MT93L16

CMOS MT93L16 Low-Voltage Acoustic Echo Canceller Preliminary Information Features • • • • • • • • • • • • • Contains two echo cancellers: 112ms acoustic echo canceller + 16ms line echo canceller Works with low cost voice codec. ITU-T G.711 or signed mag µ/A-Law, or linear 2’s comp Each port may operate in different format Advanced NLP design - full duplex speech with no switched loss on audio paths Fast re-convergence time: tracks changing echo environment quickly Adaptation algorithm converges even during Double-Talk Designed for exceptional performance in high background noise environments Provides protection against narrow-band signal divergence Howling prevention stops uncontrolled oscillation in high loop gain conditions Offset nulling of all PCM channels Serial micro-controller interface ST-BUS, GCI, or variable-rate SSI PCM interfaces User gain control provided for speaker path (-24dB to +21dB in 3dB steps) DS5068 ISSUE3 July 1999 Ordering Information MT93L16AQ 36 Pin QSOP -40 °C to + 85 °C • • • • • • • AGC on speaker path Handles up to 0 dB acoustic echo return loss and 0dB line ERL Transparent data transfer and mute options 20 MHz master clock operation Low power mode during PCM Bypass Bootloadable for future factory software upgrades 2.7V to 3.6V supply voltage; 5V-tolerant inputs Applications • • • • Full duplex speaker-phone for digital telephone Echo cancellation for video conferencing Handsfree in automobile environment Full duplex speaker-phone for PC Limiter Sin MD1 NBSD S1 µ/A-Law/ Linear Offset Null + + S2 ADV NLP Program RAM S3 Program ROM Linear/ µ/A-Law Micro Interface Sout DATA1 DATA2 ACOUSTIC ECHO PATH CONTROL UNIT Adaptive Filter Line ECho Path PORT 1 PORT 2 Howling Double Talk Detector R3 R2 Adaptive Filter NBSD R1 Controller MD2 Rout Linear/ µ/A-Law Limiter SCLK -24 -> +21dB AGC User Gain ADV NLP CS + + Offset Null µ/A-Law/ Linear Rin VDD VSS RESET FORMAT ENA2 ENA1 LAW F0i BCLK/C4i MCLK Figure 1 - Functional Block Diagram 1 MT93L16 ENA1 MD1 ENA2 MD2 Rin Sin IC MCLK IC IC IC LAW FORMAT RESET NC NC SCLK CS 1 36 35 2 34 3 33 4 5 32 31 6 30 7 29 8 9 QSOP 28 27 10 26 11 25 12 24 13 23 14 22 15 21 16 20 17 19 18 IC IC IC MCLK2 NC VSS VDD2 VSS2 IC IC BCLK/C4i F0i Rout Sout VDD NC DATA1 DATA2 Preliminary Information Figure 2 - Pin Connections Pin Description Pin # 1 Name ENA1 Description SSI Enable Strobe / ST-BUS & GCI Mode for Rin/Sout (Input). This pin has dual functions depending on whether SSI or ST-BUS/GCI is selected. For SSI, this strobe must be present for frame synchronization. This is an active high channel enable strobe, 8 or 16 data bits wide, enabling serial PCM data transfer for on Rin/Sout pins. Strobe period is 125 microseconds. For ST-BUS or GCI, this pin, in conjunction with the MD1 pin, selects the proper mode for Rin/Sout pins (see ST-BUS and GCI Operation description). ST-BUS & GCI Mode for Rin/Sout (Input). When in ST-BUS or GCI operation, this pin, in conjunction with the ENA1 pin, will select the proper mode for Rin/Sout pins (see ST-BUS and GCI Operation description). Connect this pin to Vss in SSI mode. SSI Enable Strobe / ST-BUS & GCI Mode for Sin/Rout (Input).This pin has dual functions depending on whether SSI or ST-BUS/GCI is selected. For SSI, this is an active high channel enable strobe, 8 or 16 data bits wide, enabling serial PCM data transfer on Sin/Rout pins. Strobe period is 125 microseconds. For ST-BUS/GCI, this pin, in conjunction with the MD2 pin, selects the proper mode for Sin/Rout pins (see ST-BUS and GCI Operation description). ST-BUS & GCI Mode for Sin/Rout (Input).When in ST-BUS or GCI operation, this pin in conjunction with the ENA2 pin, selects the proper mode for Sin/Rout pins (see ST-BUS and GCI Operation description). Connect this pin to Vss in SSI mode. Receive PCM Signal Input (Input). 128 kbit/s to 4096 kbit/s serial PCM input stream. Data may be in either companded or 2’s complement linear format. This is the Receive Input channel from the line (or network) side. Data bits are clocked in following SSI, GCI or STBUS timing requirements. Send PCM Signal Input (Input). 128 kbit/s to 4096 kbit/s serial PCM input stream. Data may be in either companded or 2’s complement linear format. This is the Send Input channel (from the microphone). Data bits are clocked in following SSI,GCI or ST-BUS timing requirements. Internal Connection (Input): Must be tied to Vss. Master Clock (Input): Nominal 20 MHz Master Clock input (may be asynchronous relative to 8KHz frame signal.) Tie together with MCLK2 (pin 33). Internal Connection (Input): Must be tied to Vss. A/µ Law Select (Input). When low, selects µ−Law companded PCM. When high, selects ALaw companded PCM. This control is for both serial pcm ports. 2 MD1 3 ENA2 4 MD2 5 Rin 6 Sin 7 8 9,10,11 12 13 IC MCLK IC LAW FORMAT ITU-T/Sign Mag (Input). When low, selects sign-magnitude PCM code. When high, selects ITU-T (G.711) PCM code. This control is for both serial pcm ports. 2 Preliminary Information Pin Description (continued) Pin # 14 15, 16 17 18 19 Name RESET NC SCLK CS DATA2 Description MT93L16 Reset / Power-down (Input). An active low resets the device and puts the MT93L16 into a low-power stand-by mode. No Connect (Output). These pins should be left un-connected. Serial Port Synchronous Clock (Input). Data clock for the serial microport interface. Serial Port Chip Select (Input). Enables serial microport interface data transfers. Active low. Serial Data Receive (Input). In Motorola/National serial microport operation, the DATA2 pin is used for receiving data. In Intel serial microport operation, the DATA2 pin is not used and must be tied to Vss or Vdd. Serial Data Port (Bidirectional). In Motorola/National serial microport operation, the DATA1 pin is used for transmitting data. In Intel serial microport operation, the DATA1 pin is used for transmitting and receiving data. No Connect (Output). This pin should be left un-connected. Positive Power Supply (Input). Nominally 3.3 volts. Send PCM Signal Output (Output). 