data that was there before is overwritten. 700 32099I–01/2012 AT32UC3L016/32/64 30.6.6 Interrupts To make the CPU able to do other things while waiting for the aWire UART user interface to finish its operations the aWire UART user interface supports generating interrupts. All status bits in the Status Register can be used as interrupt sources, except the SR.BUSY and SR.CENABLED bits. To enable an interrupt the user must write a one to the corresponding bit in the Interrupt Enable Register (IER). Upon the next zero to one transition of this SR bit the aWire UART user interface will flag this interrupt to the CPU. To clear the interrupt the user must write a one to the corresponding bit in the Status Clear Register (SCR). Interrupts can be disabled by writing a one to the corresponding bit in the Interrupt Disable Register (IDR). The interrupt Mask Register (IMR) can be read to check if an interrupt is enabled or disabled. 30.6.7 Using the Peripheral DMA Controller To relieve the CPU of data transfers the aWire UART user interface support using the Peripheral DMA controller. To transmit using the Peripheral DMA Controller do the following: 1. Setup the aWire UART user interface in transmit mode. 2. Setup the Peripheral DMA Controller with buffer address and length, use byte as transfer size. 3. Enable the Peripheral DMA Controller. 4. Wait until the Peripheral DMA Controller is done. To receive using the Peripheral DMA Controller do the following: 1. Setup the aWire UART user interface in receive mode 2. Setup the Peripheral DMA Controller with buffer address and length, use byte as transfer size. 3. Enable the Peripheral DMA Controller. 4. Wait until the Peripheral DMA Controller is ready. 701 32099I–01/2012 AT32UC3L016/32/64 30.7 User Interface Table 30-2. Note: aWire UART user interface Register Memory Map Offset Register Register Name Access Reset 0x00 Control Register CTRL Read/Write 0x00000000 0x04 Status Register SR Read-only 0x00000000 0x08 Status Clear Register SCR Write-only - 0x0C Interrupt Enable Register IER Write-only - 0x10 Interrupt Disable Register IDR Write-only - 0x14 Interrupt Mask Register IMR Read-only 0x00000000 0x18 Receive Holding Register RHR Read-only 0x00000000 0x1C Transmit Holding Register THR Read/Write 0x00000000 0x20 Baud Rate Register BRR Read/Write 0x00000000 0x24 Version Register VERSION Read-only -(1) 0x28 Clock Request Register CLKR Read/Write 0x00000000 1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter. 702 32099I–01/2012 AT32UC3L016/32/64 30.7.1 Name: Control Register CTRL Access Type: Read/Write Offset: 0x00 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - MODE • MODE: aWire UART user interface mode Table 30-3. aWire UART user interface Modes MODE Mode Description 0 Disabled 1 Receive 2 Transmit 3 Receive with resync. 703 32099I–01/2012 AT32UC3L016/32/64 30.7.2 Name: Status Register SR Access Type: Read-only Offset: 0x04 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - TRMIS - - OVERRUN DREADYINT READYINT 7 6 5 4 3 2 1 0 - - - - - CENABLED - BUSY • TRMIS: Transmit Mismatch 0: No transfers mismatches. 1: The transceiver was active when receiving. This bit is set when the transceiver is active when receiving. This bit is cleared when corresponding bit in SCR is written to one. • OVERRUN: Data Overrun 0: No data overwritten in RHR. 1: Data in RHR has been overwritten before it has been read. This bit is set when data in RHR is overwritten before it has been read. This bit is cleared when corresponding bit in SCR is written to one. • DREADYINT: Data Ready Interrupt 0: No new data in the RHR. 1: New data received and placed in the RHR. This bit is set when new data is received and placed in the RHR. This bit is cleared when corresponding bit in SCR is written to one. • READYINT: Ready Interrupt 0: The interface has not generated an ready interrupt. 1: The interface has had a transition from busy to not busy. This bit is set when the interface has transition from busy to not busy. This bit is cleared when corresponding bit in SCR is written to one. • CENABLED: Clock Enabled 0: The aWire clock is not enabled. 1: The aWire clock is enabled. 704 32099I–01/2012 AT32UC3L016/32/64 This bit is set when the clock is disabled. This bit is cleared when the clock is enabled. • BUSY: Synchronizer Busy 0: The asynchronous interface is ready to accept more data. 1: The asynchronous interface is busy and will block writes to CTRL, BRR, and THR. This bit is set when the asynchronous interface becomes busy. This bit is cleared when the asynchronous interface becomes ready. 705 32099I–01/2012 AT32UC3L016/32/64 30.7.3 Name: Status Clear Register SCR Access Type: Write-only Offset: 0x08 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - TRMIS - - OVERRUN DREADYINT READYINT 7 6 5 4 3 2 1 0 - - - - - - - - Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in SR and the corresponding interrupt request. 706 32099I–01/2012 AT32UC3L016/32/64 30.7.4 Name: Interrupt Enable Register IER Access Type: Write-only Offset: 0x0C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - TRMIS - - OVERRUN DREADYINT READYINT 7 6 5 4 3 2 1 0 - - - - - - - - Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will set the corresponding bit in IMR. 707 32099I–01/2012 AT32UC3L016/32/64 30.7.5 Name: Interrupt Disable Register IDR Access Type: Write-only Offset: 0x10 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - TRMIS - - OVERRUN DREADYINT READYINT 7 6 5 4 3 2 1 0 - - - - - - - - Writing a zero to a bit in this register has no effect. Writing a one to a bit in this register will clear the corresponding bit in IMR. 708 32099I–01/2012 AT32UC3L016/32/64 30.7.6 Name: Interrupt Mask Register IMR Access Type: Read-only Offset: 0x14 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - TRMIS - - OVERRUN DREADYINT READYINT 7 6 5 4 3 2 1 0 - - - - - - - - 0: The corresponding interrupt is disabled. 1: The corresponding interrupt is enabled. A bit in this register is cleared when the corresponding bit in IDR is written to one. A bit in this register is set when the corresponding bit in IER is written to one. 709 32099I–01/2012 AT32UC3L016/32/64 30.7.7 Name: Receive Holding Register RHR Access Type: Read-only Offset: 0x18 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 RXDATA • RXDATA: Received Data The last byte received. 710 32099I–01/2012 AT32UC3L016/32/64 30.7.8 Name: Transmit Holding Register THR Access Type: Read/Write Offset: 0x1C Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 TXDATA • TXDATA: Transmit Data The data to send. 711 32099I–01/2012 AT32UC3L016/32/64 30.7.9 Name: Baud Rate Register BRR Access Type: Read/Write Offset: 0x20 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 3 2 1 0 BR[15:8] 7 6 5 4 BR[7:0] • BR: Baud Rate The baud rate ( f br ) of the transmission, calculated using the following formula ( f aw is the RC120M frequency): 8f aw f br = ----------BR BR should not be set to a value smaller than 32. Writing a value to this field will update the baud rate of the transmission. Reading this field will give the current baud rate of the transmission. 712 32099I–01/2012 AT32UC3L016/32/64 30.7.10 Name: Version Register VERSION Access Type: Read-only Offset: 0x24 Reset Value: 0x00000200 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - 7 6 5 4 VERSION[11:8] 3 2 1 0 VERSION[7:0] • VERSION: Version Number Version number of the module. No functionality associated. 713 32099I–01/2012 AT32UC3L016/32/64 30.7.11 Name: Clock Request Register CLKR Access Type: Read/Write Offset: 0x28 Reset Value: 0x00000000 31 30 29 28 27 26 25 24 - - - - - - - - 23 22 21 20 19 18 17 16 - - - - - - - - 15 14 13 12 11 10 9 8 - - - - - - - - 7 6 5 4 3 2 1 0 - - - - - - - CLKEN • CLKEN: Clock Enable 0: The aWire clock is disabled. 1: The aWire clock is enabled. Writing a zero to this bit will disable the aWire clock. Writing a one to this bit will enable the aWire clock. 714 32099I–01/2012 AT32UC3L016/32/64 30.8 Module Configuration The specific configuration for each aWire instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details. Table 30-4. AW Clocks Clock Name Description CLK_AW Clock for the AW bus interface Table 30-5. Register Reset Values Register Reset Value VERSION 0x00000210 715 32099I–01/2012 AT32UC3L016/32/64 31. Programming and Debugging 31.1 Overview The AT32UC3L016/32/64 supports programming and debugging through two interfaces, JTAG and aWire. JTAG is an industry standard interface and allows boundary scan for PCB testing, as well as daisy-chaining of multiple devices on the PCB. aWire is an Atmel proprietary protocol which offers higher throughput and robust communication, and does not require application pins to be reserved. Either interface provides access to the internal Service Access Bus (SAB), which offers a bridge to the High Speed Bus, giving access to memories and peripherals in the device. By using this bridge to the bus system, the flash and fuses can thus be programmed by accessing the Flash Controller in the same manner as the CPU. The SAB also provides access to the Nexus-compliant On-Chip Debug (OCD) system in the device, which gives the user non-intrusive run-time control of the program execution. Additionally, trace information can be output on the Auxiliary (AUX) debug port or buffered in internal RAM for later retrieval by JTAG or aWire. 31.2 Service Access Bus The AVR32 architecture offers a common interface for access to On-Chip Debug, programming, and test functions. These are mapped on a common bus called the Service Access Bus (SAB), which is linked to the JTAG and aWire port through a bus master module, which also handles synchronization between the debugger and SAB clocks. When accessing the SAB through the debugger there are no limitations on debugger frequency compared to device frequency, although there must be an active system clock in order for the SAB accesses to complete. If the system clock is switched off in sleep mode, activity on the debugger will restart the system clock automatically, without waking the device from sleep. Debuggers may optimize the transfer rate by adjusting the frequency in relation to the system clock. This ratio can be measured with debug protocol specific instructions. The Service Access Bus uses 36 address bits to address memory or registers in any of the slaves on the bus. The bus supports sized accesses of bytes (8 bits), halfwords (16 bits), or words (32 bits). All accesses must be aligned to the size of the access, i.e. halfword accesses must have the lowest address bit cleared, and word accesses must have the two lowest address bits cleared. 31.2.1 SAB Address Map The SAB gives the user access to the internal address space and other features through a 36 bits address space. The 4 MSBs identify the slave number, while the 32 LSBs are decoded within the slave’s address space. The SAB slaves are shown in Table 31-1. Table 31-1. SAB Slaves, Addresses and Descriptions Slave Address [35:32] Description Unallocated 0x0 Intentionally unallocated OCD 0x1 OCD registers HSB 0x4 HSB memory space, as seen by the CPU 716 32099I–01/2012 AT32UC3L016/32/64 Table 31-1. 31.2.2 SAB Slaves, Addresses and Descriptions Slave Address [35:32] Description HSB 0x5 Alternative mapping for HSB space, for compatibility with other 32-bit AVR devices. Memory Service Unit 0x6 Memory Service Unit registers Reserved Other Unused SAB Security Restrictions The Service Access bus can be restricted by internal security measures. A short description of the security measures are found in the table below. 31.2.2.1 Security measure and control location A security measure is a mechanism to either block or allow SAB access to a certain address or address range. A security measure is enabled or disabled by one or several control signals. This is called the control location for the security measure. These security measures can be used to prevent an end user from reading out the code programmed in the flash, for instance. Table 31-2. SAB Security Measures Security Measure Control Location Description Secure mode FLASHCDW SECURE bits set Allocates a portion of the flash for secure code. This code cannot be read or debugged. The User page is also locked. Security bit FLASHCDW security bit set Programming and debugging not possible, very restricted access. User code programming FLASHCDW UPROT + security bit set Restricts all access except parts of the flash and the flash controller for programming user code. Debugging is not possible unless an OS running from the secure part of the flash supports it. Below follows a more in depth description of what locations are accessible when the security measures are active. Table 31-3. Secure Mode SAB Restrictions Name Address Start Address End Access Secure flash area 0x580000000 0x580000000 + (USERPAGE[15:0] 1.62V 470 Cbus = 400pF, VVDD > 2.0V Max Units ns 400 kHz Note: 1. VVDD corresponds to either VVDDIN or VVDDIO, depending on the supply for the pin. Refer to Section 3.2 on page 9 for details. 32.6 Oscillator Characteristics 32.6.1 Oscillator 0 (OSC0) Characteristics 32.6.1.1 Digital Clock Characteristics The following table describes the characteristics for the oscillator when a digital clock is applied on XIN. Table 32-11. Digital Clock Characteristics Symbol Parameter fCPXIN XIN clock frequency tCPXIN XIN clock duty cycle tSTARTUP Startup time CIN 32.6.1.2 XIN input capacitance Conditions Min Typ Max 40 Units 50 MHz 60 % 0 TQFP48 package 7.0 QFN48 package 6.7 TLLGA48 package 6.7 cycles pF Crystal Oscillator Characteristics The following table describes the characteristics for the oscillator when a crystal is connected between XIN and XOUT as shown in Figure 32-3. The user must choose a crystal oscillator where the crystal load capacitance CL is within the range given in the table. The exact value of CL can be found in the crystal datasheet. The capacitance of the external capacitors (CLEXT) can then be computed as follows: C LEXT = 2 ( C L – C i ) – C PCB where CPCB is the capacitance of the PCB and Ci is the internal equivalent load capacitance. Table 32-12. Crystal Oscillator Characteristics Symbol Parameter fOUT Crystal oscillator frequency CL Crystal load capacitance Ci Internal equivalent load capacitance Conditions Min Typ Max Unit 0.45 10 16 MHz 6 18 pF 2 779 32099I–01/2012 AT32UC3L016/32/64 Table 32-12. Crystal Oscillator Characteristics Symbol Parameter tSTARTUP Conditions Min (1) Startup time SCIF.OSCCTRL.GAIN = 2 Current consumption Notes: Max (2) 30 000 Active mode, f = 0.45MHz, SCIF.OSCCTRL.GAIN = 0 IOSC Typ Unit cycles 30 µA Active mode, f = 10MHz, SCIF.OSCCTRL.GAIN = 2 170 1. Please refer to the SCIF chapter for details. 2. Nominal crystal cycles. 3. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. Figure 32-3. Oscillator Connection CLEXT XOUT UC3L Ci CL XIN CLEXT 32.6.2 32KHz Crystal Oscillator (OSC32K) Characteristics Figure 32-3 and the equation above also applies to the 32KHz oscillator connection. The user must choose a crystal oscillator where the crystal load capacitance CL is within the range given in the table. The exact value of CL can then be found in the crystal datasheet. Table 32-13. 32 KHz Crystal Oscillator Characteristics Symbol Parameter fOUT Crystal oscillator frequency tSTARTUP Startup time CL Crystal load capacitance(2) Ci Internal equivalent load capacitance IOSC32 Current consumption RS Equivalent series resistance Conditions Min Typ Max 32 768 Hz (1) RS = 60kOhm, CL = 9pF 30 000 6 Unit cycles 12.5 pF 2 0.9 (2) 32 768Hz 35 µA 85 kOhm 780 32099I–01/2012 AT32UC3L016/32/64 Notes: 1. Nominal crystal cycles. 2. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. 32.6.3 Digital Frequency Locked Loop (DFLL) Characteristics Table 32-14. Digital Frequency Locked Loop Characteristics Symbol Parameter Conditions Output frequency fOUT (2) Reference frequency fREF (2) FINE resolution FINE > 100, all COARSE values tSTARTUP tLOCK Notes: Startup time Lock time (2) Max Unit 40 150 MHz 8 150 kHz 0.25 % Fine lock, fREF = 32kHz, SSG disabled 0.1 0.5 Accurate lock, fREF = 32kHz, dither clk RCSYS/2, SSG disabled 0.06 0.5 Fine lock, fREF = 8-150kHz, SSG disabled 0.2 1 Accurate lock, fREF = 8-150kHz, dither clk RCSYS/2, SSG disabled 0.1 1 Power consumption IDFLL Typ See Figure 324 Frequency drift over voltage and temperature Accuracy(2) Min % 22 Within 90% of final values µA/MHz 100 fREF = 32kHz, fine lock, SSG disabled 600 fREF = 32kHz, accurate lock, dithering clock = RCSYS/2, SSG disabled 1100 µs 1. Spread Spectrum Generator (SSG) is disabled by writing a zero to the EN bit in the SCIF.DFLL0SSG register. 2. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. 781 32099I–01/2012 AT32UC3L016/32/64 Figure 32-4. DFLL Open Loop Frequency Variation(1) DFLL Open Loop Frequency variation 160 150 Frequencies (MHz) 140 130 1,98V 120 1,8V 1.62V 110 100 90 80 -40 -20 0 20 40 60 80 Tem pera ture Note: 1. The plot shows a typical behaviour for coarse = 99 and fine = 255 in open loop mode. 32.6.4 120MHz RC Oscillator (RC120M) Characteristics Table 32-15. Internal 120MHz RC Oscillator Characteristics Symbol Parameter Conditions (1) fOUT Output frequency IRC120M Current consumption tSTARTUP Startup time Note: VVDDCORE = 1.8V Min Typ Max Unit 88 120 152 MHz 1.85 mA 3 µs 1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. 782 32099I–01/2012 AT32UC3L016/32/64 32.6.5 32kHz RC Oscillator (RC32K) Characteristics Table 32-16. 32kHz RC Oscillator Characteristics Symbol Parameter Conditions (1) Min Typ Max Unit 20 32 44 kHz fOUT Output frequency IRC32K Current consumption 0.6 µA tSTARTUP Startup time 100 µs Note: 1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. 