128 kbit/s to 4096 kbit/s serial PCM output stream. Data may be in either companded or 2’s complement linear PCM format. This is the Send Out signal after acoustic echo cancellation and non-linear processing. Data bits are clocked out following SSI, ST-BUS, or GCI timing requirements. Receive PCM Signal Output (Output). 128 kbit/s to 4096 kbit/s serial PCM output stream. Data may be in either companded or 2’s complement linear PCM format. This is the Receive out signal after line echo cancellation non-linear processing, AGC, and gain control. Data bits are clocked out following SSI, ST-BUS, or GCI timing requirements. Frame Pulse (Input). In ST-BUS (or GCI) operation, this is an active-low (or active-high) frame alignment pulse, respectively. SSI operation is enabled by connecting this pin to Vss. 20 DATA1 21 22 23 NC VDD Sout 24 Rout 25 26 F0i BCLK/C4i Bit Clock/ST-BUS Clock (Input). In SSI operation, BCLK pin is a 128 kHz to 4.096 MHz bit clock. This clock must be synchronous with ENA1, and ENA2 enable strobes. In ST-BUS or GCI operation, C4i pin must be connected to the 4.096MHz (C4) system clock. IC VSS2 VDD2 VSS NC MCLK2 IC Internal Connection (Input). Tie to Vss. Digital Ground (Input): Nominally 0 volts. Positive Power Supply (Input): Nominally 3.3 volts (tie together with VDD, pin 22). Digital Ground (Input): Nominally 0 volts (tie together with VSS2, pin 29). No Connect (Output). This pin should be left un-connected. Master Clock (Input): Nominal 20MHz master clock (tie together with MCLK, pin 8). Internal Connection (Input). Tie to Vss. 27, 28 29 30 31 32 33 34,35,36 Notes: 1. All inputs have CMOS compatible, 5V-tolerant logic levels. 2. All outputs have CMOS logic levels. Rout, Sout, and DATA1 are 5V-tolerant when tristated (to withstand other 5V drivers on a shared bus). Glossary Double-Talk Near-end Single-Talk Far-end Single-Talk ADV NLP Howling Narrowband NBSD Noise-Gating Offset Nulling Reverberation time ERL ERLE AGC Simultaneous signals present on Rin and Sin. Signals only present at Sin input. Signals only present at Rin input. Advanced Non-Linear-Processor Oscillation caused by feedback from acoustic and line echo paths Any mono or dual sinusoidal signals Narrow Band Signal Detector Audible switching of background noise Removal of DC component The time duration before an echo level decays to -60dBm Echo Return Loss Echo Return Loss Enhancement Automatic Gain Control 3 MT93L16 Functional Description The MT93L16 device contains two echo cancellers, as well as the many control functions necessary to operate the echo cancellers. One canceller is for acoustic speaker to microphone echo, and one for line echo cancellation. The MT93L16 provides clear signal transmission in both audio path directions to ensure reliable voice communication, even with low level signals. The MT93L16 does not use variable attenuators during double-talk or single-talk periods of speech, as do many other acoustic echo cancellers for speaker-phones. Instead, the MT93L16 provides high performance full-duplex operation similar to network echo cancellers, so that users experience clear speech and un-interrupted background signals during the conversation. This prevents subjective sound quality problems associated with “noise gating” or “noise contrasting”. The MT93L16 uses an advanced adaptive filter algorithm that is double-talk stable, which means that convergence takes place even while both parties are talking 1. This algorithm allows continual tracking of changes in the echo path, regardless of doubletalk, as long as a reference signal is available for the echo canceller. (1. Patent Pending) Preliminary Information • PCM encoder/decoder compatible with µ/ALaw ITU-T G.711, µ/A-Law Sign-Mag or linear 2’s complement coding. Automatic gain control on the receive speaker path. • Adaptation Speed Control The adaptation speed of the acoustic echo canceller is designed to optimize the convergence speed versus divergence caused by interfering near-end signals. Adaptation speed algorithm takes into account many different factors such as relative double-talk condition, far end signal power, echo path change, and noise levels to achieve fast convergence. Advanced Non-Linear Processor (ADV-NLP)2 (2. Patent Pending) After echo cancellation, there is likely to be residual echo which needs to be removed so that it will not be audible. The MT93L16 uses an NLP to remove low level residual echo signals which are not comprised of background noise. The operation of the NLP depends upon a dynamic activation threshold, as well as a double-talk detector which disables the NLP during double-talk periods. The MT93L16 keeps the perceived noise level constant, without the need for any variable attenuators or gain switching that causes audible “noise gating”. The noise level is constant and identical to the original background noise even when the NLP is activated. For each audio path, the NLP can be disabled by setting the NLP- bit to 1 in the LEC or AEC control registers. Narrow Band Signal Detector (NBSD) (3. Patent Pending) The echo tail cancellation capability of the acoustic echo canceller has been sized appropriately (112ms) to cancel echo in an average sized office with a reverberation time of less than 112ms. The 16ms line echo canceller is sufficient to ensure a high ERLE for most line circuits. In addition to the echo cancellers, the following functions are supported: • Control of adaptive filter convergence speed during periods of double-talk, far end singletalk, and near-end echo path changes. • Control of Non-Linear Processor thresholds for suppression of residual non-linear echo. • Howling detector to identify when instability is starting to occur, and to take action to prevent oscillation. • Narrow-Band Detector for preventing adaptive filter divergence caused by narrow-band signals • Offset Nulling filters for removal of DC components in PCM channels. • Limiters that introduce controlled saturation levels. • Serial controller interface compatible with Motorola, National and Intel microcontrollers. 