32.6.6 System RC Oscillator (RCSYS) Characteristics Table 32-17. System RC Oscillator Characteristics Symbol Parameter Conditions fOUT Output frequency Calibrated at 85°C 32.7 Min Typ Max Unit 111.6 115 118.4 kHz Flash Characteristics Table 32-18 gives the device maximum operating frequency depending on the number of flash wait states and the flash read mode. The FSW bit in the FLASHCDW FSR register controls the number of wait states used when accessing the flash memory. Table 32-18. Maximum Operating Frequency Flash Wait States Read Mode Maximum Operating Frequency 1 50MHz High speed read mode 0 25MHz 1 30MHz Normal read mode 0 15MHz Table 32-19. Flash Characteristics Symbol Parameter tFPP Page programming time tFPE Page erase time tFFP Fuse programming time tFEA Full chip erase time (EA) tFCE JTAG chip erase time (CHIP_ERASE) Conditions Min Typ Max Unit 5 5 fCLK_HSB = 50MHz 1 ms 5 fCLK_HSB = 115kHz 170 783 32099I–01/2012 AT32UC3L016/32/64 Table 32-20. Flash Endurance and Data Retention Symbol Parameter NFARRAY Array endurance (write/page) 100k NFFUSE General Purpose fuses endurance (write/bit) 10k tRET Data retention 15 32.8 Conditions Min Typ Max Unit cycles years Analog Characteristics 32.8.1 Voltage Regulator Characteristics Table 32-21. VREG Electrical Characteristics Symbol Parameter VVDDIN Input voltage range VVDDCORE Output voltage, calibrated value Min Typ Max 1.98 3.3 3.6 Units V Output voltage accuracy IOUT DC output current(1) IVREG Static current of internal regulator Note: Condition VVDDIN >= 1.98V 1.8 IOUT = 0.1mA to 60mA, VVDDIN > 2.2V 2 IOUT = 0.1mA to 60mA, VVDDIN = 1.98V to 2.2V 4 % Normal mode 60 Low power mode 1 mA Normal mode 20 Low power mode 6 µA 1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. Table 32-22. Decoupling Requirements Symbol Parameter CIN1 Input regulator capacitor 1 33 CIN2 Input regulator capacitor 2 100 CIN3 Input regulator capacitor 3 10 µF COUT1 Output regulator capacitor 1 100 nF COUT2 Output regulator capacitor 2 2.2 Note: Condition Typ Techno. Units nF Tantalum 0.5 3.0V, fADC = 6MHz, 8-bit resolution mode, low impedance source 460 cycles kSPS VADVREFP Reference voltage range VADVREFP = VVDDANA IADC Current consumption on VVDDANA ADC Clock = 6MHz 300 IADVREFP Current consumption on ADVREFP pin fADC = 6MHz 250 Note: Max 1.62 1.98 V µA These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. 32.8.6.1 Inputs and Sample and Hold Aquisition Time Table 32-29. Analog Inputs Symbol Parameter VADn Input Voltage Range CONCHIP Internal Capacitance(1) RONCHIP Note: Conditions Internal Resistance (1) 10-bit mode 8-bit mode Min 0 Typ Max Units VADVREFP V 21.5 pF VVDDIO = 3.0V to 3.6V, VVDDCORE = 1.8V 2.55 VVDDIO = VVDDCORE = 1.62V to 1.98V 55.3 kOhm 1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. An analog voltage input must be able to charge the sample and hold (S/H) capacitor in the ADC in order to achieve maximum accuracy. Seen externally the ADC input consists of a resistor ( R ONCHIP ) and a capacitor ( CONCHIP ). In addition the resistance ( R SOURCE ) and capacitance ( C SOURCE ) of the PCB and source must be taken into account when calculating the sample and hold time. Figure 32-7 shows the ADC input channel equivalent circuit. 788 32099I–01/2012 AT32UC3L016/32/64 Figure 32-7. ADC Input RSOURCE Positive Input RONCHIP CSOURCE VIN CONCHIP ADCVREFP/2 The minimum sample and hold time (in ns) can be found using this formula: t SAMPLEHOLD ≥ ( R ONCHIP + R OFFCHIP ) × ( C ONCHIP + C OFFCHIP ) × ln ( 2 n+1 ) Where n is the number of bits in the conversion. t SAMPLEHOLD is defined by the SHTIM field in the ADCIFB ACR register. Please refer to the ADCIFB chapter for more information. 32.8.6.2 Applicable Conditions and Derating Data Table 32-30. Transfer Characteristics 10-bit Resolution Mode Parameter Conditions Min Resolution Offset error Max 10 Integral non-linearity Differential non-linearity Typ Units Bit +/-2 ADC clock frequency = 6MHz -0.9 1 +/-4 Gain error LSB +/-4 Table 32-31. Transfer Characteristics 8-bit Resolution Mode Parameter Conditions Min Resolution Offset error Gain error Max 8 Integral non-linearity Differential non-linearity Typ Units Bit +/-0.5 ADC clock frequency = 6MHz -0.23 0.25 +/-1 LSB +/-1 789 32099I–01/2012 AT32UC3L016/32/64 32.8.7 Temperature Sensor Characteristics Table 32-32. Temperature Sensor Characteristics(1) Symbol Parameter Condition Min Typ Max Units Gradient 1 mV/°C ITS Current consumption 0.5 µA tSTARTUP Startup time 0 µs Note: 1. The Temperature Sensor is not calibrated. The accuracy of the Temperature Sensor is governed by the ADC accuracy. 32.8.8 Analog Comparator Characteristics Table 32-33. Analog Comparator Characteristics Symbol Parameter Condition Min Typ Max Positive input voltage range -0.2 VVDDIO + 0.3 Negative input voltage range -0.2 VVDDIO - 0.6 Units V Statistical offset VACREFN = 1.0V, fAC = 12MHz, filter length = 2, hysteresis = 0(1) Clock frequency for GCLK4 fAC Throughput rate(3) fAC = 12MHz Propagation delay Delay from input change to Interrupt Status Register Changes IAC Current consumption All channels, VDDIO = 3.3V, fA = 3MHz tSTARTUP Startup time Input current per pin Notes: 20 1 ⎛ + 3⎞⎠ × t CLKACIFB ⎝ t---------------------------------------CLKACIFB × f AC mV 12 MHz 12 000 000 Comparisons per second ns 420 µA 3 cycles 0.2 µA/MHz(2) 1. AC.CONFn.FLEN and AC.CONFn.HYS fields, refer to the Analog Comparator Interface chapter. 2. Referring to fAC. 3. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. 790 32099I–01/2012 AT32UC3L016/32/64 32.8.9 Capacitive Touch Characteristics 32.8.9.1 Discharge Current Source Table 32-34. DICS Characteristics Symbol Parameter RREF Internal resistor 120 kOhm k Trim step size 0.7 % 32.8.9.2 Min Typ Max Unit Strong Pull-up Pull-down Table 32-35. Strong Pull-up Pull-down Parameter Min Typ Pull-down resistor 1 Pull-up resistor 1 Max Unit kOhm 791 32099I–01/2012 AT32UC3L016/32/64 32.9 Timing Characteristics 32.9.1 Startup, Reset, and Wake-up Timing The startup, reset, and wake-up timings are calculated using the following formula: t = t CONST + N CPU × t CPU Where t CONST and N CPU are found in Table 32-36. t CPU is the period of the CPU clock. If another clock source than RCSYS is selected as CPU clock the startup time of the oscillator, t OSCSTART , must added to the wake-up time in the stop, deepstop, and static sleep modes. Please refer to the source for the CPU clock in the ”Oscillator Characteristics” on page 779 for more details about oscillator startup times. Table 32-36. Maximum Reset and Wake-up Timing(1) Max t CONST (in µs) Max N CPU Parameter Measuring Startup time from power-up, using regulator Time from VDDIN crossing the VPOT+ threshold of POR33 to the first instruction entering the decode stage of CPU. VDDCORE is supplied by the internal regulator. 2210 0 Startup time from power-up, no regulator Time from VDDIN crossing the VPOT+ threshold of POR33 to the first instruction entering the decode stage of CPU. VDDCORE is connected to VDDIN. 1810 0 Startup time from reset release Time from releasing a reset source (except POR18, POR33, and SM33) to the first instruction entering the decode stage of CPU. 170 0 Idle 0 19 Frozen 0 110 0 110 27 + t OSCSTART 116 Deepstop 27 + t OSCSTART 116 Static 97 + t OSCSTART 116 1180 0 Standby Wake-up Stop Wake-up from shutdown Note: From wake-up event to the first instruction of an interrupt routine entering the decode stage of the CPU. From wake-up event to the first instruction entering the decode stage of the CPU. 1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. 32.9.2 RESET_N Timing Table 32-37. RESET_N Waveform Parameters(1) Symbol Parameter tRESET RESET_N minimum pulse length Note: Conditions Min 10 Max Units ns 1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. 792 32099I–01/2012 AT32UC3L016/32/64 32.9.3 USART in SPI Mode Timing 32.9.3.1 Master mode Figure 32-8. USART in SPI Master Mode With (CPOL= CPHA= 0) or (CPOL= CPHA= 1) SPCK MISO USPI0 USPI1 MOSI USPI2 Figure 32-9. USART in SPI Master Mode With (CPOL= 0 and CPHA= 1) or (CPOL= 1 and CPHA= 0) SPCK MISO USPI3 USPI4 MOSI USPI5 Table 32-38. USART in SPI Mode Timing, Master Mode(1) Symbol Parameter Conditions USPI0 MISO setup time before SPCK rises USPI1 MISO hold time after SPCK rises USPI2 SPCK rising to MOSI delay USPI3 MISO setup time before SPCK falls USPI4 MISO hold time after SPCK falls USPI5 SPCK falling to MOSI delay Notes: Min 30.0+ VVDDIO from 3.0V to 3.6V, maximum external capacitor = 40pF Max Units tSAMPLE(2) 0 8.5 ns 25.5 + tSAMPLE(2) 0 13.6 1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. t SPCK 1⎞ 2. Where: t SAMPLE = t SPCK – ⎛ -------------------------------------- × t CLKUSART ⎝ 2×t 2⎠ CLKUSART 793 32099I–01/2012 AT32UC3L016/32/64 Maximum SPI Frequency, Master Output The maximum SPI master output frequency is given by the following formula: 1 f CLKSPI × 2 f SPCKMAX = MIN (f PINMAX,------------, -----------------------------) SPIn 9 Where SPIn is the MOSI delay, USPI2 or USPI5 depending on CPOL and NCPHA. f PINMAX is the maximum frequency of the SPI pins. Please refer to the I/O Pin Characteristics section for the maximum frequency of the pins. f CLKSPI is the maximum frequency of the CLK_SPI. Refer to the SPI chapter for a description of this clock. Maximum SPI Frequency, Master Input The maximum SPI master input frequency is given by the following formula: f CLKSPI × 2 1 f SPCKMAX = MIN (------------------------------------,-----------------------------) SPIn + t VALID 9 Where SPIn is the MISO setup and hold time, USPI0 + USPI1 or USPI3 + USPI4 depending on CPOL and NCPHA. T VALID is the SPI slave response time. Please refer to the SPI slave datasheet for T VALID . f CLKSPI is the maximum frequency of the CLK_SPI. Refer to the SPI chapter for a description of this clock. 32.9.3.2 Slave mode Figure 32-10. USART in SPI Slave Mode With (CPOL= 0 and CPHA= 1) or (CPOL= 1 and CPHA= 0) SPCK MISO USPI6 MOSI USPI7 USPI8 Figure 32-11. USART in SPI Slave Mode With (CPOL= CPHA= 0) or (CPOL= CPHA= 1) SPCK MISO USPI9 MOSI USPI10 USPI11 794 32099I–01/2012 AT32UC3L016/32/64 Figure 32-12. USART in SPI Slave Mode NPCS Timing USPI12 USPI13 USPI14 USPI15 SPCK, CPOL=0 SPCK, CPOL=1 NSS Table 32-39. USART in SPI mode Timing, Slave Mode(1) Symbol Parameter Conditions USPI6 SPCK falling to MISO delay Max Units 27.6 USPI7 MOSI setup time before SPCK rises USPI8 MOSI hold time after SPCK rises USPI9 SPCK rising to MISO delay USPI10 MOSI setup time before SPCK falls USPI11 MOSI hold time after SPCK falls USPI12 NSS setup time before SPCK rises USPI13 NSS hold time after SPCK falls USPI14 NSS setup time before SPCK falls USPI15 NSS hold time after SPCK rises Notes: Min tSAMPLE(2) + tCLK_USART 0 VVDDIO from 3.0V to 3.6V, maximum external capacitor = 40pF 27.2 tSAMPLE(2) + tCLK_USART ns 0 25.0 0 25.0 0 1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. t SPCK 1 2. Where: t SAMPLE = t SPCK – ⎛ -----------------------------------+ ---⎞ × t CLKUSART ⎝ 2×t 2⎠ CLKUSART Maximum SPI Frequency, Slave Input Mode The maximum SPI slave input frequency is given by the following formula: f CLKSPI × 2 1 f SPCKMAX = MIN (-----------------------------,------------) 9 SPIn Where SPIn is the MOSI setup and hold time, USPI7 + USPI8 or USPI10 + USPI11 depending on CPOL and NCPHA. f CLKSPI is the maximum frequency of the CLK_SPI. Refer to the SPI chapter for a description of this clock. Maximum SPI Frequency, Slave Output Mode 795 32099I–01/2012 AT32UC3L016/32/64 The maximum SPI slave output frequency is given by the following formula: f CLKSPI × 2 1 f SPCKMAX = MIN (-----------------------------, f PINMAX,------------------------------------) 9 SPIn + t SETUP Where SPIn is the MISO delay, USPI6 or USPI9 depending on CPOL and NCPHA. T SETUP is the SPI master setup time. Please refer to the SPI master datasheet for T SETUP . f CLKSPI is the maximum frequency of the CLK_SPI. Refer to the SPI chapter for a description of this clock. f PINMAX is the maximum frequency of the SPI pins. Please refer to the I/O Pin Characteristics section for the maximum frequency of the pins. 32.9.4 SPI Timing 32.9.4.1 Master mode Figure 32-13. SPI Master Mode With (CPOL= NCPHA= 0) or (CPOL= NCPHA= 1) SPCK MISO SPI0 SPI1 MOSI SPI2 Figure 32-14. SPI Master Mode With (CPOL= 0 and NCPHA= 1) or (CPOL= 1 and NCPHA= 0) SPCK MISO SPI3 SPI4 MOSI SPI5 796 32099I–01/2012 AT32UC3L016/32/64 Table 32-40. SPI Timing, Master Mode(1) Symbol Parameter SPI0 MISO setup time before SPCK rises SPI1 MISO hold time after SPCK rises SPI2 SPCK rising to MOSI delay SPI3 MISO setup time before SPCK falls SPI4 MISO hold time after SPCK falls SPI5 SPCK falling to MOSI delay Note: Conditions Min Max Units 28.4 + (tCLK_SPI)/2 VVDDIO from 3.0V to 3.6V, maximum external capacitor = 40pF 0 7.1 ns 22.8 + (tCLK_SPI)/2 0 11.0 1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. Maximum SPI Frequency, Master Output The maximum SPI master output frequency is given by the following formula: 1 f SPCKMAX = MIN (f PINMAX,------------) SPIn Where SPIn is the MOSI delay, SPI2 or SPI5 depending on CPOL and NCPHA. f PINMAX is the maximum frequency of the SPI pins. Please refer to the I/O Pin Characteristics section for the maximum frequency of the pins. Maximum SPI Frequency, Master Input The maximum SPI master input frequency is given by the following formula: 1 f SPCKMAX = -----------------------------------SPIn + t VALID Where SPIn is the MISO setup and hold time, SPI0 + SPI1 or SPI3 + SPI4 depending on CPOL and NCPHA. t VALID is the SPI slave response time. Please refer to the SPI slave datasheet for t VALID . 32.9.4.2 Slave mode Figure 32-15. SPI Slave Mode With (CPOL= 0 and NCPHA= 1) or (CPOL= 1 and NCPHA= 0) SPCK MISO SPI6 MOSI SPI7 SPI8 797 32099I–01/2012 AT32UC3L016/32/64 Figure 32-16. SPI Slave Mode With (CPOL= NCPHA= 0) or (CPOL= NCPHA= 1) SPCK MISO SPI9 MOSI SPI10 SPI11 Figure 32-17. SPI Slave Mode NPCS Timing SPI12 SPI13 SPI14 SPI15 SPCK, CPOL=0 SPCK, CPOL=1 NPCS Table 32-41. SPI Timing, Slave Mode(1) Symbol Parameter SPI6 SPCK falling to MISO delay SPI7 MOSI setup time before SPCK rises SPI8 MOSI hold time after SPCK rises SPI9 SPCK rising to MISO delay SPI10 MOSI setup time before SPCK falls SPI11 MOSI hold time after SPCK falls SPI12 NPCS setup time before SPCK rises SPI13 NPCS hold time after SPCK falls 0.2 SPI14 NPCS setup time before SPCK falls 2.2 SPI15 NPCS hold time after SPCK rises Note: Conditions Min Max Units 30.8 0 4.1 VVDDIO from 3.0V to 3.6V, maximum external capacitor = 40pF 29.9 0 ns 3.5 1.9 0 1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. Maximum SPI Frequency, Slave Input Mode 798 32099I–01/2012 AT32UC3L016/32/64 The maximum SPI slave input frequency is given by the following formula: 1 f SPCKMAX = MIN (f CLKSPI,------------) SPIn Where SPIn is the MOSI setup and hold time, SPI7 + SPI8 or SPI10 + SPI11 depending on CPOL and NCPHA. f CLKSPI is the maximum frequency of the CLK_SPI. Refer to the SPI chapter for a description of this clock. Maximum SPI Frequency, Slave Output Mode The maximum SPI slave output frequency is given by the following formula: 1 f SPCKMAX = MIN (f PINMAX,------------------------------------) SPIn + t SETUP Where SPIn is the MISO delay, SPI6 or SPI9 depending on CPOL and NCPHA. t SETUP is the SPI master setup time. Please refer to the SPI master datasheet for t SETUP . f PINMAX is the maximum frequency of the SPI pins. Please refer to the I/O Pin Characteristics section for the maximum frequency of the pins. 32.9.5 TWIM/TWIS Timing Figure 32-42 shows the TWI-bus timing requirements and the compliance of the device with them. Some of these requirements (tr and tf) are met by the device without requiring user intervention. Compliance with the other requirements (tHD-STA, tSU-STA, tSU-STO, tHD-DAT, tSU-DAT-TWI, tLOWTWI, tHIGH, and fTWCK) requires user intervention through appropriate programming of the relevant TWIM and TWIS user interface registers. Please refer to the TWIM and TWIS sections for more information. Table 32-42. TWI-Bus Timing Requirements Minimum Symbol Parameter Mode Requirement Standard(1) tr TWCK and TWD rise time tf TWCK and TWD fall time tHD-STA (Repeated) START hold time tSU-STA (Repeated) START set-up time tSU-STO STOP set-up time tHD-DAT Data hold time Maximum Device Requirement Device - 1000 20 + 0.1Cb 300 - 300 20 + 0.1Cb 300 Unit ns Fast(1) Standard ns Fast Standard 4 Fast 0.6 Standard 4.7 Fast 0.6 Standard 4.0 Fast 0.6 Standard Fast 0.3(2) tclkpb - μs tclkpb - μs 4tclkpb - μs 3.45() 2tclkpb 0.9() 15tprescaled + tclkpb μs 799 32099I–01/2012 AT32UC3L016/32/64 Table 32-42. TWI-Bus Timing Requirements Minimum Symbol Parameter tSU-DAT-TWI Data set-up time tSU-DAT tLOW-TWI Mode Requirement Standard 250 Fast 100 - Device - Standard 4.7 Fast 1.3 TWCK LOW period tLOW - tHIGH TWCK HIGH period fTWCK TWCK frequency - Standard 4.0 Fast 0.6 Standard Notes: Maximum Requirement Unit 2tclkpb - ns tclkpb - - 4tclkpb - μs tclkpb - - 8tclkpb - μs 100 - Fast Device 400 1 -----------------------12t clkpb kHz 1. Standard mode: f TWCK ≤ 100 kHz ; fast mode: f TWCK > 100 kHz . 2. A device must internally provide a hold time of at least 300 ns for TWD with reference to the falling edge of TWCK. Notations: Cb = total capacitance of one bus line in pF tclkpb = period of TWI peripheral bus clock tprescaled = period of TWI internal prescaled clock (see chapters on TWIM and TWIS) The maximum tHD;DAT has only to be met if the device does not stretch the LOW period (tLOW-TWI) of TWCK. 800 32099I–01/2012 AT32UC3L016/32/64 32.9.6 JTAG Timing Figure 32-18. JTAG Interface Signals JTAG2 TCK JTAG0 JTAG1 TMS/TDI JTAG3 JTAG4 JTAG7 JTAG8 TDO JTAG5 JTAG6 Boundary Scan Inputs Boundary Scan Outputs JTAG9 JTAG10 Table 32-43. JTAG Timings(1) Symbol Parameter JTAG0 TCK Low Half-period 23.2 JTAG1 TCK High Half-period 8.8 JTAG2 TCK Period 32.0 JTAG3 TDI, TMS Setup before TCK High JTAG4 TDI, TMS Hold after TCK High JTAG5 TDO Hold Time JTAG6 TCK Low to TDO Valid JTAG7 Boundary Scan Inputs Setup Time JTAG8 Boundary Scan Inputs Hold Time 5.0 JTAG9 Boundary Scan Outputs Hold Time 8.7 JTAG10 TCK to Boundary Scan Outputs Valid Note: Conditions VVDDIO from 3.0V to 3.6V, maximum external capacitor = 40pF Min Max Units 3.9 0.6 ns 4.5 23.2 0 17.7 1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production. 801 32099I–01/2012 AT32UC3L016/32/64 33. Mechanical Characteristics 33.1 33.1.1 Thermal Considerations Thermal Data Table 33-1 summarizes the thermal resistance data depending on the package. Table 33-1. 