3 Single or multi-frequency tones (e.g. DTMF, or signalling tones) present in the reference input of an echo canceller for a prolonged period of time may cause the adaptive filter to diverge. The Narrow Band Signal Detector (NBSD) is designed to prevent this divergence by detecting single or multi-tones of arbitrary frequency, phase, and amplitude. When narrow band signals are detected, the filter adaptation process is stopped but the echo canceller continues to cancel echo. The NBSD can be disabled by setting the NB- bit to 1 in the MC control registers. 4 Preliminary Information Howling Detector (HWLD) (4. Patent Pending) MT93L16 The AGC can be disabled by setting the AGC- bit to 1 in MC control register. Mute Function A pcm mute function is provided for independent control of the Receive and Send audio paths. Setting the MUTE_R or MUTE_S bit in the MC register, causes quiet code to be transmitted on the Rout or Sout paths respectively. Quiet code is defined according to the following table. LINEAR SIGN/ 16 bits MAGNITUDE 2’s µ-Law complement A-Law +Zero (quiet code) 0000h 80h CCITT (G.711) 4 The Howling detector is part of an Anti-Howling control, designed to prevent oscillation as a result of positive feedback in the audio paths. The HWLD can be disabled by setting the AH- bit to 1 in the (MC) control register. Offset Null Filter To ensure robust performance of the adaptive filters at all times, any DC offset that may be present on either the Rin signal or the Sin signal, is removed by highpass filters. These filters have a corner frequency placed at 40Hz. The offset null filters can be disabled by setting the HPF- bit to 1 in the LEC or AEC control registers. Limiters To prevent clipping in the echo paths, two limiters with variable thresholds are provided at the outputs. The Rout limiter threshold is in Rout Limiter Register 1 and 2. The Sout limiter threshold is in Sout Limiter Register. Both output limiters are always enabled. User Gain The user gain function provides the ability for users to adjust the audio gain in the receive path (speaker path). This gain is adjustable from -24dB to +21dB in 3dB steps. It is important to use ONLY this user gain function to adjust the speaker volume. The user gain function in the MT93L16 is optimally placed between the two echo cancellers such that no reconvergence is necessary after gain changes. The gain can be accessed through Receive Gain Control Register. AGC The AGC function is provided to limit the volume in the speaker path. The gain of the speaker path is automatically reduced during the following conditions: • When clipping of the receive signal occurs. • When initial convergence of the acoustic echo canceller detects unusually large echo return. • When howling is detected. µ-Law FFh A-Law D5h Table 1 - Quiet PCM Code Assignment Bypass Control A PCM bypass function is provided to allow transparent transmission of pcm data through the MT93L16. When the bypass function is active, pcm data passes transparently from Rin to Rout and from Sin to Sout, with bit-wise integrity preserved. When the Bypass function is selected, most internal functions are powered down to provide low power consumption. The BYPASS control bit is located in the main control MC register. Adaptation Enable/Disable Adaptation control bits are located in the AEC and LEC control registers. When the ADAPT- bit is set to 1, the adaptive filter is frozen at the current state. In this state, the device continues to cancel echo with the current echo model. When the ADAPT- bit is set to 0, the adaptive filter is continually updated. This allows the echo canceller to adapt and track changes in the echo path. This is the normal operating state. MT93L16 Throughput Delay In all modes, voice channels always have 2 frames of delay. In ST-BUS/GCI operation, the D and C channels have a delay of one frame. 5 MT93L16 Power Down / Reset Holding the RESET pin at logic low will keep the MT93L16 device in a power-down state. In this state all internal clocks are halted, and the DATA1, Sout and Rout pins are tristated. The user should hold the RESET pin low for at least 200 msec following power-up. This will insure that the device powers up in a proper state. Following any return of RESET to logic high, the user must wait for 8 complete 8 KHz frames prior to writing to the device registers. During this time, the initialization routines will execute and set the MT93L16 to default operation (program execution from ROM using default register values). Preliminary Information high) frame pulse is applied to the F0i pin, the device will assume GCI operation. If F0i is tied continuously to Vss, the device will assume SSI operation. Figures 11 to 13 show timing diagrams of these 3 PCM-interface operation conventions. ST-BUS and GCI Operation The ST-BUS PCM interface conforms to Mitel’s STBUS standard, with an active-low frame pulse. Input data is clocked in by the rising edge of the bit clock (C4i) three-quarters of the way into the bitcell, and output data bit boundaries (Rout, Sout) occur every second falling edge of the bit clock (see Figure 11.) The GCI PCM interface corresponds to the GCI standard commonly used in Europe, with an activehigh frame pulse. Input data is clocked in by the falling edge of the bit clock (C4i) three-quarters of the way into the bitcell, and output data bit boundaries (Rout, Sout) occur every second rising edge of the bit clock (see Figure 12.) Either of these interfaces (STBUS or GCI) can be used to transport 8 bit companded PCM data (using one timeslot) or 16 bit 2’s complement linear PCM data (using two timeslots). The MD1/ENA1 pins select the timeslot on the Rin/Sout port while the MD2/ENA2 pin selects the timeslot on the Sin/Rout port, as in Table 2. Figures 3 to 6 illustrate the timeslot allocation for each of these four modes. PCM Data I/O The PCM data transfer for the MT93L16 is provided through two PCM ports. One port consists of Rin and Sout pins while the second port consists of Sin and Rout pins. The data are transferred through these ports according to either ST-BUS, GCI, or SSI conventions, and the device automatically detects the correct convention. The device determines the convention by monitoring the signal applied to