33.1.2 Thermal Resistance Data Symbol Parameter Condition Package Typ θJA Junction-to-ambient thermal resistance Still Air TQFP48 63.2 θJC Junction-to-case thermal resistance TQFP48 21.8 θJA Junction-to-ambient thermal resistance QFN48 28.3 θJC Junction-to-case thermal resistance QFN48 2.5 θJA Junction-to-ambient thermal resistance TLLGA48 25.4 θJC Junction-to-case thermal resistance TLLGA48 12.7 Still Air Still Air Unit °C/W °C/W °C/W Junction Temperature The average chip-junction temperature, TJ, in °C can be obtained from the following: 1. T J = T A + ( P D × θ JA ) 2. T J = T A + ( P D × ( θ HEATSINK + θ JC ) ) where: • θJA = package thermal resistance, Junction-to-ambient (°C/W), provided in Table 33-1. • θJC = package thermal resistance, Junction-to-case thermal resistance (°C/W), provided in Table 33-1. • θHEAT SINK = cooling device thermal resistance (°C/W), provided in the device datasheet. • PD = device power consumption (W) estimated from data provided in the Section 32.4 on page 771. • TA = ambient temperature (°C). From the first equation, the user can derive the estimated lifetime of the chip and decide if a cooling device is necessary or not. If a cooling device is to be fitted on the chip, the second equation should be used to compute the resulting average chip-junction temperature TJ in °C. 802 32099I–01/2012 AT32UC3L016/32/64 33.2 Package Drawings Figure 33-1. TQFP-48 Package Drawing Table 33-2. Device and Package Maximum Weight 140 Table 33-3. mg Package Characteristics Moisture Sensitivity Level Table 33-4. MSL3 Package Reference JEDEC Drawing Reference MS-026 JESD97 Classification E3 803 32099I–01/2012 AT32UC3L016/32/64 Figure 33-2. QFN-48 Package Drawing Note: The exposed pad is not connected to anything internally, but should be soldered to ground to increase board level reliability. Table 33-5. Device and Package Maximum Weight 140 Table 33-6. mg Package Characteristics Moisture Sensitivity Level Table 33-7. MSL3 Package Reference JEDEC Drawing Reference M0-220 JESD97 Classification E3 804 32099I–01/2012 AT32UC3L016/32/64 Figure 33-3. TLLGA-48 Package Drawing Table 33-8. Device and Package Maximum Weight 39.3 Table 33-9. mg Package Characteristics Moisture Sensitivity Level MSL3 Table 33-10. Package Reference JEDEC Drawing Reference N/A JESD97 Classification E4 805 32099I–01/2012 AT32UC3L016/32/64 33.3 Soldering Profile Table 33-11 gives the recommended soldering profile from J-STD-20. Table 33-11. Soldering Profile Profile Feature Green Package Average Ramp-up Rate (217°C to Peak) 3°C/s max Preheat Temperature 175°C ±25°C 150-200°C Time Maintained Above 217°C 60-120s Time within 5°C of Actual Peak Temperature 30s Peak Temperature Range 260°C Ramp-down Rate 6°C/s max Time 25°C to Peak Temperature 8 minutes max A maximum of three reflow passes is allowed per component. 806 32099I–01/2012 AT32UC3L016/32/64 34. Ordering Information Table 34-1. Device Ordering Information Ordering Code Carrier Type AT32UC3L064-AUTES ES AT32UC3L064-AUT Tray AT32UC3L064-AUR Tape & Reel Package Package Type Temperature Operating Range TQFP 48 JESD97 Classification E3 AT32UC3L064 AT32UC3L064-ZAUES ES AT32UC3L064-ZAUT Tray AT32UC3L064-ZAUR Tape & Reel AT32UC3L064-D3HES ES AT32UC3L064-D3HT Tray AT32UC3L064-D3HR Tape & Reel AT32UC3L032-AUT Tray AT32UC3L032-AUR Tape & Reel QFN 48 TLLGA 48 JESD97 Classification E4 TQFP 48 Industrial (-40°C to 85°C) JESD97 Classification E3 AT32UC3L032-ZAUT Tray AT32UC3L032-ZAUR Tape & Reel AT32UC3L032-D3HT Tray AT32UC3L032-D3HR Tape & Reel AT32UC3L032 QFN 48 TLLGA 48 AT32UC3L016-AUT Tray AT32UC3L016-AUR Tape & Reel AT32UC3L016-ZAUT Tray AT32UC3L016-ZAUR Tape & Reel AT32UC3L016-D3HT Tray AT32UC3L016-D3HR Tape & Reel JESD97 Classification E4 TQFP 48 JESD97 Classification E3 AT32UC3L016 QFN 48 TLLGA 48 JESD97 Classification E4 807 32099I–01/2012 AT32UC3L016/32/64 35. Errata 35.1 35.1.1 Rev. E Processor and Architecture 1. Hardware breakpoints may corrupt MAC results Hardware breakpoints on MAC instructions may corrupt the destination register of the MAC instruction. Fix/Workaround Place breakpoints on earlier or later instructions. 2. Privilege violation when using interrupts in application mode with protected system stack If the system stack is protected by the MPU and an interrupt occurs in application mode, an MPU DTLB exception will occur. Fix/Workaround Make a DTLB Protection (Write) exception handler which permits the interrupt request to be handled in privileged mode. 35.1.2 FLASHCDW 1. Flash self programming may fail in one wait state mode Writes in flash and user pages may fail if executing code is located in address space mapped to flash, and the flash controller is configured in one wait state mode (the Flash Wait State bit in the Flash Control Register (FCR.FWS) is one). Fix/Workaround Solution 1: Configure the flash controller in zero wait state mode (FCR.FWS=0). Solution 2: Configure the HMATRIX master 1 (CPU Instruction) to use the unlimited burst length transfer mode (MCFG1.ULBT=0), and the HMATRIX slave 0 (FLASHCDW) to use the maximum slot cycle limit (SCFG0.SLOT_CYCLE=255). 35.1.3 Power Manager 1. Clock sources will not be stopped in Static mode if the difference between CPU and PBx division factor is larger than 4 If the division factor between the CPU/HSB and PBx frequencies is more than 4 when entering a sleep mode where the system RC oscillator (RCSYS) is turned off, the high speed clock sources will not be turned off. This will result in a significantly higher power consumption during the sleep mode. Fix/Workaround Before going to sleep modes where RCSYS is stopped, make sure the division factor between CPU/HSB and PBx frequencies is less than or equal to 4. 2. Clock Failure Detector (CFD) can be issued while turning off the CFD While turning off the CFD, the CFD bit in the Status Register (SR) can be set. This will change the main clock source to RCSYS. Fix/Workaround Solution 1: Enable CFD interrupt. If CFD interrupt is issues after turning off the CFD, switch back to original main clock source. Solution 2: Only turn off the CFD while running the main clock on RCSYS. 3. Sleepwalking in idle and frozen sleep mode will mask all other PB clocks 808 32099I–01/2012 AT32UC3L016/32/64 If the CPU is in idle or frozen sleep mode and a module is in a state that triggers sleep walking, all PB clocks will be masked except the PB clock to the sleepwalking module. Fix/Workaround Mask all clock requests in the PM.PPCR register before going into idle or frozen mode. 35.1.4 SCIF 1. PCLKSR.OSC32RDY bit might not be cleared after disabling OSC32K In some cases the OSC32RDY bit in the PCLKSR register will not be cleared when OSC32K is disabled. Fix/Workaround When re-enabling the OSC32K, read the PCLKSR.OSC32RDY bit. If this bit is: 0: Follow normal procedures. 1: Ignore the PCLKSR.OSC32RDY and ISR.OSC32RDY bit. Use the Frequency Meter (FREQM) to determine if the OSC32K clock is ready. The OSC32K clock is ready when the FREQM measures a non-zero frequency. 2. The RC32K output on PA20 is not always permanently disabled The RC32K output on PA20 may sometimes re-appear. Fix/Workaround Before using RC32K for other purposes, the following procedure has to be followed in order to properly disable it: - Run the CPU on RCSYS - Disable the output to PA20 by writing a zero to PM.PPCR.RC32OUT - Enable RC32K by writing a one to SCIF.RC32KCR.EN, and wait for this bit to be read as one - Disable RC32K by writing a zero to SCIF.RC32KCR.EN, and wait for this bit to be read as zero. 35.1.5 AST 1. Reset may set status bits in the AST If a reset occurs and the AST is enabled, the SR.ALARM0, SR.PER0, and SR.OVF bits may be set. Fix/Workaround If the part is reset and the AST is used, clear all bits in the Status Register before entering sleep mode. 2. AST wake signal is released one AST clock cycle after the BUSY bit is cleared After writing to the Status Clear Register (SCR) the wake signal is released one AST clock cycle after the BUSY bit in the Status Register (SR.BUSY) is cleared. If entering sleep mode directly after the BUSY bit is cleared the part will wake up immediately. Fix/Workaround Read the Wake Enable Register (WER) and write this value back to the same register. Wait for BUSY to clear before entering sleep mode. 35.1.6 WDT 1. Clearing the Watchdog Timer (WDT) counter in second half of timeout period will issue a Watchdog reset If the WDT counter is cleared in the second half of the timeout period, the WDT will immediately issue a Watchdog reset. Fix/Workaround 809 32099I–01/2012 AT32UC3L016/32/64 Use twice as long timeout period as needed and clear the WDT counter within the first half of the timeout period. If the WDT counter is cleared after the first half of the timeout period, you will get a Watchdog reset immediately. If the WDT counter is not cleared at all, the time before the reset will be twice as long as needed. 2. WDT Control Register does not have synchronization feedback When writing to the Timeout Prescale Select (PSEL), Time Ban Prescale Select (TBAN), Enable (EN), or WDT Mode (MODE) fieldss of the WDT Control Register (CTRL), a synchronizer is started to propagate the values to the WDT clcok domain. This synchronization takes a finite amount of time, but only the status of the synchronization of the EN bit is reflected back to the user. Writing to the synchronized fields during synchronization can lead to undefined behavior. Fix/Workaround -When writing to the affected fields, the user must ensure a wait corresponding to 2 clock cycles of both the WDT peripheral bus clock and the selected WDT clock source. -When doing writes that changes the EN bit, the EN bit can be read back until it reflects the written value. 35.1.7 GPIO 1. Clearing GPIO interrupt may fail Writing a one to the GPIO.IFRC register to clear an interrupt will be ignored if interrupt is enabled for the corresponding port. Fix/Workaround Disable the interrupt, clear it by writing a one to GPIO.IFRC, then enable the interrupt. 35.1.8 SPI 1. SPI data transfer hangs with CSR0.CSAAT==1 and MR.MODFDIS==0 When CSR0.CSAAT==1 and mode fault detection is enabled (MR.MODFDIS==0), the SPI module will not start a data transfer. Fix/Workaround Disable mode fault detection by writing a one to MR.MODFDIS. 2. Disabling SPI has no effect on the SR.TDRE bit Disabling SPI has no effect on the SR.TDRE bit whereas the write data command is filtered when SPI is disabled. Writing to TDR when SPI is disabled will not clear SR.TDRE. If SPI is disabled during a PDCA transfer, the PDCA will continue to write data to TDR until its buffer is empty, and this data will be lost. Fix/Workaround Disable the PDCA, add two NOPs, and disable the SPI. To continue the transfer, enable the SPI and PDCA. 3. SPI disable does not work in SLAVE mode SPI disable does not work in SLAVE mode. Fix/Workaround Read the last received data, then perform a software reset by writing a one to the Software Reset bit in the Control Register (CR.SWRST). 4. SPI bad serial clock generation on 2nd chip_select when SCBR=1, CPOL=1, and NCPHA=0 When multiple chip selects (CS) are in use, if one of the baudrates equal 1 while one (CSRn.SCBR=1) of the others do not equal 1, and CSRn.CPOL=1 and CSRn.NCPHA=0, then an additional pulse will be generated on SCK. 810 32099I–01/2012 AT32UC3L016/32/64 Fix/Workaround When multiple CS are in use, if one of the baudrates equals 1, the others must also equal 1 if CSRn.CPOL=1 and CSRn.NCPHA=0. 5. SPI mode fault detection enable causes incorrect behavior When mode fault detection is enabled (MR.MODFDIS==0), the SPI module may not operate properly. Fix/Workaround Always disable mode fault detection before using the SPI by writing a one to MR.MODFDIS. 35.1.9 TWI 1. TWIM SR.IDLE goes high immediately when NAK is received When a NAK is received and there is a non-zero number of bytes to be transmitted, SR.IDLE goes high immediately and does not wait for the STOP condition to be sent. This does not cause any problem just by itself, but can cause a problem if software waits for SR.IDLE to go high and then immediately disables the TWIM by writing a one to CR.MDIS. Disabling the TWIM causes the TWCK and TWD pins to go high immediately, so the STOP condition will not be transmitted correctly. Fix/Workaround If possible, do not disable the TWIM. If it is absolutely necessary to disable the TWIM, there must be a software delay of at least two TWCK periods between the detection of SR.IDLE==1 and the disabling of the TWIM. 2. TWIM TWALM polarity is wrong The TWALM signal in the TWIM is active high instead of active low. Fix/Workaround Use an external inverter to invert the signal going into the TWIM. When using both TWIM and TWIS on the same pins, the TWALM cannot be used. 3. TWIS may not wake the device from sleep mode If the CPU is put to a sleep mode (except Idle and Frozen) directly after a TWI Start condition, the CPU may not wake upon a TWIS address match. The request is NACKed. Fix/Workaround When using the TWI address match to wake the device from sleep, do not switch to sleep modes deeper than Frozen. Another solution is to enable asynchronous EIC wake on the TWIS clock (TWCK) or TWIS data (TWD) pins, in order to wake the system up on bus events. 4. SMBALERT bit may be set after reset The SMBus Alert (SMBALERT) bit in the Status Register (SR) might be erroneously set after system reset. Fix/Workaround After system reset, clear the SR.SMBALERT bit before commencing any TWI transfer. 5. Clearing the NAK bit before the BTF bit is set locks up the TWI bus When the TWIS is in transmit mode, clearing the NAK Received (NAK) bit of the Status Register (SR) before the end of the Acknowledge/Not Acknowledge cycle will cause the TWIS to attempt to continue transmitting data, thus locking up the bus. Fix/Workaround Clear SR.NAK only after the Byte Transfer Finished (BTF) bit of the same register has been set. 811 32099I–01/2012 AT32UC3L016/32/64 6. TWIS stretch on Address match error When the TWIS stretches TWCK due to a slave address match, it also holds TWD low for the same duration if it is to be receiving data. When TWIS releases TWCK, it releases TWD at the same time. This can cause a TWI timing violation. Fix/Workaround None. 35.1.10 PWMA 1. BUSY bit is never cleared after writes to the Control Register (CR) When writing a non-zero value to CR.TOP, CR.SPREAD, or CR.TCLR when the PWMA is disabled (CR.EN == 0), the BUSY bit in the Status Register (SR.BUSY) will be set, but never cleared. Fix/Workaround When writing a non-zero value to CR.TOP, CR.SPREAD, or CR.TCLR, make sure the PWMA is enabled, or simultaneously enable the PWMA by writing a one to CR.EN. 2. Incoming peripheral events are discarded during duty cycle register update Incoming peripheral events to all applied channels will be discarded if a duty cycle update is received from the user interface in the same PWMA clock period. Fix/Workaround Ensure that duty cycle writes from the user interface are not performed in a PWMA period when an incoming peripheral event is expected. 35.1.11 ADCIFB 1. Using STARTUPTIME larger than 0x1F will freeze the ADC Writing a value larger than 0x1F to the Startup Time field in the ADC Configuration Register (ACR.STARTUP) will freeze the ADC, and the Busy Status bit in the Status Register (SR.BUSY) will never be cleared. Fix/Workaround Do not write values larger than 0x1F to ACR.STARTUP. 35.1.12 CAT 1. CAT asynchronous wake will be delayed by one AST event period If the CAT detects a condition the should asynchronously wake the device in static mode, the asynchronous wake will not occur until the next AST event. For example, if the AST is generating events to the CAT every 50ms, and the CAT detects a touch at t=9200ms, the asynchronous wake will occur at t=9250ms. Fix/Workaround None. 2. CAT QMatrix sense capacitors discharged prematurely At the end of a QMatrix burst charging sequence that uses different burst count values for different Y lines, the Y lines may be incorrectly grounded for up to n-1 periods of the peripheral bus clock, where n is the ratio of the PB clock frequency to the GCLK_CAT frequency. This results in premature loss of charge from the sense capacitors and thus increased variability of the acquired count values. Fix/Workaround Enable the 1kOhm drive resistors on all implemented QMatrix Y lines (CSA 1, 3, 5, 7, 9, 11, 13, and/or 15) by writing ones to the corresponding odd bits of the CSARES register. 3. CAT module does not terminate QTouch burst on detect 812 32099I–01/2012 AT32UC3L016/32/64 The CAT module does not terminate a QTouch burst when the detection voltage is reached on the sense capacitor. This can cause the sense capacitor to be charged more than necessary. Depending on the dielectric absorption characteristics of the capacitor, this can lead to unstable measurements. Fix/Workaround Use the minimum possible value for the MAX field in the ATCFG1, TG0CFG1, and TG1CFG1 registers. 4. Autonomous CAT acquisition must be longer than AST source clock period When using the AST to trigger CAT autonomous touch acquisition in sleep modes where the CAT bus clock is turned off, the CAT will start several acquisitions if the period of the AST source clock is larger than one CAT acquisition. One AST clock period after the AST trigger, the CAT clock will automatically stop and the CAT acquisition can be stopped prematurely, ruining the result. Fix/Workaround Always ensure that the ATCFG1.max field is set so that the duration of the autonomous touch acquisition is greater than one clock period of the AST source clock. 35.1.13 aWire 1. aWire CPU clock speed robustness The aWire memory speed request command counter warps at clock speeds below approximately 5kHz. Fix/Workaround None. 2. The aWire debug interface is reset after leaving Shutdown mode If the aWire debug mode is used as debug interface and the program enters Shutdown mode, the aWire interface will be reset when the part receives a wakeup either from the WAKE_N pin or the AST. Fix/Workaround None. 35.1.14 CHIP 1. Increased Power Consumption in VDDIO in sleep modes If OSC0 is enabled in crystal mode when entering a sleep mode where the OSC0 is disabled, this will lead to an increased power consumption in VDDIO. Fix/Workaround Solution 1: Disable OSC0 by writing a zero to the Oscillator Enable bit in the System Control Interface (SCIF) Oscillator Control Register (SCIF.OSC0CTRL.OSCEN) before going to a sleep mode where OSC0 is disabled. Solution 2: Pull down or up XIN0 or XOUT0 with 1MOhm resistor. 35.1.15 I/O Pins 1. PA17 has low ESD tolerance PA17 only tolerates 500V ESD pulses (Human Body Model). Fix/Workaround Care must be taken during manufacturing and PCB design. 813 32099I–01/2012 AT32UC3L016/32/64 35.2 35.2.1 Rev. D Processor and Architecture 1. Hardware breakpoints may corrupt MAC results Hardware breakpoints on MAC instructions may corrupt the destination register of the MAC instruction. Fix/Workaround Place breakpoints on earlier or later instructions. 2. Privilege violation when using interrupts in application mode with protected system stack If the system stack is protected by the MPU and an interrupt occurs in application mode, an MPU DTLB exception will occur. Fix/Workaround Make a DTLB Protection (Write) exception handler which permits the interrupt request to be handled in privileged mode. 35.2.2 FLASHCDW 1. Flash self programming may fail in one wait state mode Writes in flash and user pages may fail if executing code is located in address space mapped to flash, and the flash controller is configured in one wait state mode (the Flash Wait State bit in the Flash Control Register (FCR.FWS) is one). Fix/Workaround Solution 1: Configure the flash controller in zero wait state mode (FCR.FWS=0). Solution 2: Configure the HMATRIX master 1 (CPU Instruction) to use the unlimited burst length transfer mode (MCFG1.ULBT=0), and the HMATRIX slave 0 (FLASHCDW) to use the maximum slot cycle limit (SCFG0.SLOT_CYCLE=255). 35.2.3 Power Manager 1. Clock sources will not be stopped in Static mode if the difference between CPU and PBx division factor is larger than 4 If the division factor between the CPU/HSB and PBx frequencies is more than 4 when entering a sleep mode where the system RC oscillator (RCSYS) is turned off, the high speed clock sources will not be turned off. This will result in a significantly higher power consumption during the sleep mode. Fix/Workaround Before going to sleep modes where RCSYS is stopped, make sure the division factor between CPU/HSB and PBx frequencies is less than or equal to 4. 2. External reset in Shutdown mode If an external reset is asserted while the device is in Shutdown mode, the Power Manager will register this as a Power-on reset (POR), and not as a SLEEP reset, in the Reset Cause register (RCAUSE) Fix/Workaround None. 3. Disabling POR33 may generate spurious resets Depending on operating conditions, POR33 may generate a spurious reset in one of the following cases: - When POR33 is disabled from the user interface - When SM33 supply monitor is enabled 814 32099I–01/2012 AT32UC3L016/32/64 - When entering Shutdown mode while debugging the chip using JTAG or aWire interface In the listed cases, writing a one to the bit VREGCR.POR33MASK in the System Control Interface (SCIF) to mask the POR33 reset will be ineffective Fix/Workaround - Do not disable POR33 using the user interface - Do not use the SM33 supply monitor - Do not enter Shutdown mode if a debugger is connected to the chip 4. Instability when exiting sleep walking If all the following operating conditions are true, exiting sleep walking might lead to instability: - The OSC0 is enabled in external clock mode (OSCCTRL0.OSCEN == 1 and OSCCTRL0.MODE == 0) - A sleep mode where the OSC0 is automatically disabled is entered - The device enters sleep walking Fix/Workaround Do not run OSC0 in external clock mode if sleepwalking is expected to be used. 5. Clock Failure Detector (CFD) can be issued while turning off the CFD While turning off the CFD, the CFD bit in the Status Register (SR) can be set. This will change the main clock source to RCSYS. Fix/Workaround Solution 1: Enable CFD interrupt. If CFD interrupt is issues after turning off the CFD, switch back to original main clock source. Solution 2: Only turn off the CFD while running the main clock on RCSYS. 6. Sleepwalking in idle and frozen sleep mode will mask all other PB clocks If the CPU is in idle or frozen sleep mode and a module is in a state that triggers sleep walking, all PB clocks will be masked except the PB clock to the sleepwalking module. Fix/Workaround Mask all clock requests in the PM.PPCR register before going into idle or frozen mode. 35.2.4 SCIF 1. PCLKSR.OSC32RDY bit might not be cleared after disabling OSC32K In some cases the OSC32RDY bit in the PCLKSR register will not be cleared when OSC32K is disabled. Fix/Workaround When re-enabling the OSC32K, read the PCLKSR.OSC32RDY bit. If this bit is: 0: Follow normal procedures. 1: Ignore the PCLKSR.OSC32RDY and ISR.OSC32RDY bit. Use the Frequency Meter (FREQM) to determine if the OSC32K clock is ready. The OSC32K clock is ready when the FREQM measures a non-zero frequency. 2. The RC32K output on PA20 is not always permanently disabled The RC32K output on PA20 may sometimes re-appear. Fix/Workaround Before using RC32K for other purposes, the following procedure has to be followed in order to properly disable it: - Run the CPU on RCSYS - Disable the output to PA20 by writing a zero to PM.PPCR.RC32OUT - Enable RC32K by writing a one to SCIF.RC32KCR.EN, and wait for this bit to be read as one 815 32099I–01/2012 AT32UC3L016/32/64 - Disable RC32K by writing a zero to SCIF.RC32KCR.EN, and wait for this bit to be read as zero. 35.2.5 AST 1. Reset may set status bits in the AST If a reset occurs and the AST is enabled, the SR.ALARM0, SR.PER0, and SR.OVF bits may be set. Fix/Workaround If the part is reset and the AST is used, clear all bits in the Status Register before entering sleep mode. 2. AST wake signal is released one AST clock cycle after the BUSY bit is cleared After writing to the Status Clear Register (SCR) the wake signal is released one AST clock cycle after the BUSY bit in the Status Register (SR.BUSY) is cleared. If entering sleep mode directly after the BUSY bit is cleared the part will wake up immediately. Fix/Workaround Read the Wake Enable Register (WER) and write this value back to the same register. Wait for BUSY to clear before entering sleep mode. 35.2.6 WDT 1. Clearing the Watchdog Timer (WDT) counter in second half of timeout period will issue a Watchdog reset If the WDT counter is cleared in the second half of the timeout period, the WDT will immediately issue a Watchdog reset. Fix/Workaround Use twice as long timeout period as needed and clear the WDT counter within the first half of the timeout period. If the WDT counter is cleared after the first half of the timeout period, you will get a Watchdog reset immediately. If the WDT counter is not cleared at all, the time before the reset will be twice as long as needed. 2. WDT Control Register does not have synchronization feedback When writing to the Timeout Prescale Select (PSEL), Time Ban Prescale Select (TBAN), Enable (EN), or WDT Mode (MODE) fieldss of the WDT Control Register (CTRL), a synchronizer is started to propagate the values to the WDT clcok domain. This synchronization takes a finite amount of time, but only the status of the synchronization of the EN bit is reflected back to the user. Writing to the synchronized fields during synchronization can lead to undefined behavior. Fix/Workaround -When writing to the affected fields, the user must ensure a wait corresponding to 2 clock cycles of both the WDT peripheral bus clock and the selected WDT clock source. -When doing writes that changes the EN bit, the EN bit can be read back until it reflects the written value. 35.2.7 GPIO 1. Clearing GPIO interrupt may fail Writing a one to the GPIO.IFRC register to clear an interrupt will be ignored if interrupt is enabled for the corresponding port. Fix/Workaround Disable the interrupt, clear it by writing a one to GPIO.IFRC, then enable the interrupt. 816 32099I–01/2012 AT32UC3L016/32/64 35.2.8 SPI 1. SPI data transfer hangs with CSR0.CSAAT==1 and MR.MODFDIS==0 When CSR0.CSAAT==1 and mode fault detection is enabled (MR.MODFDIS==0), the SPI module will not start a data transfer. Fix/Workaround Disable mode fault detection by writing a one to MR.MODFDIS. 2. Disabling SPI has no effect on the SR.TDRE bit Disabling SPI has no effect on the SR.TDRE bit whereas the write data command is filtered when SPI is disabled. Writing to TDR when SPI is disabled will not clear SR.TDRE. If SPI is disabled during a PDCA transfer, the PDCA will continue to write data to TDR until its buffer is empty, and this data will be lost. Fix/Workaround Disable the PDCA, add two NOPs, and disable the SPI. To continue the transfer, enable the SPI and PDCA. 3. SPI disable does not work in SLAVE mode SPI disable does not work in SLAVE mode. Fix/Workaround Read the last received data, then perform a software reset by writing a one to the Software Reset bit in the Control Register (CR.SWRST). 4. SPI bad serial clock generation on 2nd chip_select when SCBR=1, CPOL=1, and NCPHA=0 When multiple chip selects (CS) are in use, if one of the baudrates equal 1 while one (CSRn.SCBR=1) of the others do not equal 1, and CSRn.CPOL=1 and CSRn.NCPHA=0, then an additional pulse will be generated on SCK. Fix/Workaround When multiple CS are in use, if one of the baudrates equals 1, the others must also equal 1 if CSRn.CPOL=1 and CSRn.NCPHA=0. 5. SPI mode fault detection enable causes incorrect behavior When mode fault detection is enabled (MR.MODFDIS==0), the SPI module may not operate properly. Fix/Workaround Always disable mode fault detection before using the SPI by writing a one to MR.MODFDIS. 35.2.9 TWI 1. TWIM SR.IDLE goes high immediately when NAK is received When a NAK is received and there is a non-zero number of bytes to be transmitted, SR.IDLE goes high immediately and does not wait for the STOP condition to be sent. This does not cause any problem just by itself, but can cause a problem if software waits for SR.IDLE to go high and then immediately disables the TWIM by writing a one to CR.MDIS. Disabling the TWIM causes the TWCK and TWD pins to go high immediately, so the STOP condition will not be transmitted correctly. Fix/Workaround If possible, do not disable the TWIM. If it is absolutely necessary to disable the TWIM, there must be a software delay of at least two TWCK periods between the detection of SR.IDLE==1 and the disabling of the TWIM. 2. TWIM TWALM polarity is wrong 817 32099I–01/2012 AT32UC3L016/32/64 The TWALM signal in the TWIM is active high instead of active low. Fix/Workaround Use an external inverter to invert the signal going into the TWIM. When using both TWIM and TWIS on the same pins, the TWALM cannot be used. 3. TWIS may not wake the device from sleep mode If the CPU is put to a sleep mode (except Idle and Frozen) directly after a TWI Start condition, the CPU may not wake upon a TWIS address match. The request is NACKed. Fix/Workaround When using the TWI address match to wake the device from sleep, do not switch to sleep modes deeper than Frozen. Another solution is to enable asynchronous EIC wake on the TWIS clock (TWCK) or TWIS data (TWD) pins, in order to wake the system up on bus events. 4. SMBALERT bit may be set after reset The SMBus Alert (SMBALERT) bit in the Status Register (SR) might be erroneously set after system reset. Fix/Workaround After system reset, clear the SR.SMBALERT bit before commencing any TWI transfer. 5. Clearing the NAK bit before the BTF bit is set locks up the TWI bus When the TWIS is in transmit mode, clearing the NAK Received (NAK) bit of the Status Register (SR) before the end of the Acknowledge/Not Acknowledge cycle will cause the TWIS to attempt to continue transmitting data, thus locking up the bus. Fix/Workaround Clear SR.NAK only after the Byte Transfer Finished (BTF) bit of the same register has been set. 6. TWIS stretch on Address match error When the TWIS stretches TWCK due to a slave address match, it also holds TWD low for the same duration if it is to be receiving data. When TWIS releases TWCK, it releases TWD at the same time. This can cause a TWI timing violation. Fix/Workaround None. 35.2.10 PWMA 1. BUSY bit is never cleared after writes to the Control Register (CR) When writing a non-zero value to CR.TOP, CR.SPREAD, or CR.TCLR when the PWMA is disabled (CR.EN == 0), the BUSY bit in the Status Register (SR.BUSY) will be set, but never cleared. Fix/Workaround When writing a non-zero value to CR.TOP, CR.SPREAD, or CR.TCLR, make sure the PWMA is enabled, or simultaneously enable the PWMA by writing a one to CR.EN. 2. Incoming peripheral events are discarded during duty cycle register update Incoming peripheral events to all applied channels will be discarded if a duty cycle update is received from the user interface in the same PWMA clock period. Fix/Workaround Ensure that duty cycle writes from the user interface are not performed in a PWMA period when an incoming peripheral event is expected. 818 32099I–01/2012 AT32UC3L016/32/64 35.2.11 ADCIFB 1. Using STARTUPTIME larger than 0x1F will freeze the ADC Writing a value larger than 0x1F to the Startup Time field in the ADC Configuration Register (ACR.STARTUP) will freeze the ADC, and the Busy Status bit in the Status Register (SR.BUSY) will never be cleared. Fix/Workaround Do not write values larger than 0x1F to ACR.STARTUP. 35.2.12 CAT 1. CAT asynchronous wake will be delayed by one AST event period If the CAT detects a condition the should asynchronously wake the device in static mode, the asynchronous wake will not occur until the next AST event. For example, if the AST is generating events to the CAT every 50ms, and the CAT detects a touch at t=9200ms, the asynchronous wake will occur at t=9250ms. Fix/Workaround None. 2. CAT QMatrix sense capacitors discharged prematurely At the end of a QMatrix burst charging sequence that uses different burst count values for different Y lines, the Y lines may be incorrectly grounded for up to n-1 periods of the peripheral bus clock, where n is the ratio of the PB clock frequency to the GCLK_CAT frequency. This results in premature loss of charge from the sense capacitors and thus increased variability of the acquired count values. Fix/Workaround Enable the 1kOhm drive resistors on all implemented QMatrix Y lines (CSA 1, 3, 5, 7, 9, 11, 13, and/or 15) by writing ones to the corresponding odd bits of the CSARES register. 3. CAT module does not terminate QTouch burst on detect The CAT module does not terminate a QTouch burst when the detection voltage is reached on the sense capacitor. This can cause the sense capacitor to be charged more than necessary. Depending on the dielectric absorption characteristics of the capacitor, this can lead to unstable measurements. Fix/Workaround Use the minimum possible value for the MAX field in the ATCFG1, TG0CFG1, and TG1CFG1 registers. 4. Autonomous CAT acquisition must be longer than AST source clock period When using the AST to trigger CAT autonomous touch acquisition in sleep modes where the CAT bus clock is turned off, the CAT will start several acquisitions if the period of the AST source clock is larger than one CAT acquisition. One AST clock period after the AST trigger, the CAT clock will automatically stop and the CAT acquisition can be stopped prematurely, ruining the result. Fix/Workaround Always ensure that the ATCFG1.max field is set so that the duration of the autonomous touch acquisition is greater than one clock period of the AST source clock. 35.2.13 aWire 1. aWire CPU clock speed robustness The aWire memory speed request command counter warps at clock speeds below approximately 5kHz. Fix/Workaround 819 32099I–01/2012 AT32UC3L016/32/64 None. 2. The aWire debug interface is reset after leaving Shutdown mode If the aWire debug mode is used as debug interface and the program enters Shutdown mode, the aWire interface will be reset when the part receives a wakeup either from the WAKE_N pin or the AST. Fix/Workaround None. 35.2.14 CHIP 1. In 3.3V Single Supply Mode, the Analog Comparator inputs affects the device’s ability to start When using the 3.3V Single Supply Mode the state of the Analog Comparator input pins can affect the device’s ability to release POR reset. This is due to an interaction between the Analog Comparator input pins and the POR circuitry. The issue is not present in the 1.8V Supply Mode or the 3.3V Supply Mode with 1.8V Regulated I/O Lines. Fix/Workaround ACREFN (pin PA16) must be connected to GND until the POR reset is released and the Analog Comparator inputs should not be driven higher than 1.0 V until the POR reset is released. 2. Increased Power Consumption in VDDIO in sleep modes If OSC0 is enabled in crystal mode when entering a sleep mode where the OSC0 is disabled, this will lead to an increased power consumption in VDDIO. Fix/Workaround Solution 1: Disable OSC0 by writing a zero to the Oscillator Enable bit in the System Control Interface (SCIF) Oscillator Control Register (SCIF.OSC0CTRL.OSCEN) before going to a sleep mode where OSC0 is disabled. Solution 2: Pull down or up XIN0 or XOUT0 with 1MOhm resistor. 35.2.15 I/O Pins 1. PA17 has low ESD tolerance PA17 only tolerates 500V ESD pulses (Human Body Model). Fix/Workaround Care must be taken during manufacturing and PCB design. 35.3 Rev. C Not sampled. 35.4 35.4.1 Rev. B Processor and Architecture 1. RETS behaves incorrectly when MPU is enabled RETS behaves incorrectly when MPU is enabled and MPU is configured so that system stack is not readable in unprivileged mode. Fix/Workaround Make system stack readable in unprivileged mode, or return from supervisor mode using rete instead of rets. This requires: 1. Changing the mode bits from 001 to 110 before issuing the instruction. Updating the mode bits to the desired value must be done using a single mtsr instruction so it is done 820 32099I–01/2012 AT32UC3L016/32/64 atomically. Even if this step is generally described as not safe in the UC technical reference manual, it is safe in this very specific case. 2. Execute the RETE instruction. 