the F0i pin. When a valid ST-BUS (active low) frame pulse is applied to the F0i pin, the MT93L16 will assume ST-BUS operation. When a valid GCI (active C4i start of frame (stbus & GCI) F0i (ST-BUS) 0 1 2 3 4 B F0i (GCI) PORT1 Rin 76 543210 EC Sout PORT2 Sin 76 543210 76 543210 EC Rout 76 543210 outputs = High impedance inputs = don’t care In ST-BUS/GCI Mode 1, echo canceller I/O channels are assigned to ST-BUS/GCI timeslot 0. Note that the user can configure PORT1 and PORT2 into different modes. Figure 3 - ST-BUS and GCI 8-Bit Companded PCM I/O on Timeslot 0 (Mode 1) 6 Preliminary Information MT93L16 C4i start of frame (stbus & GCI) F0i (ST-BUS) 0 1 2 3 4 B F0i (GCI) PORT1 Rin 76 543210 EC Sout 76 543210 PORT2 Sin 76 543210 EC Rout outputs = High impedance inputs = don’t care 76 543210 In ST-BUS/GCI Mode 2, echo canceller I/O channels are assigned to ST-BUS/GCI timeslot 2. Note that the user can configure PORT1 and PORT2 into different modes. Figure 4 - ST-BUS and GCI 8-Bit Companded PCM I/O on Timeslot 2 (Mode 2) C4i start of frame (stbus & GCI) F0i (ST-BUS) 0 1 2 3 4 D F0i (GCI) PORT1 Rin C B 765432107654321076543210 EC Sout 765432107654321076543210 PORT2 Sin 765432107654321076543210 EC Rout 765432107654321076543210 outputs = High impedance inputs = don’t care indicates that an input channel is bypassed to an output channel ST-BUS/GCI Mode 3 supports connection to 2B+D devices where timeslots 0 and 1 transport D and C channels and echo canceller (EC) I/O channels are assigned to ST-BUS timeslot 2 (B). Both PORT1 and PORT2 must be configured in Mode 3. Figure 5 - ST-BUS and GCI 8-Bit Companded PCM I/O with D and C channels (Mode 3) 7 MT93L16 C4i start of frame (stbus & GCI) F0i (stbus) Preliminary Information F0i (GCI) Rin PORT1 Sout S 1413 12 11 10 9 8 7 6 5 4 3 2 1 0 EC S 1413 12 11 10 9 8 7 6 5 4 3 2 1 0 Sin PORT2 Rout S 1413 12 11 10 9 8 7 6 5 4 3 2 1 0 EC S 1413 12 11 10 9 8 7 6 5 4 3 2 1 0 outputs = High impedance inputs = don’t care ST-BUS/GCI Mode 4 allows 16 bit 2’s complement linear data to be transferred using ST-BUS/GCI I/O timing. Note that PORT1 and PORT2 need not necessarily both be in mode 4. Figure 6 - ST-BUS and GCI 16-Bit 2’s complement linear PCM I/O (Mode 4) In SSI operation, the frame boundary is determined by the rising edge of the ENA1 enable strobe (see Figure 7). The other enable strobe (ENA2) is used for parsing input/output data and it must pulse within 125 microseconds of the rising edge of ENA1. In SSI operation, the enable strobes may be a mixed combination of 8 or 16 BCLK cycles allowing the flexibility to mix 2’s complement linear data on one port (e.g., Rin/Sout) with companded data on the other port (e.g., Sin/Rout). Enable Strobe Pin 1 1 ENA1 ENA2 Designated PCM I/O Port Line Side Echo Path (PORT 1) Acoustic Side Echo Path (PORT 2) PORT1 Rin/Sout Enable Pins MD1 ENA1 ST-BUS/GCI Mode Selection PORT2 Sin/Rout Enable Pins MD2 ENA2 0 0 1 0 1 0 Mode 1. 8 bit companded PCM I/O on timeslot 0 Mode 2. 8 bit companded PCM I/O on timeslot 2. Mode 3. 8 bit companded PCM I/O on timeslot 2. Includes D & C channel bypass in timeslots 0 & 1. Mode 4. 16 bit 2’s complement linear PCM I/O on timeslots 0 & 1. 0 0 1 0 1 0 1 1 Table 2 - ST-BUS & GCI Mode Select SSI Operation Table 3 - SSI Enable Strobe Pins PCM Law and Format Control (LAW, FORMAT) The SSI PCM interface consists of data input pins (Rin, Sin), data output pins (Sout, Rout), a variable rate bit clock (BCLK), and two enable pins (ENA1, ENA2) to provide strobes for data transfers. The active high enable may be either 8 or 16 BCLK cycles in duration. Automatic detection of the data type (8 bit companded or 16 bit 2’s complement linear) is accomplished internally. The data type cannot change dynamically from one frame to the next. The PCM companding/coding law used by the MT93L16 is controlled through the LAW and FORMAT pins. ITU-T G.711 companding curves for µ-Law and A-Law are selected by the LAW pin. PCM coding ITU-T G.711 and Sign-Magnitude are selected by the FORMAT pin. See Table 4. 8 Preliminary Information MT93L16 BCLK start of frame (SSI) PORT1 ENA1 Rin 8 or 16 bits EC Sout 8 or 16 bits PORT2 ENA2 Sin 8 or 16 bits EC Rout outputs = High impedance inputs = don’t care Note that the two ports are independent so that, for example, PORT1 can operate with 8-bit enable strobes and PORT2 can operate with 16-bit enable strobes. 8 or 16 bits Figure 7 - SSI Operation Sign-Magnitude FORMAT=0 ITU-T (G.711) FORMAT=1 µ-LAW LAW = 0 1000 0000 1111 1111 0111 1111 0000 0000 A-LAW LAW =1 1010 1010 1101 0101 0101 0101 0010 1010 Bit Clock (BCLK/ C4i ) The BCLK/C4i pin is used to clock the PCM data for GCI and ST-BUS (C4i) interfaces, as well as for the SSI (BCLK) interface. In SSI operation, the bit rate is determined by the BCLK frequency. This input must contain either eight or sixteen clock cycles within the valid enable strobe window. BCLK may be any rate between 128 KHz to 4.096 MHz and can be discontinuous outside of the enable strobe windows defined by ENA1, ENA2 pins. Incoming PCM data (Rin, Sin) are sampled on the falling edge of BCLK while outgoing PCM data (Sout, Rout) are clocked out on the rising edge of BCLK. See Figure 13. In ST-BUS and GCI operation, connect the system C4 (4.096MHz) clock to the C4i pin. Master Clock (MCLK) A nominal 20MHz, continuously-running master clock (MCLK) is required. MCLK may be asynchronous with the 8KHz frame. 