2. Hardware breakpoints may corrupt MAC results Hardware breakpoints on MAC instructions may corrupt the destination register of the MAC instruction. Fix/Workaround Place breakpoints on earlier or later instructions. 3. Privilege violation when using interrupts in application mode with protected system stack If the system stack is protected by the MPU and an interrupt occurs in application mode, an MPU DTLB exception will occur. Fix/Workaround Make a DTLB Protection (Write) exception handler which permits the interrupt request to be handled in privileged mode. 35.4.2 PDCA 1. PCONTROL.CHxRES is non-functional PCONTROL.CHxRES is non-functional. Counters are reset at power-on, and cannot be reset by software. Fix/Workaround Software needs to keep history of performance counters. 2. Transfer error will stall a transmit peripheral handshake interface If a transfer error is encountered on a channel transmitting to a peripheral, the peripheral handshake of the active channel will stall and the PDCA will not do any more transfers on the affected peripheral handshake interface. Fix/Workaround Disable and then enable the peripheral after the transfer error. 3. VERSION register reads 0x120 The VERSION register reads 0x120 instead of 0x122. Fix/Workaround None. 35.4.3 FLASHCDW 1. Fuse Programming Programming fuses does not work. Fix/Workaround Do not program fuses. All fuses will be erased during chip erase command. 2. Chip Erase When performing a chip erase, the device may report that it is protected (IR=0x11) and that the erase failed, even if it was successful. Fix/Workaround Perform a reset before any further read and programming. 3. Wait 500 ns before reading from the flash after switching read mode After switching between normal read mode and high-speed read mode, the application must wait at least 500ns before attempting any access to the flash. 821 32099I–01/2012 AT32UC3L016/32/64 Fix/Workaround Solution 1: Make sure that the appropriate instructions are executed from RAM, and that a waiting-loop is executed from RAM waiting 500ns or more before executing from flash. Solution 2. Execute from flash with a clock with period longer than 500ns. This guarantees that no new read access is attempted before the flash has had time to settle in the new read mode. 4. Flash self programming may fail in one wait state mode Writes in flash and user pages may fail if executing code is located in address space mapped to flash, and the flash controller is configured in one wait state mode (the Flash Wait State bit in the Flash Control Register (FCR.FWS) is one). Fix/Workaround Solution 1: Configure the flash controller in zero wait state mode (FCR.FWS=0). Solution 2: Configure the HMATRIX master 1 (CPU Instruction) to use the unlimited burst length transfer mode (MCFG1.ULBT=0), and the HMATRIX slave 0 (FLASHCDW) to use the maximum slot cycle limit (SCFG0.SLOT_CYCLE=255). 5. VERSION register reads 0x100 The VERSION register reads 0x100 instead of 0x102. Fix/Workaround None. 35.4.4 SAU 1. The SR.IDLE bit reads as zero The IDLE bit in the Status Register (SR.IDLE) reads as zero. Fix/Workaround None. 2. Open Mode is not functional The Open Mode is not functional. Fix/Workaround None. 3. VERSION register reads 0x100 The VERSION register reads 0x100 instead of 0x110. Fix/Workaround None. 35.4.5 HMATRIX 1. In the PRAS and PRBS registers, the MxPR fields are only two bits In the PRAS and PRBS registers, the MxPR fields are only two bits wide, instead of four bits. The unused bits are undefined when reading the registers. Fix/Workaround Mask undefined bits when reading PRAS and PRBS. 35.4.6 Power Manager 1. CONFIG register reads 0x4F The CONFIG register reads 0x4F instead of 0x43. Fix/Workaround None. 822 32099I–01/2012 AT32UC3L016/32/64 2. It is not possible to mask the request clock requests It is not possible to mask the request clock requests using PPCR. Fix/Workaround None. 3. Static mode cannot be entered if the WDT is using OSC32 If the WDT is using OSC32 as clock source and the user tries to enter Static mode, the Deepstop mode will be entered instead. Fix/Workaround None. 4. Clock Failure Detector (CFD) does not work Clock Failure Detector (CFD) does not work. Fix/Workaround None. 5. WCAUSE register should not be used The WCAUSE register should not be used. Fix/Workaround None. 6. PB writes via debugger in sleep modes are blocked during sleepwalking During sleepwalking, PB writes performed by a debugger will be discarded by all PB modules except the module that is requesting the clock. Fix/Workaround None. 7. Clock sources will not be stopped in Static mode if the difference between CPU and PBx division factor is larger than 4 If the division factor between the CPU/HSB and PBx frequencies is more than 4 when entering a sleep mode where the system RC oscillator (RCSYS) is turned off, the high speed clock sources will not be turned off. This will result in a significantly higher power consumption during the sleep mode. Fix/Workaround Before going to sleep modes where RCSYS is stopped, make sure the division factor between CPU/HSB and PBx frequencies is less than or equal to 4. 8. Disabling POR33 may generate spurious resets Depending on operating conditions, POR33 may generate a spurious reset in one of the following cases: - When POR33 is disabled from the user interface - When SM33 supply monitor is enabled - When entering Shutdown mode while debugging the chip using JTAG or aWire interface In the listed cases, writing a one to the bit VREGCR.POR33MASK in the System Control Interface (SCIF) to mask the POR33 reset will be ineffective Fix/Workaround - Do not disable POR33 using the user interface - Do not use the SM33 supply monitor - Do not enter Shutdown mode if a debugger is connected to the chip 9. Instability when exiting sleep walking If all the following operating conditions are true, exiting sleep walking might lead to instability: 823 32099I–01/2012 AT32UC3L016/32/64 - The OSC0 is enabled in external clock mode (OSCCTRL0.OSCEN == 1 and OSCCTRL0.MODE == 0) - A sleep mode where the OSC0 is automatically disabled is entered - The device enters sleep walking Fix/Workaround Do not run OSC0 in external clock mode if sleepwalking is expected to be used. 10. Sleepwalking in idle and frozen sleep mode will mask all other PB clocks If the CPU is in idle or frozen sleep mode and a module is in a state that triggers sleep walking, all PB clocks will be masked except the PB clock to the sleepwalking module. Fix/Workaround Mask all clock requests in the PM.PPCR register before going into idle or frozen mode. 11. VERSION register reads 0x400 The VERSION register reads 0x400 instead of 0x411. Fix/Workaround None. 35.4.7 SCIF 1. The DFLL should be slowed down before disabling it The frequency of the DFLL should be set to minimum before disabling it. Fix/Workaround Before disabling the DFLL the value of the COARSE register should be zero. 2. Writing to ICR masks new interrupts received in the same clock cycle Writing to ICR masks any new SCIF interrupt received in the same clock cycle, regardless of write value. Fix/Workaround For every interrupt except BODDET, SM33DET, and VREGOK the PCLKSR register can be read to detect new interrupts. BODDET, SM33DET and VREGOK interrupts will not be generated if they occur whilst writing to the ICR register. 3. FINE value for DFLL is not correct when dithering is disabled In open loop mode, the FINE value used by the DFLL DAC is offset by two compared to the value written to the DFLL0CONF.FINE field. The value used by the DFLL DAC is DFLL0CONF.FINE-0x002. If DFLL0CONF.FINE is written to 0x000, 0x001 or 0x002 the value used by the DFLL DAC will be 0x1FE, 0x1FF, or 0x000 respectively. Fix/Workaround Write the desired value added by two to the DFLL0CONF.FINE field. 4. BODVERSION register reads 0x100 The BODVERSION register reads 0x100 instead of 0x101 Fix/Workaround None. 5. VREGCR.DEEPMODEDISABLE bit is not readable VREGCR.DEEPMODEDISABLE bit is not readable. Fix/Workaround None. 6. DFLL step size should be seven or lower when below 30MHz If max step size is above seven, the DFLL might not lock at the correct frequency if the target frequency is below 30MHz. 824 32099I–01/2012 AT32UC3L016/32/64 Fix/Workaround If the target frequency is below 30MHz, use a max step size (DFLL0MAXSTEP.MAXSTEP) of seven or lower. 7. Generic clock sources are kept running in sleep modes If a clock is used as a source for a generic clock when going to a sleep mode where clock sources are stopped, the source of the generic clock will be kept running. Please refer to Power Manager chapter for details about sleep modes. Fix/Workaround Disable generic clocks before going to sleep modes where clock sources are stopped to save power. 8. DFLL clock is unstable with a fast reference clock The DFLL clock can be unstable when a fast clock is used as a reference clock in closed loop mode. Fix/Workaround Use the 32KHz crystal oscillator clock, or a clock with a similar frequency, as DFLLIF reference clock. 9. DFLLIF indicates coarse lock too early The DFLLIF might indicate coarse lock too early, the DFLL will lose coarse lock and regain it later. Fix/Workaround Use max step size (DFLL0MAXSTEP.MAXSTEP) of 4 or higher. 10. DFLLIF dithering does not work The DFLLIF dithering does not work. Fix/Workaround None. 11. DFLLIF might lose fine lock when dithering is disabled When dithering is disabled and fine lock has been acquired, the DFLL might lose the fine lock resulting in up to 20% over-/undershoot. Fix/Workaround Solution 1: When the DFLL is used as main clock source, the target frequency of the DFLL should be 20% below the maximum operating frequency of the CPU. Don’t use the DFLL as clock source for frequency sensitive applications. Solution 2: Do not use the DFLL in closed loop mode. 12. GCLK5 is non-functional GCLK5 is non-functional. Fix/Workaround None. 13. BRIFA is non-functional BRIFA is non-functional. Fix/Workaround None. 14. SCIF VERSION register reads 0x100 SCIFVERSION register reads 0x100 instead of 0x102. Fix/Workaround None. 825 32099I–01/2012 AT32UC3L016/32/64 15. BODVERSION register reads 0x100 BODVERSION register reads 0x100 instead of 0x101. Fix/Workaround None. 16. DFLLVERSION register reads 0x200 DFLLVERSION register reads 0x200 instead of 0x201. Fix/Workaround None. 17. RCCRVERSION register reads 0x100 RCCRVERSION register reads 0x100 instead of 0x101. Fix/Workaround None. 18. OSC32VERSION register reads 0x100 OSC32VERSION register reads 0x100 instead of 0x101. Fix/Workaround None. 19. VREGVERSION register reads 0x100 VREGVERSION register reads 0x100 instead of 0x101. Fix/Workaround None. 20. RC120MVERSION register reads 0x100 RC120MVERSION register reads 0x100 instead of 0x101. Fix/Workaround None. 35.4.8 AST 1. AST wake signal is released one AST clock cycle after the BUSY bit is cleared After writing to the Status Clear Register (SCR) the wake signal is released one AST clock cycle after the BUSY bit in the Status Register (SR.BUSY) is cleared. If entering sleep mode directly after the BUSY bit is cleared the part will wake up immediately. Fix/Workaround Read the Wake Enable Register (WER) and write this value back to the same register. Wait for BUSY to clear before entering sleep mode. 35.4.9 WDT 1. Clearing the WDT in window mode In window mode, if the WDT is cleared 2TBAN CLK_WDT cycles after entering the window, the counter will be cleared, but will not exit the window. If this occurs, the SR.WINDOW bit will not be cleared after clearing the WDT. Fix/Workaround Check SR.WINDOW immediately after clearing the WDT. If set then clear the WDT once more. 2. Clearing the Watchdog Timer (WDT) counter in second half of timeout period will issue a Watchdog reset If the WDT counter is cleared in the second half of the timeout period, the WDT will immediately issue a Watchdog reset. 826 32099I–01/2012 AT32UC3L016/32/64 Fix/Workaround Use twice as long timeout period as needed and clear the WDT counter within the first half of the timeout period. If the WDT counter is cleared after the first half of the timeout period, you will get a Watchdog reset immediately. If the WDT counter is not cleared at all, the time before the reset will be twice as long as needed. 3. VERSION register reads 0x400 The VERSION register reads 0x400 instead of 0x402. Fix/Workaround None. 4. WDT Control Register does not have synchronization feedback When writing to the Timeout Prescale Select (PSEL), Time Ban Prescale Select (TBAN), Enable (EN), or WDT Mode (MODE) fieldss of the WDT Control Register (CTRL), a synchronizer is started to propagate the values to the WDT clcok domain. This synchronization takes a finite amount of time, but only the status of the synchronization of the EN bit is reflected back to the user. Writing to the synchronized fields during synchronization can lead to undefined behavior. Fix/Workaround -When writing to the affected fields, the user must ensure a wait corresponding to 2 clock cycles of both the WDT peripheral bus clock and the selected WDT clock source. -When doing writes that changes the EN bit, the EN bit can be read back until it reflects the written value. 35.4.10 FREQM 1. Measured clock (CLK_MSR) sources 15-17 are shifted CLKSEL = 14 selects the RC120M AW clock, CLKSEL = 15 selects the RC120M clock, and CLKSEL = 16 selects the RC32K clock as source for the measured clock (CLK_MSR). Fix/Workaround None. 2. GCLK5 can not be used as source for the CLK_MSR The frequency for GCLK5 can not be measured by the FREQM. Fix/Workaround None. 35.4.11 GPIO 1. GPIO interrupt can not be cleared when interrupts are disabled The GPIO interrupt can not be cleared unless the interrupt is enabled for the pin. Fix/Workaround Enable interrupt for the corresponding pin, then clear the interrupt. 2. VERSION register reads 0x210 The VERSION register reads 0x210 instead of 0x211. Fix/Workaround None. 35.4.12 USART 1. The RTS output does not function correctly in hardware handshaking mode 827 32099I–01/2012 AT32UC3L016/32/64 The RTS signal is not generated properly when the USART receives data in hardware handshaking mode. When the Peripheral DMA receive buffer becomes full, the RTS output should go high, but it will stay low. Fix/Workaround Do not use the hardware handshaking mode of the USART. If it is necessary to drive the RTS output high when the Peripheral DMA receive buffer becomes full, use the normal mode of the USART. Configure the Peripheral DMA Controller to signal an interrupt when the receive buffer is full. In the interrupt handler code, write a one to the RTSDIS bit in the USART Control Register (CR). This will drive the RTS output high. After the next DMA transfer is started and a receive buffer is available, write a one to the RTSEN bit in the USART CR so that RTS will be driven low. 35.4.13 SPI 1. SPI data transfer hangs with CSR0.CSAAT==1 and MR.MODFDIS==0 When CSR0.CSAAT==1 and mode fault detection is enabled (MR.MODFDIS==0), the SPI module will not start a data transfer. Fix/Workaround Disable mode fault detection by writing a one to MR.MODFDIS. 2. Disabling SPI has no effect on the SR.TDRE bit Disabling SPI has no effect on the SR.TDRE bit whereas the write data command is filtered when SPI is disabled. Writing to TDR when SPI is disabled will not clear SR.TDRE. If SPI is disabled during a PDCA transfer, the PDCA will continue to write data to TDR until its buffer is empty, and this data will be lost. Fix/Workaround Disable the PDCA, add two NOPs, and disable the SPI. To continue the transfer, enable the SPI and PDCA. 3. SPI disable does not work in SLAVE mode SPI disable does not work in SLAVE mode. Fix/Workaround Read the last received data, then perform a software reset by writing a one to the Software Reset bit in the Control Register (CR.SWRST). 4. SPI bad serial clock generation on 2nd chip_select when SCBR=1, CPOL=1, and NCPHA=0 When multiple chip selects (CS) are in use, if one of the baudrates equal 1 while one (CSRn.SCBR=1) of the others do not equal 1, and CSRn.CPOL=1 and CSRn.NCPHA=0, then an additional pulse will be generated on SCK. Fix/Workaround When multiple CS are in use, if one of the baudrates equals 1, the others must also equal 1 if CSRn.CPOL=1 and CSRn.NCPHA=0. 5. SPI mode fault detection enable causes incorrect behavior When mode fault detection is enabled (MR.MODFDIS==0), the SPI module may not operate properly. Fix/Workaround Always disable mode fault detection before using the SPI by writing a one to MR.MODFDIS. 35.4.14 TWI 1. TWI pins are not SMBus compliant 828 32099I–01/2012 AT32UC3L016/32/64 The TWI pins draw current when they are supplied with 3.3V and the part is left unpowered. Fix/Workaround None. 2. PA21, PB04, and PB05 are not 5V tolerant Pins PA21, PB04, and PB05 are only 3.3V tolerant. Fix/Workaround None. 3. PB04 SMBALERT function should not be used The SMBALERT function from TWIMS0 should not be selected on pin PB04. Fix/Workaround None. 4. TWIM STOP bit in IMR always reads as zero The STOP bit in IMR always reads as zero. Fix/Workaround None. 5. Disabled TWIM drives TWD and TWCK low When the TWIM is disabled, it drives the TWD and TWCK signals with logic level zero. This can lead to communication problems with other devices on the TWI bus. Fix/Workaround Enable the TWIM first and then enable the TWD and TWCK peripheral pins in the GPIO controller. If it is necessary to disable the TWIM, first disable the TWD and TWCK peripheral pins in the GPIO controller and then disable the TWIM. 6. TWIM SR.IDLE goes high immediately when NAK is received When a NAK is received and there is a non-zero number of bytes to be transmitted, SR.IDLE goes high immediately and does not wait for the STOP condition to be sent. This does not cause any problem just by itself, but can cause a problem if software waits for SR.IDLE to go high and then immediately disables the TWIM by writing a one to CR.MDIS. Disabling the TWIM causes the TWCK and TWD pins to go high immediately, so the STOP condition will not be transmitted correctly. Fix/Workaround If possible, do not disable the TWIM. If it is absolutely necessary to disable the TWIM, there must be a software delay of at least two TWCK periods between the detection of SR.IDLE==1 and the disabling of the TWIM. 7. TWIM TWALM polarity is wrong The TWALM signal in the TWIM is active high instead of active low. Fix/Workaround Use an external inverter to invert the signal going into the TWIM. When using both TWIM and TWIS on the same pins, the TWALM cannot be used. 8. TWIS CR.STREN does not work in deep sleep modes When the device is in Stop, DeepStop, or Static mode, address reception will not wake device if both CR.SOAM and CR.STREN are one. Fix/Workaround Do not write both CR.STREN and CR.SOAM to one if the device needs to wake from deep sleep modes. 9. TWI0.TWCK on PB05 is non-functional TWI0.TWCK on PB05 is non-functional. 829 32099I–01/2012 AT32UC3L016/32/64 Fix/Workaround Use TWI0.TWCK on other pins. 10. TWIM Version Register reads zero TWIM Version Register (VR) reads zero instead of 0x101 Fix/Workaround None. 11. TWIS Version Register reads zero TWIS Version Register (VR) reads zero instead of 0x112 Fix/Workaround None. 12. SMBALERT bit may be set after reset The SMBus Alert (SMBALERT) bit in the Status Register (SR) might be erroneously set after system reset. Fix/Workaround After system reset, clear the SR.SMBALERT bit before commencing any TWI transfer. 13. Clearing the NAK bit before the BTF bit is set locks up the TWI bus When the TWIS is in transmit mode, clearing the NAK Received (NAK) bit of the Status Register (SR) before the end of the Acknowledge/Not Acknowledge cycle will cause the TWIS to attempt to continue transmitting data, thus locking up the bus. Fix/Workaround Clear SR.NAK only after the Byte Transfer Finished (BTF) bit of the same register has been set. 14. TWIS stretch on Address match error When the TWIS stretches TWCK due to a slave address match, it also holds TWD low for the same duration if it is to be receiving data. When TWIS releases TWCK, it releases TWD at the same time. This can cause a TWI timing violation. Fix/Workaround None. 35.4.15 PWMA 1. PARAMETER register reads 0x2424 The PARAMETER register reads 0x2424 instead of 0x24. Fix/Workaround None. 2. Writing to the duty cycle registers when the timebase counter overflows can give an undefined result The duty cycle registers will be corrupted if written when the timebase counter overflows. If the duty cycle registers are written exactly when the timebase counter overflows at TOP, the duty cycle registers may become corrupted. Fix/Workaround Write to the duty cycle registers only directly after the Timebase Overflow bit in the status register is set. 3. Open Drain mode does not work The open drain mode does not work. Fix/Workaround None. 830 32099I–01/2012 AT32UC3L016/32/64 4. BUSY bit is never cleared after writes to the Control Register (CR) When writing a non-zero value to CR.TOP, CR.SPREAD, or CR.TCLR when the PWMA is disabled (CR.EN == 0), the BUSY bit in the Status Register (SR.BUSY) will be set, but never cleared. Fix/Workaround When writing a non-zero value to CR.TOP, CR.SPREAD, or CR.TCLR, make sure the PWMA is enabled, or simultaneously enable the PWMA by writing a one to CR.EN. 5. Incoming peripheral events are discarded during duty cycle register update Incoming peripheral events to all applied channels will be discarded if a duty cycle update is received from the user interface in the same PWMA clock period. Fix/Workaround Ensure that duty cycle writes from the user interface are not performed in a PWMA period when an incoming peripheral event is expected. 6. VERSION register reads 0x100 The VERSION register reads 0x100 instead of 0x101. Fix/Workaround None. 35.4.16 TC 1. When the main clock is RCSYS, TIMER_CLOCK5 is equal to CLK_PBA When the main clock is generated from RCSYS, TIMER_CLOCK5 is equal to CLK_PBA and not CLK_PBA/128. Fix/Workaround None. 35.4.17 ADCIFB 1. Pendetect in sleep modes without CLK_ADCIFB will not wake the system The pendetect will not wake the system from a sleep mode if the clock for the ADCIFB (CLK_ADCIFB) is turned off. Fix/Workaround Use a sleep mode where CLK_ADCIFB is not turned off to wake the part using pendetect. 2. 8-bit mode is not working Do not use the ADCIFB 8-bit mode. Fix/Workaround Use the 10-bit mode and shift right by two bits. 3. Using STARTUPTIME larger than 0x1F will freeze the ADC Writing a value larger than 0x1F to the Startup Time field in the ADC Configuration Register (ACR.STARTUP) will freeze the ADC, and the Busy Status bit in the Status Register (SR.BUSY) will never be cleared. Fix/Workaround Do not write values larger than 0x1F to ACR.STARTUP. 4. ADC channels six to eight are non-functional ADC channels six to eight are non-functional. Fix/Workaround None. 5. VERSION register reads 0x100 831 32099I–01/2012 AT32UC3L016/32/64 The VERSION register reads 0x100 instead of 0x101. Fix/Workaround None. 35.4.18 ACIFB 1. Generic clock sources in sleep modes. The ACIFB should not use RC32K or CLK_1K as generic clock source if the chip uses sleep modes. Fix/Workaround None. 2. Negative offset The static offset of the analog comparator is approximately -50mV Fix/Workaround None. 3. CONFW.WEVSRC and CONFW.WEVEN are not correctly described in the user interface CONFW.WEVSRC is only two bits instead of three bits wide. Only values 0, 1, and 2 can be written to this register. CONFW.WEVEN is in bit position 10 instead of 11. Fix/Workaround Only write values 0, 1, and 2 to CONFW.WEVSRC. When reading CONFW.WEVSRC, disregard the third bit. Read/write bit 10 to access CONFW.WEVEN. 4. VERSION register reads 0x200 The VERSION register reads 0x200 instead of 0x212. Fix/Workaround None. 35.4.19 CAT 1. Switch off discharge current when reaching 0V The discharge current will switch off when reaching MGCFG1.MAX, not when reaching 0V. Fix/Workaround None. 2. CAT external capacitors are not clamped to ground when CAT is idle The CAT module does not clamp the external capacitors to ground when it is idle. The capacitors are left floating, so they could accumulate small amounts of charge. Fix/Workaround None. 3. DISHIFT field is stuck at zero The DISHIFT field in the MGCFG1, TGACFG1, TGBCFG1, and ATCFG1 registers is stuck at zero and cannot be written to a different value. Capacitor discharge time will only be determined by the DILEN field. Fix/Workaround None. 4. MGCFG2.CONSEN field is stuck at zero The CONSEN field in the MGCFG2 register is stuck at zero and cannot be written to a different value. The CAT consensus filter does not function properly, so termination of QMatrix data acquisition is controlled only by the MAX field in MGCFG1. 832 32099I–01/2012 AT32UC3L016/32/64 Fix/Workaround None. 5. MGCFG2.ACCTRL bit is stuck at zero The ACCTRL bit in the MGCFG2 register is stuck at zero and cannot be written to one. The analog comparators will be constantly enabled. Fix/Workaround None. 6. CAT asynchronous wake will be delayed by one AST event period If the CAT detects a condition the should asynchronously wake the device in static mode, the asynchronous wake will not occur until the next AST event. For example, if the AST is generating events to the CAT every 50ms, and the CAT detects a touch at t=9200ms, the asynchronous wake will occur at t=9250ms. Fix/Workaround None. 7. VERSION register reads 0x100 The VERSION register reads 0x100 instead of 0x200. Fix/Workaround None. 8. CAT module does not terminate QTouch burst on detect The CAT module does not terminate a QTouch burst when the detection voltage is reached on the sense capacitor. This can cause the sense capacitor to be charged more than necessary. Depending on the dielectric absorption characteristics of the capacitor, this can lead to unstable measurements. Fix/Workaround Use the minimum possible value for the MAX field in the ATCFG1, TG0CFG1, and TG1CFG1 registers. 9. Autonomous CAT acquisition must be longer than AST source clock period When using the AST to trigger CAT autonomous touch acquisition in sleep modes where the CAT bus clock is turned off, the CAT will start several acquisitions if the period of the AST source clock is larger than one CAT acquisition. One AST clock period after the AST trigger, the CAT clock will automatically stop and the CAT acquisition can be stopped prematurely, ruining the result. Fix/Workaround Always ensure that the ATCFG1.max field is set so that the duration of the autonomous touch acquisition is greater than one clock period of the AST source clock. 35.4.20 GLOC 1. GLOC is non-functional Glue Logic Controller (GLOC) is non-functional. Fix/Workaround None. 35.4.21 aWire 1. SAB multiaccess reads are not working Reading more than one word, halfword, or byte in one command is not working correctly. Fix/Workaround Split the access into several single word, halfword, or byte accesses. 833 32099I–01/2012 AT32UC3L016/32/64 2. If a reset happens during the last SAB write, the aWire will stall If a reset happens during the last word, halfword or byte write the aWire will wait forever for an acknowledge from the SAB. Fix/Workaround Reset the aWire by keeping the RESET_N line low for 100ms. 3. aWire enable does not work in Static mode aWire enable does not work in Static mode. Fix/Workaround None. 4. aWire CPU clock speed robustness The aWire memory speed request command counter warps at clock speeds below approximately 5kHz. Fix/Workaround None. 5. The aWire debug interface is reset after leaving Shutdown mode If the aWire debug mode is used as debug interface and the program enters Shutdown mode, the aWire interface will be reset when the part receives a wakeup either from the WAKE_N pin or the AST. Fix/Workaround None. 6. aWire PB mapping and PB clock mask number The aWire PB has a different PB address and PB clock mask number. Fix/Workaround Use aWire PB address 0xFFFF6C00 and PB clock (PBAMASK) 24. 7. VERSION register reads 0x200 The VERSION register reads 0x200 instead of 0x210. Fix/Workaround None. 35.4.22 Chip 1. WAKE_N pin can only wake up the chip from Shutdown mode It is not possible to wake up the chip from any other sleep mode than Shutdown using the WAKE_N pin. If the WAKE_N pin is asserted during a sleep mode other than Shutdown, nothing will happen. Fix/Workaround Use an EIC pin to wake up from sleep modes higher than Shutdown. 2. Power consumption in static mode is too high Power consumption in static mode is too high when PA21 is high. Fix/Workaround Ensure PA21 is low. 3. Shutdown mode is not functional Do not enter Shutdown mode. Fix/Workaround None. 4. VDDIN current consumption increase above 1.8V 834 32099I–01/2012 AT32UC3L016/32/64 When VDDIN increases above 1.8V, current on VDDIN increases with up to 40uA. Fix/Workaround None. 5. Increased Power Consumption in VDDIO in sleep modes If OSC0 is enabled in crystal mode when entering a sleep mode where the OSC0 is disabled, this will lead to an increased power consumption in VDDIO. Fix/Workaround Solution 1: Disable OSC0 by writing a zero to the Oscillator Enable bit in the System Control Interface (SCIF) Oscillator Control Register (SCIF.OSC0CTRL.OSCEN) before going to a sleep mode where OSC0 is disabled. Solution 2: Pull down or up XIN0 and XOUT0 with 1MOhm resistor. 35.4.23 I/O Pins 1. PB10 is not 3.3V tolerant. PB10 should be grounded on the PCB and left unused. Fix/Workaround None. 2. Analog multiplexing consumes extra power Current consumption on VDDIO increases when the voltage on analog inputs is close to VDDIO/2. Fix/Workaround None. 3. PA02, PB01, PB04, PB05, and RESET_N have half of the pull-up strength Pins PA02, PB01, PB04, PB05, and RESET_N have half of the specified pull-up strength. Fix/Workaround None. 4. JTAG is enabled at power up The JTAG function on pins PA00, PA01, PA02, and PA03, are enabled after startup. Normal I/O module functionality is not possible on these pins. Fix/Workaround Add a 10kOhm pullup on the reset line. 5. MCKO and MDO[3] are swapped in the AUX1 mapping When using the OCD AUX1 mapping of trace signals MDO[3] is located on pin PB05 and MCKO is located on PB01. Fix/Workaround Swap pins PB01 and PB05 if using OCD AUX1. 835 32099I–01/2012 AT32UC3L016/32/64 36. Datasheet Revision History Please note that the referring page numbers in this section are referred to this document. The referring revision in this section are referring to the document revision. 36.1 36.2 Rev. I - 01/2012 1. Overview - Block diagram: CAT SMP corrected from I/O to output. SPI NPCS corrected from output to I/O. 2. Package and Pinout: PRND signal removed from Signal Descriptions List table and GPIO Controller Function Multiplexing table 3. ADCIFB: PRND signal removed from block diagram. 4. Electrical Characteristics: Added pin input capacitance CIN for TLLGA package for “all normal I/O pins except PA05 , PA07, PA17, PA20, PA21, PB04, PB05”. 5. Errata: Added more errata for TWI and CAT modules. Rev. H - 12/2011 1. Mechanical Characteristics: Updated the note related to the QFN48 Package Drawing.Updated thermal reistance data for TLLGA48 package. Updated package drawings for TQFP48 and QFN48 packages. Soldering profile updated (Time maintained above 217°C.) 2. Memories: Local bus address map corrected: The address offset for port 1 registers is 0x100, not 0x200. 3. SCIF DFLL: Removed “not” from “DFLLnSTEP.CSTEP and DFLLnSTEP.FSTEP should not be lower than 50% of the maximum value of DFLLnCONF.COARSE and DFLLnCONF.FINE”. 4. SCIF VREG POR descriptions updated. POR33 bits added in VREGCR. 5. SCIF VREG: Removed reference to flash fuses for CALIB field, user is recommended not writing to SELVDD field. Removed references to Electrical Characteristics. 6. SCIF: Fuses text removed from some submodules (SM33, VREG). 7. SCIF VREG: Flash recalibration is always done at POR. 8. SCIF SM33: Enabling SM33 will disable the POR33 detector. 9, Erratum regarding OSC32 disabling is not valid for RevB. 10. Flash Controller: Serial number address updated. 11. Block diagram and ADCIFB: Removed PRND signal from block diagram and ADCIFB block diagram. 12. USART: Added CTSIC bit description. 13. Power Manager: Updated Clock division and Clock ready flag sections. 14. ADCIFB: Added DMA section in Product Dependencies. 836 32099I–01/2012 AT32UC3L016/32/64 36.3 36.4 36.5 15. Electrcal Characteristics: Updated SPI timing data. 16. Electrical Characteristics, I/O Pin Characteristics: Added Input capacitance for TLLGA48 package. 17. Errata: Removed erratum regarding SPI RDR.PCS field, as the PCS field has been removed (refer to Section 36.8 on page 838). Rev. G - 06/2011 1. FLASHCDW: FSR register is a Read-only register. Added info about QPRUP. 2. PM: Clarified POR33 masking requirements before shutdown. Added more info about wakeup sources. Added AWEN description. PPCR register reset value corrected. 3. SAU: SR.IDLE definition and reset value corrected. 4. DFLL: Open- and closed-loop operation clarified. 5. OSC32: Added note about OSC32RDY bit not always cleared when disabling OSC32. 6. USART: Major cleanup. 1. Features: Removed superfluous R mark. 2. Package and Pinout, GPIO function multiplexing:TWIMS0-TWCK on PA20 removed. ADCIFBAD[3] on PA17 removed, number of ADC channels are 8, not 9. These were removed from rev. C, but reappeared in rev. E. 1. Overview: Added missing signals in block diagram. 2. Package and pinout: Added note about TWI, SMBUS and 5V tolerant pads in peripheral multiplexing. Added CAT DIS signal to signal descriptions. Removed TBD on ADVREFP minimum voltage. 3. Memories: Added SAU slave address to physical memory map. 4. Supply and startup considerations: VDDIN is using GND as ground pin. Clarified references to PORs in startup considerations. 5. FLASHCDW: Added serial number location to module configuration section. 6. PM: Added more info about the WAKE_N pin. Added info about CLK_PM, Updated the selection main clock source section. 7. SCIF: Major chapter update. 8. AST: Updated digital tuner formula and conditions. 9. GPIO: Updated GPER reset value and added more registers with non-zero reset value. 10. CAT: Added info about VDIVEN and discharge current formula. Rev. F- 11/2010 Rev. E- 10/2010 837 32099I–01/2012 AT32UC3L016/32/64 36.6 11. ADCIFB: Fixed Sample and Hold time formula. 12. GLOC: Added info about pullup control and renamed LUTCR register to CR. 13. TC: Added features and version register. 14. SAU: Added OPEN bit to config register. Added description of unlock fields. 15. TWIS: SCR is Write-only. Improved explanation of slave transmitter mode. Updated data transfer diagrams. 16. Electrical Characteristics: Added more values. Added notes on simulated and characterized values. Added pin capacitance, rise, and fall times. Added timing characteristics. Removed all TBDs. Added ADC analog input characteristics. Symbol cleanup. 17. Errata: Updated errata list. Rev. D - 06/2010 1. 36.7 36.8 Ordering Information: Ordering code for TQFP ES changed from AT32UC3L064-AUES to AT32UC3L064-AUTES. TLLGA48 Tray option added. Rev. C - 06/2010 1. Features and Description: Added QTouch library support. 2. USART: Description of unimplemented features removed. 3. Electrical Characteristics: Power Consumption numbers updated. Flash timing numbers added. Rev. B - 05/2010 1. Package and Pinout: Added pinout figure for TLLGA48 package. 2. Package and Pinout, GPIO function multiplexing:TWIMS0-TWCK on PA20 removed. ADCIFBAD[3] on PA17 removed, number of ADC channels are 8, not 9. 3. I/O Lines Considerations: Added: Following pins have high-drive capability: PA02, PA06, PA08, PA09, and PB01. Some TWI0 pins are SMBUS compliant (PA21, PB04, PB05). 4. HMATRIX Masters: PDCA is master 4, not master 3. SAU is master 3, not master 4. 5. SAU: IDLE bit added in the Status Register. 6. PDCA: Number of PDCA performance monitors is device dependent. 