9 PCM Code µ/A-LAW LAW = 0 or 1 + Full Scale + Zero - Zero - Full Scale 1111 1111 1000 0000 0000 0000 0111 1111 Table 4 - Companded PCM Linear PCM The 16-bit 2’s complement PCM linear coding permits a dynamic range beyond that which is specified in ITU-T G.711 for companded PCM. The echo-cancellation algorithm will accept 16 bits 2’s complement linear code which gives a maximum signal level of +15dBm0. MT93L16 Microport The serial microport provides access to all MT93L16 internal read and write registers, plus write-only access to the bootloadable program RAM (see next section for bootload description.) This microport is compatible with Intel MCS-51 (mode 0), Motorola SPI (CPOL=0, CPHA=0), and National Semiconductor Microwire specifications. The microport consists of a transmit/receive data pin (DATA1), a receive data pin (DATA2), a chip select pin (CS) and a synchronous data clock pin (SCLK). The MT93L16 automatically adjusts its internal timing and pin configuration to conform to Intel or Motorola/National requirements. The microport dynamically senses the state of the SCLK pin each time CS pin becomes active (i.e. high to low transition). If SCLK pin is high during CS activation, then Intel mode 0 timing is assumed. In this case DATA1 pin is defined as a bi-directional (transmit/ receive) serial port and DATA2 is internally disconnected. If SCLK is low during CS activation, then Motorola/National timing is assumed and DATA1 is defined as the data transmit pin while DATA2 becomes the data receive pin. The MT93L16 supports Motorola half-duplex processor mode (CPOL=0 and CPHA=0). This means that during a write to the MT93L16, by the Motorola processor, output data from the DATA1 pin must be ignored. This also means that input data on the DATA2 pin is ignored by the MT93L16 during a valid read by the Motorola processor. All data transfers through the microport are two bytes long. This requires the transmission of a Command/ Address byte followed by the data byte to be written to or read from the addressed register. CS must remain low for the duration of this two-byte transfer. As shown in Figures 8 and 9, the falling edge of CS indicates to the MT93L16 that a microport transfer is about to begin. The first 8 clock cycles of SCLK after the falling edge of CS are always used to receive the Command/Address byte from the microcontroller. The Command/Address byte contains information detailing whether the second byte transfer will be a read or a write operation and at what address. The next 8 clock cycles are used to transfer the data byte between the MT93L16 and the microcontroller. At the end of the two-byte transfer, CS is brought high again to terminate the session. The rising edge of CS will tri-state the DATA1 pin. The DATA1 pin will remain tristated as long as CS is high. Intel processors utilize Least Significant Bit (LSB) first transmission while Motorola/National processors use Most Significant Bit (MSB) first transmission. Preliminary Information The MT93L16 microport automatically accommodates these two schemes for normal data bytes. However, to ensure timely decoding of the R/W and address information, the Command/ Address byte is defined differently for Intel and Motorola/National operations. Refer to the relative timing diagrams of Figure 8 and Figure 9. Receive data bits are sampled on the rising edge of SCLK while transmit data is clocked out on the falling edge of SCLK. Detailed microport timing is shown in Figure 14 and Figure 15. Bootload Process and Execution from RAM A bootloadable program RAM (BRAM) is available on the MT93L16 to support factory-issued software upgrades to the built-in algorithm. To make use of this bootload feature, users must include 4096 X 8bits of memory in their microcontroller system (i.e. external to the MT93L16), from which the MT93L16 can be bootloaded. Registers and program data are loaded into the MT93L16 in the same fashion via the serial microport. Both employ the same command / address / data byte specification described in the previous section on serial microport. Either intel or motorola mode may be transparently used for bootloading. There are also two registers relevant to bootloading (BRC=control and SIG=signature, see Register Summary). The effect of these register values on device operation is summarized in Table 5. Bootload mode is entered and exited by writing to the bootload bit in the Bootload RAM Control (BRC) register at address 3fh (see Register Summary). During bootload mode, any serial microport "write" (R/W command bit =0) to an address other than that of the BRC register will contribute to filling the program BRAM. Call these transactions "BRAM-fill" writes. Although a command/address byte must still precede each data byte (as described for the serial microport), the values of the address fields for these "BRAM-fill" writes are ignored (except for the value 3fh, which designates the BRC register.) Instead, addresses are internally generated by the MT93L16 for each "BRAM-fill" write. Address generation for "BRAM-fill" writes resumes where it left off following any read transaction while bootload mode is enabled. The first 4096 such "BRAM-fill" writes while bootload is enabled will load the memory, but further ones after that are ignored. Following the write of the first 4096 bytes, the program BRAM will be filled. Before bootload mode is disabled, it is recommended that users then read back the value from the signature register (SIG) and compare it to the one supplied by the factory along with the code. Equality verifies that the correct data has been loaded. The signature calculation uses an 8-bit MISR which only incorporates input from "BRAM-fill" 10 Preliminary Information FUNCTIONAL DESCRIPTION FOR USING THE BOOTABLE RAM MT93L16 BOOTLOAD MODE - Microport Access is to bootload RAM (BRAM) R/W BRC Register Bits W W R R Address 3fh (= 1 1 1 1 1 1 b) other than 3fh 1x xxxxb 0x xxxxb Data Writes "data" to BRC reg. - Bootload frozen; BRAM contents are NOT affected. Writes "data" to next byte in BRAM (bootloading.) Reads back "data" = BRC reg value. - Bootload frozen; BRAM contents are NOT affected. Reads back "data" = SIG reg value. - Bootload frozen; BRAM contents are NOT affected. C3C2C1C0 X100 NON-BOOTLOAD MODE - Microport Access is to device registers (DREGs) BRC Register Bits R/W W R Address any (= a5 a4 a3 a2 a1 a0 b) any (= a5 a4 a3 a2 a1 a0 b) Data Writes "data" to corresponding DREG. Reads back "data" = corresponding DREG value. C3C2C1C0 X000 PROGRAM EXECUTION MODES C3C2C1C0 0000 Execute program in ROM, bootload mode disabled. - BRAM address counter reset to initial (ready) state. - SIG reg reseeded to initial (ready) state Execute program in ROM, while bootloading the RAM. - BRAM address counter increments on microport writes (except to 3fh) - SIG reg