7. Peripheral Event System: Chapter updated. 8. PM: Bits in RCAUSE registers removed and renamed (JTAGHARD and AWIREHARD renamed to JTAG and AWIRE respectively, JTAG and AWIRE removed. BOD33 bit removed). 9. PM: RCAUSE.BOD33 bit removed. SM33 reset will be detected as a POR reset. 10. PM: WDT can be used as wake-up source if WDT is clocked from 32KHz oscillator. 838 32099I–01/2012 AT32UC3L016/32/64 36.9 11. PM: Entering Shutdown mode description updated. 12. SCIF: DFLL output frequency is 40-150MHz, not 20-150MHz or 30-150MHz. 13. SCIF: Temperature sensor is connected to ADC channel 9, not 7. 14. SCIF: Updated the oscillator connection figure for OSC0 15. GPIO: Removed unimplemented features (pull-down, buskeeper, drive strength, slew rate, Schmidt trigger, open drain). 16. SPI: RDR.PCS field removed (RDR[19:16]). 17. TWIS: Figures updated. 18. ADCIFB: The sample and hold time and the startup time formulas have been corrected (ADC Configuration Register). 19. ADCIFB: Updated ADC signal names. 20. ACIFB: CONFW.WEVSRC is bit 8-10, CONFW.EWEVEN is bit 11. CONF.EVENP and CONF.EVENN bits are swapped. 21. CAT: Matrix size is 16 by 8, not 18 by 8. 22. Electrical Characteristics: General update. 23. Mechanical Characteristics: Added numbers for package drawings. 24. Mechanical Characteristics: In the TQFP-48 package drawing the Lead Coplanarity is 0.102mm, not 0.080mm. 25. Ordering Information: Ordering code for TLLGA-48 package updated. Rev. A – 06/2009 1. Initial revision. 839 32099I–01/2012 AT32UC3L016/32/64 Table of Contents Features ..................................................................................................... 1 1 Description ............................................................................................... 3 2 Overview ................................................................................................... 5 3 4 5 6 7 2.1 Block Diagram ...................................................................................................5 2.2 Configuration Summary .....................................................................................6 Package and Pinout ................................................................................. 7 3.1 Package .............................................................................................................7 3.2 Peripheral Multiplexing on I/O lines ...................................................................9 3.3 Signal Descriptions ..........................................................................................13 3.4 I/O Line Considerations ...................................................................................16 Processor and Architecture .................................................................. 18 4.1 Features ..........................................................................................................18 4.2 AVR32 Architecture .........................................................................................18 4.3 The AVR32UC CPU ........................................................................................19 4.4 Programming Model ........................................................................................23 4.5 Exceptions and Interrupts ................................................................................27 Memories ................................................................................................ 32 5.1 Embedded Memories ......................................................................................32 5.2 Physical Memory Map .....................................................................................32 5.3 Peripheral Address Map ..................................................................................33 5.4 CPU Local Bus Mapping .................................................................................34 Supply and Startup Considerations ..................................................... 36 6.1 Supply Considerations .....................................................................................36 6.2 Startup Considerations ....................................................................................40 Peripheral DMA Controller (PDCA) ...................................................... 41 7.1 Features ..........................................................................................................41 7.2 Overview ..........................................................................................................41 7.3 Block Diagram .................................................................................................42 7.4 Product Dependencies ....................................................................................42 7.5 Functional Description .....................................................................................43 7.6 Performance Monitors .....................................................................................45 7.7 User Interface ..................................................................................................47 i 32099I–01/2012 AT32UC3L016/32/64 7.8 8 9 Module Configuration ......................................................................................75 Flash Controller (FLASHCDW) ............................................................. 77 8.1 Features ..........................................................................................................77 8.2 Overview ..........................................................................................................77 8.3 Product Dependencies ....................................................................................77 8.4 Functional Description .....................................................................................78 8.5 Flash Commands ............................................................................................83 8.6 General-purpose Fuse Bits ..............................................................................85 8.7 Security Bit ......................................................................................................88 8.8 User Interface ..................................................................................................89 8.9 Fuse Settings ...................................................................................................99 8.10 Bootloader Configuration ...............................................................................103 8.11 Serial Number ................................................................................................103 8.12 Module Configuration ....................................................................................103 Secure Access Unit (SAU) .................................................................. 104 9.1 Features ........................................................................................................104 9.2 Overview ........................................................................................................104 9.3 Block Diagram ...............................................................................................105 9.4 Product Dependencies ..................................................................................106 9.5 Functional Description ...................................................................................106 9.6 User Interface ................................................................................................110 9.7 Module Configuration ....................................................................................125 10 HSB Bus Matrix (HMATRIXB) .............................................................. 126 10.1 Features ........................................................................................................126 10.2 Overview ........................................................................................................126 10.3 Product Dependencies ..................................................................................126 10.4 Functional Description ...................................................................................126 10.5 User Interface ................................................................................................130 10.6 Module Configuration ....................................................................................138 11 Interrupt Controller (INTC) .................................................................. 140 11.1 Features ........................................................................................................140 11.2 Overview ........................................................................................................140 11.3 Block Diagram ...............................................................................................140 11.4 Product Dependencies ..................................................................................141 11.5 Functional Description ...................................................................................141 ii 32099I–01/2012 AT32UC3L016/32/64 11.6 User Interface ................................................................................................144 11.7 Module Configuration ....................................................................................148 12 Power Manager (PM) ............................................................................ 151 12.1 Features ........................................................................................................151 12.2 Overview ........................................................................................................151 12.3 Block Diagram ...............................................................................................152 12.4 I/O Lines Description .....................................................................................152 12.5 Product Dependencies ..................................................................................152 12.6 Functional Description ...................................................................................153 12.7 User Interface ................................................................................................162 12.8 Module Configuration ....................................................................................185 13 System Control Interface (SCIF) ......................................................... 186 13.1 Features ........................................................................................................186 13.2 Overview ........................................................................................................186 13.3 I/O Lines Description .....................................................................................186 13.4 Product Dependencies ..................................................................................186 13.5 Functional Description ...................................................................................187 13.6 User Interface ................................................................................................203 13.7 Module Configuration ....................................................................................247 14 Asynchronous Timer (AST) ................................................................ 250 14.1 Features ........................................................................................................250 14.2 Overview ........................................................................................................250 14.3 Block Diagram ...............................................................................................251 14.4 Product Dependencies ..................................................................................251 14.5 Functional Description ...................................................................................252 14.6 User Interface ................................................................................................258 14.7 Module Configuration ....................................................................................279 15 Watchdog Timer (WDT) ....................................................................... 280 15.1 Features ........................................................................................................280 15.2 Overview ........................................................................................................280 15.3 Block Diagram ...............................................................................................280 15.4 Product Dependencies ..................................................................................280 15.5 Functional Description ...................................................................................281 15.6 User Interface ................................................................................................286 15.7 Module Configuration ....................................................................................292 iii 32099I–01/2012 AT32UC3L016/32/64 16 External Interrupt Controller (EIC) ..................................................... 293 16.1 Features ........................................................................................................293 16.2 Overview ........................................................................................................293 16.3 Block Diagram ...............................................................................................293 16.4 I/O Lines Description .....................................................................................294 16.5 Product Dependencies ..................................................................................294 16.6 Functional Description ...................................................................................294 16.7 User Interface ................................................................................................298 16.8 Module Configuration ....................................................................................314 17 Frequency Meter (FREQM) .................................................................. 315 17.1 Features ........................................................................................................315 17.2 Overview ........................................................................................................315 17.3 Block Diagram ...............................................................................................315 17.4 Product Dependencies ..................................................................................315 17.5 Functional Description ...................................................................................316 17.6 User Interface ................................................................................................318 17.7 Module Configuration ....................................................................................329 18 General-Purpose Input/Output Controller (GPIO) ............................. 331 18.1 Features ........................................................................................................331 18.2 Overview ........................................................................................................331 18.3 Block Diagram ...............................................................................................331 18.4 I/O Lines Description .....................................................................................332 18.5 Product Dependencies ..................................................................................332 18.6 Functional Description ...................................................................................333 18.7 User Interface ................................................................................................338 18.8 Module Configuration ....................................................................................361 19 Universal Synchronous Asynchronous Receiver Transmitter (USART) 362 19.1 Features ........................................................................................................362 19.2 Overview ........................................................................................................362 19.3 Block Diagram ...............................................................................................363 19.4 I/O Lines Description ....................................................................................364 19.5 Product Dependencies ..................................................................................364 19.6 Functional Description ...................................................................................365 19.7 User Interface ................................................................................................391 iv 32099I–01/2012 AT32UC3L016/32/64 19.8 Module Configuration ....................................................................................413 20 Serial Peripheral Interface (SPI) ......................................................... 414 20.1 Features ........................................................................................................414 20.2 Overview ........................................................................................................414 20.3 Block Diagram ...............................................................................................415 20.4 Application Block Diagram .............................................................................415 20.5 I/O Lines Description .....................................................................................416 20.6 Product Dependencies ..................................................................................416 20.7 Functional Description ...................................................................................416 20.8 User Interface ................................................................................................427 20.9 Module Configuration ....................................................................................454 21 Two-wire Master Interface (TWIM) ...................................................... 455 21.1 Features ........................................................................................................455 21.2 Overview ........................................................................................................455 21.3 List of Abbreviations ......................................................................................456 21.4 Block Diagram ...............................................................................................456 21.5 Application Block Diagram .............................................................................457 21.6 I/O Lines Description .....................................................................................457 21.7 Product Dependencies ..................................................................................457 21.8 Functional Description ...................................................................................459 21.9 User Interface ................................................................................................471 21.10 Module Configuration ....................................................................................488 22 Two-wire Slave Interface (TWIS) ......................................................... 489 22.1 Features ........................................................................................................489 22.2 Overview ........................................................................................................489 22.3 List of Abbreviations ......................................................................................490 22.4 Block Diagram ...............................................................................................490 22.5 Application Block Diagram .............................................................................491 22.6 I/O Lines Description .....................................................................................491 22.7 Product Dependencies ..................................................................................491 22.8 Functional Description ...................................................................................492 22.9 User Interface ................................................................................................502 22.10 Module Configuration ....................................................................................518 23 Pulse Width Modulation Controller (PWMA) ..................................... 519 23.1 Features ........................................................................................................519 v 32099I–01/2012 AT32UC3L016/32/64 23.2 Overview ........................................................................................................519 23.3 Block Diagram ...............................................................................................520 23.4 I/O Lines Description .....................................................................................520 23.5 Product Dependencies ..................................................................................520 23.6 Functional Description ...................................................................................521 23.7 User Interface ................................................................................................526 23.8 Module Configuration ....................................................................................540 24 Timer/Counter (TC) .............................................................................. 541 24.1 Features ........................................................................................................541 24.2 Overview ........................................................................................................541 24.3 Block Diagram ...............................................................................................542 24.4 I/O Lines Description .....................................................................................542 24.5 Product Dependencies ..................................................................................542 24.6 Functional Description ...................................................................................543 24.7 User Interface ................................................................................................558 24.8 Module Configuration ....................................................................................581 25 Peripheral Event System ..................................................................... 583 25.1 Features ........................................................................................................583 25.2 Overview ........................................................................................................583 25.3 Peripheral Event System Block Diagram .......................................................583 25.4 Functional Description ...................................................................................583 25.5 Application Example ......................................................................................586 26 ADC Interface (ADCIFB) ...................................................................... 587 26.1 Features ........................................................................................................587 26.2 Overview ........................................................................................................587 26.3 Block Diagram ...............................................................................................588 26.4 I/O Lines Description .....................................................................................589 26.5 Product Dependencies ..................................................................................589 26.6 Functional Description ...................................................................................590 26.7 Resistive Touch Screen .................................................................................594 26.8 Operating Modes ...........................................................................................600 26.9 User Interface ................................................................................................602 26.10 Module Configuration ....................................................................................621 27 Analog Comparator Interface (ACIFB) ............................................... 622 27.1 Features ........................................................................................................622 vi 32099I–01/2012 AT32UC3L016/32/64 27.2 Overview ........................................................................................................622 27.3 Block Diagram ...............................................................................................623 27.4 I/O Lines Description .....................................................................................623 27.5 Product Dependencies ..................................................................................624 27.6 Functional Description ...................................................................................625 27.7 Peripheral Event Triggers ..............................................................................630 27.8 AC Test mode ................................................................................................630 27.9 User Interface ................................................................................................631 27.10 Module Configuration ....................................................................................645 28 Capacitive Touch Module (CAT) ......................................................... 646 28.1 Features ........................................................................................................646 28.2 Overview ........................................................................................................646 28.3 Block Diagram ...............................................................................................647 28.4 I/O Lines Description .....................................................................................648 28.5 Product Dependencies ..................................................................................648 28.6 Functional Description ...................................................................................650 28.7 User Interface ................................................................................................657 28.8 Module Configuration ....................................................................................688 29 Glue Logic Controller (GLOC) ............................................................ 689 29.1 Features ........................................................................................................689 29.2 Overview ........................................................................................................689 29.3 Block Diagram ...............................................................................................689 29.4 I/O Lines Description .....................................................................................690 29.5 Product Dependencies ..................................................................................690 29.6 Functional Description ...................................................................................690 29.7 User Interface ................................................................................................692 29.8 Module Configuration ....................................................................................697 30 aWire UART (AW) ................................................................................. 698 30.1 Features ........................................................................................................698 30.2 Overview ........................................................................................................698 30.3 Block Diagram ...............................................................................................698 30.4 I/O Lines Description .....................................................................................699 30.5 Product Dependencies ..................................................................................699 30.6 Functional Description ...................................................................................699 30.7 User Interface ................................................................................................702 vii 32099I–01/2012 AT32UC3L016/32/64 30.8 Module Configuration ....................................................................................715 31 Programming and Debugging ............................................................ 716 31.1 Overview ........................................................................................................716 31.2 Service Access Bus .......................................................................................716 31.3 On-Chip Debug ..............................................................................................719 31.4 JTAG and Boundary-scan (JTAG) .................................................................728 31.5 JTAG Instruction Summary ...........................................................................736 31.6 aWire Debug Interface (AW) .........................................................................753 32 Electrical Characteristics .................................................................... 770 32.1 Absolute Maximum Ratings* .........................................................................770 32.2 Supply Characteristics ...................................................................................770 32.3 Maximum Clock Frequencies ........................................................................771 32.4 Power Consumption ......................................................................................771 32.5 I/O Pin Characteristics ...................................................................................776 32.6 Oscillator Characteristics ...............................................................................779 32.7 Flash Characteristics .....................................................................................783 32.8 Analog Characteristics ...................................................................................784 32.9 Timing Characteristics ...................................................................................792 33 Mechanical Characteristics ................................................................. 802 33.1 Thermal Considerations ................................................................................802 33.2 Package Drawings .........................................................................................803 33.3 Soldering Profile ............................................................................................806 34 Ordering Information ........................................................................... 807 35 Errata ..................................................................................................... 808 35.1 Rev. E ............................................................................................................808 35.2 Rev. D ............................................................................................................814 35.3 Rev. C ............................................................................................................820 35.4 Rev. B ............................................................................................................820 36 Datasheet Revision History ................................................................ 836 36.1 Rev. I - 01/2012 .............................................................................................836 36.2 Rev. H - 12/2011 ...........................................................................................836 36.3 Rev. G - 06/2011 ...........................................................................................837 36.4 Rev. F- 11/2010 .............................................................................................837 36.5 Rev. E- 10/2010 .............................................................................................837 viii 32099I–01/2012 AT32UC3L016/32/64 36.6 Rev. D - 06/2010 ...........................................................................................838 36.7 Rev. C - 06/2010 ...........................................................................................838 36.8 Rev. B - 05/2010 ............................................................................................838 36.9 Rev. A – 06/2009 ...........................................................................................839 Table of Contents....................................................................................... i ix 32099I–01/2012 Atmel Corporation 2325 Orchard Parkway San Jose, CA 95131 USA Tel: (+1)(408) 441-0311 Fax: (+1)(408) 487-2600 www.atmel.com Atmel Asia Limited Unit 1-5 & 16, 19/F BEA Tower, Millennium City 5 418 Kwun Tong Road Kwun Tong, Kowloon HONG KONG Tel: (+852) 2245-6100 Fax: (+852) 2722-1369 Atmel Munich GmbH Business Campus Parkring 4 D-85748 Garching b. Munich GERMANY Tel: (+49) 89-31970-0 Fax: (+49) 89-3194621 Atmel Japan 16F, Shin Osaki Kangyo Bldg. 1-6-4 Osaka Shinagawa-ku Tokyo 104-0032 JAPAN Tel: (+81) 3-6417-0300 Fax: (+81) 3-6417-0370 © 2012 Atmel Corporation. All rights reserved. Atmel ®, logo and combinations thereof, AVR ®, picoPower ®, QTouch ®, AKS ® and others are registered trademarks or trademarks of Atmel Corporation or its subsidiaries. Other terms and product names may be trademarks of others. Disclaimer: The information in this document is provided in connection with Atmel products. No license, express or implied, by estoppel or otherwise, to any intellectual property right is granted by this document or in connection with the sale of Atmel products. EXCEPT AS SET FORTH IN THE ATMEL TERMS AND CONDITIONS OF SALES LOCATED ON THE ATMEL WEBSITE, ATMEL ASSUMES NO LIABILITY WHATSOEVER AND DISCLAIMS ANY EXPRESS, IMPLIED OR STATUTORY WARRANTY RELATING TO ITS PRODUCTS INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT. IN NO EVENT SHALL ATMEL BE LIABLE FOR ANY DIRECT, INDIRECT, CONSEQUENTIAL, PUNITIVE, SPECIAL OR INCIDENTAL DAMAGES (INCLUDING, WITHOUT LIMITATION, DAMAGES FOR LOSS AND PROFITS, BUSINESS INTERRUPTION, OR LOSS OF INFORMATION) ARISING OUT OF THE USE OR INABILITY TO USE THIS DOCUMENT, EVEN IF ATMEL HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. Atmel makes no representations or warranties with respect to the accuracy or completeness of the contents of this document and reserves the right to make changes to specifications and product descriptions at any time without notice. Atmel does not make any commitment to update the information contained herein. Unless specifically provided otherwise, Atmel products are not suitable for, and shall not be used in, automotive applications. Atmel products are not intended, authorized, or warranted for use as components in applications intended to support or sustain life. 32099I–01/2012