recalculates signature on microport writes (except to 3fh) Execute program in RAM, bootload mode disabled. - BRAM address counter reset to initial (ready) state. - SIG reg reseeded to initial (ready) state - NOT RECOMMENDED (Execute program in RAM, while bootloading the RAM) C3C2C1C0 0100 C3C2C1C0 1000 C3C2C1C0 1100 Table 5 - Bootload RAM Control (BRC) Register States Note: bits C1 C0 are reserved, and must be set to zero. writes. Resetting the bootload bit (C2) in the BRC register to 0 (see Register Summary) exits bootload mode, resetting the signature (SIG) register and internal address generator for the next bootload. A hardware reset (RESET=0) similarly returns the MT93L16 to the ready state for the start of a bootload. Once the program has been loaded, to begin execution from RAM, bootload mode must be disabled (BOOT bit, C2=0) and execution from RAM enabled (RAM_ROMb bit, C3=1) by setting the appropriate bits in the BRC register. During the bootload process, however, ROM program execution (RAM_ROMb bit, C3=0) should be selected. See Table 5 for the effect of the BRC register settings on Microport accesses and on program execution. Following program loading and enabling of execution from RAM, it is recommended that users set the software reset bit in the Main Control (MC) register, to ensure that the device updates the default register values to those of the new program in RAM. Note: it is important to use a software reset rather than a hardware (RESET=0) reset, as the latter will return the device to its default settings (which includes execution from program ROM instead of RAM.) To verify which code revision is currently running, users can access the Firmware Revision Code (FRC) register (see Register Summary). This register reflects the identity code (revision number) of the last program to run register initialization (which follows a software or hardware reset.) 11 MT93L16 Preliminary Information COMMAND/ADDRESS  DATA INPUT/OUTPUT DATA 1 R/W A0 A1 A2 A3 A4 A5 X D0 D1 D2 D3 D4 D5 D6 D7 SCLK  Œ CS Ž  Œ  Ž   This delay is due to internal processor timing and is equal to Tsch time. The delay is transparent to MT93L16. The MT93L16: latches receive data on the rising edge of SCLK outputs transmit data on the falling edge of SCLK The falling edge of CS indicates that a COMMAND/ADDRESS byte will be transmitted from the microprocessor. The subsequent byte is always data followed by CS returning high. A new COMMAND/ADDRESS byte may be loaded only by CS cycling high then low again. The COMMAND/ADDRESS byte contains: 1 bit - Read/Write 6 bits - Addressing Data 1 bit - Unused Figure 8 - Serial Microport Timing for Intel Mode 0 COMMAND/ADDRESS  DATA INPUT DATA 2 Receive R/W A5 A4 A3 A2 A1 A0 X D7 D6 D5 D4 D3 D2 D1 D0 DATA OUTPUT DATA 1 Transmit High Impedance D7 D6 D5 D4 D3 D2 D1 D0 SCLK  Œ CS Ž Œ  Ž    This delay is due to internal processor timing and is equal to Tsch time. The delay is transparent to MT93L16. The MT93L16: latches receive data on the rising edge of SCLK outputs transmit data on the falling edge of SCLK The falling edge of CS indicates that a COMMAND/ADDRESS byte will be transmitted from the microprocessor. The subsequent byte is always data followed by CS returning high. A new COMMAND/ADDRESS byte may be loaded only by CS cycling high then low again. The COMMAND/ADDRESS byte contains: 1 bit - Read/Write 6 bits - Addressing Data 1 bit - Unused Figure 9 - Serial Microport Timing for Motorola Mode 00 or National Microwire 12 Preliminary Information Absolute Maximum Ratings* Parameter 1 2 3 4 5 Supply Voltage Input Voltage Output Voltage Swing Continuous Current on any digital pin Storage Temperature Symbol VDD-VSS Vi Vo Ii/o TST -65 Min -0.5 VSS-0.3 VSS-0.3 Max 5.0 5.5 5.5 ±20 150 MT93L16 Units V V V mA °C mW . 6 Package Power Dissipation PD 90 (typ) * Exceeding these values may cause permanent damage. Functional operation under these conditions is not implied. Recommended Operating Conditions - Voltages are with respect to ground (VSS) unless otherwise stated Characteristics 1 2 3 4 Supply Voltage Input High Voltage Input Low Voltage Operating Temperature TA Sym VDD Min 2.7 1.4 VSS -40 Typ 3.3 Max 3.6 VDD 0.4 +85 Units V V V °C Test Conditions Echo Return Limits Characteristics 1 2 Acoustic Echo Return Line Echo Return Min Typ Max 0 0 Units dB dB Test Conditions Measured from Rout -> Sin Measured from Sout -> Rin DC Electrical Characteristics*- Voltages are with respect to ground (VSS) unless otherwise stated. Characteristics Standby Supply Current: 1 2 3 4 5 6 7 8 Operating Supply Current: Input HIGH voltage Input LOW voltage Input leakage current High level output voltage Low level output voltage High impedance leakage Output capacitance Sym ICC IDD VIH VIL IIH/IIL VOH VOL IOZ Co 1 10 0.8VDD 0.4VDD 10 0.1 0.7VDD 0.3VDD 10 Min Typ‡ 3 20 Max 70 Units µA mA V V µA V V µA pF VIN=VSS to VDD IOH=2.5mA IOL=5.0mA VIN=VSS to VDD Conditions/Notes RESET = 0 RESET = 1, clocks active 9 Input capacitance Ci 8 pF ‡ Typical figures are at 25°C and are for design aid only: not guaranteed and not subject to production testing. *DC Electrical Characteristics are over recommended temperature and supply voltage. 13 MT93L16 otherwise stated Preliminary Information AC Electrical Characteristics† - Serial Data Interfaces - Voltages are with respect to ground (VSS) unless Characteristics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 MCLK Frequency BCLK/C4i Clock High BCLK/C4i Clock Low BCLK/C4i Period SSI Enable Strobe to Data Delay (first bit) SSI Data Output Delay (excluding first bit) SSI Output Active to High Impedance SSI Enable Strobe Signal Setup SSI Enable Strobe Signal Hold SSI Data Input Setup SSI Data Input Hold ST-BUS/GCI F0i Setup ST-BUS/GCI F0i Hold ST-BUS/GCI Data Output delay ST-BUS/GCI Output Active to High Impedance ST-BUS/GCI Data Input Hold time Sym fCLK tBCH, tC4H tBLL, tC4L tBCP tSD tDD tAHZ tSSS tSSH tDIS tDIH tF0iS tF0iH tDSD tASHZ tDSH Min 19.15 90 90 240 80 80 80 10 15 10 15 20 20 80 80 20 150 150 tBCP -15 tBCP -10 7900 Typ Max 20.5 Units MHz ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns CL=150pF CL=150pF CL=150pF CL=150pF CL=150pF Test Notes 17 ST-BUS/GCI Data Input Setup time tDSS 20 † Timing is over recommended temperature and power supply voltages. 14 Preliminary Information AC Electrical Characteristics† - Microport Timing Characteristics 1 2 3 4 5 6 7 8 9 10 Input Data Setup Input Data Hold Output Data Delay Serial Clock Period SCLK Pulse Width High SCLK Pulse Width Low CS Setup-Intel CS Setup-Motorola CS Hold CS to Output High Impedance Sym tIDS tIDH tODD tSCP tSCH tSCL tCSSI tCSSM tCSH tOHZ Min 30 30 100 500 250 250 200 100 100 100 Typ Max Units ns ns ns ns ns ns ns ns ns ns MT93L16 Test Notes CL=150pF CL=150pF † Timing is over recommended temperature range and recommended power supply voltages. Characteristic CMOS reference level Input HIGH level Input LOW level Rise/Fall HIGH measurement point Rise/Fall LOW measurement point Symbol VCT VH VL VHM VLM CMOS Level 0.5*VDD 0.9*VDD 0.1*VDD 0.7*VDD 0.3*VDD Units V V V V V Table 8 - Reference Level Definition for Timing Measurements T=1/fCLK MCLK (I) VH VCT VL Notes: O. CMOS output I. CMOS input (5V tolerant) (see Table 8 for symbol definitions) Figure 10 - Master Clock - MCLK 15 MT93L16 Bit 7 Sout/Rout (O) tDSD C4i (I) VH Preliminary Information Bit 6 VCT tC4H tASHZ VCT VL tF0iS tF0iH F0i (I) VH VL input sampled tC4L VCT start of frame tDSS tDSH VCT Rin/Sin (I) VH VL Bit 7 ) Bit 7 Sout/Rout (O) tDSD C4i (I) VH input sampled Bit 6 Figure 11 -GCI Data Port Timing Bit 6 VCT tC4H tASHZ VL VCT tC4L tF0iS tF0iH F0i (I) VH VL start of frame VCT tDSS tDSH Rin/Sin (I) VH VCT VL Bit 7 Bit 6 Figure 12 - ST-BUS Data Port Timing Bit 7 Sout/Rout (O) tSD BCLK (I) VH Bit 6 Bit 5 VCT tDD tBCH tAHZ VCT tSSS tBCP input sampled VL tBCL tSSH ENA1 (I) or ENA2 (I) VH VL VCT tDIS tDIH start of frame VH Rin/Sin (1) VCT VL Bit 7 Notes: O. CMOS output Bit 6 Bit 5 (see Table 8 for symbol definitions) I. CMOS input (5V tolerant) Figure 13 - SSI Data Port Timing 16 Preliminary Information MT93L16 DATA INPUT DATA1 (I,O) tIDS tIDH SCLK (I) VH DATA OUTPUT VCT tSCH tODD tOHZ VCT VL tCSSI CS (I) VH tSCL tSCP tCSH VCT VL Notes: O. CMOS output I. CMOS input (5V tolerant) (see Table 8 for symbol definitions) Figure 14 - INTEL Serial Microport Timing DATA2 (I) (Input) VH VL VCT tIDS tIDH tSCH tSCP SCLK (I) VH VCT VL tCSSM CS (I) VH tSCL tCSH VCT VL tODD DATA1 (O) (Output) tOHZ VCT Notes: O. CMOS output I. CMOS input (5V tolerant) (see Table 8 for symbol definitions) Figure 15 - Motorola Serial Microport Timing 17 MT93L16 Register Summary Address: 00h R/W Power Up Reset 00h RESET AHAGCNBBYPASS MUTE_S MUTE_R LIMIT Preliminary Information Main Control Register (MC) 7 MSB LIMIT 6 MUTE_R 5 MUTE_S 4 BYPASS 3 NB- 2 AGC- 1 AH- 0 RESET LSB When high, the power initialization routine is executed presetting all registers to default values. This bit automatically clears itself to’0’ when reset is complete. When high, the Howling detector is disabled and when low the Howling detector is enabled. When high, AGC is disabled and when low AGC is enabled. When high, Narrowband signal detectors in Rin and Sin paths are disabled and when low the signal detectors are enabled When high, the Send and Receive paths are transparently by-passed from input to output and when low the Send and Receive paths are not bypassed When high, the Sin path is muted to quite code (after the NLP) and when low the Sin path is not muted When high, the Rin path is muted to quite code (after the NLP) and when low the Rin path is not muted When high, the 2-bit shift mode is enabled in conjunction with bit 7 of LEC register and when low 2-bit shift mode is disabled Address: 21h R/W Power Up Reset 00h ECBY ADAPTHCLR HPFINJNLPASCP- Acoustic Echo Canceller Control Register (AEC) 7 MSB P- 6 ASC- 5 NLP- 4 INJ- 3 HPF- 2 HCLR 1 ADAPT- 0 ECBY LSB When high, the Echo estimate from the filter is not subtracted from the input (Sin), when low the estimate is subtracted When high, the Echo canceller adaptation is disabled and when low the adaptation is enabled When high, Adaptive filter coefficients are cleared and when low the filter coefficients are not cleared When high, Offset nulling filter is bypassed in the Sin/Sout path and when low the Offset nulling filter in not bypassed When high, the Noise filtering process is disabled in the NLP and when low the Noise filtering process is enabled When high, the Non Linear Processor is disabled in the Sin/Sout path and when low the NLP is enabled When high, the Internal Adaptation speed control is disabled and when low the Adaptation speed is enabled When high, the Exponential weighting function for the adaptive filter is disabled and when low the weighting function is enabled Address: 01h R/W Power Up Reset 00h ECBY ADAPTHCLR HPFINJNLPASCSHFT Line Echo Canceller Control Register (LEC) MSB 7 SHFT 6 ASC- 5 NLP- 4 INJ- 3 HPF- 2 HCLR 1 ADAPT- 0 ECBY LSB When high, the Echo estimate from the filter is not substracted from the input (Rin), when low the estimate is substracted When high, the Echo canceller adaptation is disabled and when low the adaptation is enabled When high, Adaptive filter coefficients are cleared and when low the filter coefficients are not cleared When high, Offset nulling filter is bypassed in the Rin/Rout path and when low the Offset nulling filter in not bypassed When high, the Noise filtering process is disabled in the NLP and when low the Noise filtering process is enabled When high, the Non Linear Processor is disabled in the Rin/Rout path and when low the NLP is enabled When high, the Internal Adaptation speed control is disabled and when low the Adaptation speed is enabled when high the 16-bit linear mode, inputs Sin, Rin, are shift right by 2 and outputs Sout, Rout are shift left by 2. This bit is ignored when 16-bit linear mode is not selected in both ports. This bit is also ignored if bit 7 of MC register is set to zero 18 Preliminary Information Address: 22h Read Power Up Reset 00h NBS NB DT NLPDC HWLNG ACMUND - MT93L16 Acoustic Echo Canceller Status Register (ASR) (* Do not write to this register) 7 MSB - 6 ACMUND 5 HWLNG 4 - 3 NLPDC 2 DT 1 NB 0 NBS LSB When high, the Narrowband signal has been detected in the Sin/Sout path and when low, the Narrowband signal has not been detected in the Sin/Sout path LOGICAL OR of the status bit NBS + NBR from LSR Register When high the Double Talk is detected and when low, the Double talk is not detected When high, the NLP is activated and when low the NLP is not activated RESERVED. When high, Howling is occurring in the loop and when low, no Howling is detected When high, No active signal in the Rin/Rout path RESERVED. Address: 02h Read Power Up Reset 00h NBR NB DT NLPC -RESERVED. . - Line Echo Canceller Status Register (LSR) (* Do not write to this register) 6 - 5 - 4 - 3 NLPC 2 DT 1 NB 0 NBR LSB When high, a narrowband signal has been detected in the Receive (Rin) path. When low no narrowband signal is not detected in the Rin path This bit indicates a LOGICAL-OR of Status bits NBR + NBS (from ASR Register) When high, double-talk is detected and when low double-talk is not detected When high, NLP is activated and when low NLP is not activated Address: 20h R/W Power Up Reset 6Dh G0 G1 G2 G3 RESERVED Receive Gain Control Register (RGC) 7 MSB - 6 - 5 - 4 - 3 G3 2 G2 1 G1 0 GO LSB User Gain Control on the Rin/Rout path (Tolerance of gains: +/- 0.15 dB). The hexadecimal number represents G3 to G0 value in the table below. Gain Values for Receive Gain Control Register Bit G3 to G0 (RGC) 0h 1h 2h 3h -24dB -21dB -18dB -15dB 4h 5h 6h 7h -12dB -9 dB -6 dB -3 dB 8h 9h Ah Bh 0 dB + 3 dB + 6 dB +9 dB Ch Dh Eh Fh +12 dB + 15 dB + 18 dB + 21 dB 19 MT93L16 Address: 16h Read Power Up Reset 00h RIPD0 RIPD1 RIPD2 RIPD3 RIPD4 RIPD5 RIPD6 RIPD7 Address: 17h Read Power Up Reset 00h RIPD8 RIPD9 RIPD10 RIPD11 RIPD12 RIPD13 RIPD14 RIPD15 Address: 18h Read Power Up Reset 00h REPD0 REPD1 REPD2 REPD3 REPD4 REPD5 REPD6 REPD7 Address: 19h Read Power Up Reset 00h REPD8 REPD9 REPD10 REPD11 REPD12 REPD13 REPD14 REPD15 See above description See Above Description Preliminary Information Receive (Rin) Peak Detect Register 1 MSB (RIPD1) 1 RIPD1 7 RIPD 7 6 RIPD 6 5 RIPD5 4 RIPD4 3 RIPD3 2 RIPD2 0 RIPD0 LSB These peak detector registers allow the user to monitor the receive in signal (Rin) peak level at reference point R1 (see Figure #1). The information is in 16-bit 2’s complement linear coded format presented in two 8 bit registers. The high byte is in Register 2 and the low byte is in Register 1. Receive (Rin) Peak Detect Register 2 MSB (RIPD2) 1 RIPD9 7 RIPD 15 6 RIPD 5 14 RIPD13 4 RIPD12 3 RIPD11 2 RIPD10 0 RIPD8 LSB 7 MSB Receive (Rin) ERROR Peak Detect Register 1 MSB (REPD1) 1 REPD1 7 REPD 7 6 REPD 6 5 REPD 5 4 REPD 4 3 REPD 3 2 REPD 2 0 REPD 0 LSB These peak detector registers allow the user to monitor the error signal peak level at reference point R2 (see Figure #1). The information is in 16-bit 2’s complement linear coded format presented in two 8 bit registers. The high byte is in Register 2 and the low byte is in Register 1. Receive (Rin) ERROR Peak Detect Register 2 MSB (REPD2) 1 REPD 9 0 REPD8 LSB 7 REPD 6 REPD14 15 5 REPD13 4 REPD 3 REPD11 12 2 REPD 10 20 Preliminary Information MT93L16 Receive (Rout) Peak Detect Register 1 Address: 3Ah Read Power Up Reset 00h ROPD0 ROPD1 ROPD2 ROPD3 ROPD4 ROPD5 ROPD6 ROPD7 Address: 3Bh Read Power Up Reset 00h ROPD8 ROPD9 ROPD10 ROPD11 ROPD12 ROPD13 ROPD14 ROPD15 Address: 36h Read Power Up Reset 00h SIPD0 SIPD1 SIPD2 SIPD3 SIPD4 SIPD5 SIPD6 SIPD7 Address: 37h Read Power Up Reset 00h SIPD8 SIPD9 SIPD10 SIPD11 SIPD12 SIPD13 SIPD14 SIPD15 See above description MSB MSB (ROPD1) 1 ROPD1 7 ROPD 7 6 ROPD 6 5 ROPD 5 4 ROPD4 3 ROPD3 2 ROPD2 0 ROPD 0 LSB These peak detector registers allow the user to monitor the receive out signal (Rout) peak level at reference point R3 (see Figure #1). The information is in 16-bit 2’s complement linear coded format presented in two 8 bit registers. The high byte is in Register 2 and the low byte is in Register 1. Receive (Rout) Peak Detect Register 2 MSB (ROPD2) 1 ROPD 9 0 ROPD8 LSB 7 ROPD 6 ROPD 5 ROPD13 15 14 4 ROPD 3 ROPD 2 ROPD 11 12 10 See Above description Send (Sin) Peak Detect Register 1 MSB (SIPD1) SIPD2 7 SIPD 7 6 SIPD 6 5 SIPD5 4 SIPD4 3 SIPD3 2 1 SIPD1 0 SIPD0 LSB These peak detector registers allow the user to monitor the receive in signal (Sin) peak level at reference point S1 (see Figure #1). The information is in 16-bit 2’s complement linear coded format presented in two 8 bit registers. The high byte is in Register 2 and the low byte is in Register 1. Send (Sin) Peak Detect Register 2 7 SIPD 15 6 SIPD 14 5 SIPD13 (SIPD2) SIPD10 4 SIPD12 3 SIPD11 2 1 SIPD9 0 SIPD8 LSB 21 MT93L16 Address: 38h Read Power Up Reset 00h SEPD0 SEPD1 SEPD2 SEPD3 SEPD4 SEPD5 SEPD6 SEPD7 Address: 39h Read Power Up Reset 00h SEPD8 SEPD9 SEPD10 SEPD11 SEPD12 SEPD13 SEPD14 SEPD15 Address: 1Ah Read Power Up Reset 00h SOPD0 SOPD1 SOPD2 SOPD3 SOPD4 SOPD5 SOPD6 SOPD7 Address: 1Bh Read Power Up Reset 00h SOPD8 SOPD9 SOPD10 SOPD11 SOPD12 SOPD13 SOPD14 SOPD15 See Above description MSB MSB Preliminary Information Send ERROR Peak Detect Register 1 7 SEPD 7 6 SEPD 6 5 SEPD5 (SEPD1) 1 SEPD1 4 SEPD4 3 SEPD3 2 SEPD2 0 SEPD0 LSB These peak detector registers allow the user to monitor the error signal peak level in the send path at reference point S2 (see Figure #1). The information is in 16-bit 2’s complement linear coded format presented in two 8 bit registers. The high byte is in Register 2 and the low byte is in Register 1. Send ERROR Peak Detect Register 2 MSB (SEPD2) 1 SEPD 9 0 SEPD8 LSB 7 SEPD 6 SEPD 15 14 5 SEPD13 4 SEPD 3 SEPD11 12 2 SEPD 10 See Above description Send (Sout) Peak Detect Register 1 7 SOPD 7 6 SOPD 6 5 SOPD 5 4 SOPD 4 3 SOPD 3 2 (SOPD1) 1 SOPD1 SOPD2 0 SOPD 0 LSB These peak detector registers allow the user to monitor the Send out signal (Sout) peak level at reference point S3 (see Figure #1). The information is in 16-bit 2’s complement linear coded format presented in two 8 bit registers. The high byte is in Register 2 and the low byte is in Register 1. Send (Sout) Peak Detect Register 2 MSB (SOPD2) 1 SOPD 9 0 SOPD8 LSB 7 SOPD 6 SOPD 15 14 5 SOPD13 4 SOPD 3 SOPD11 12 2 SOPD 10 22 Preliminary Information MT93L16 (A_AS1) 0 A_AS 0 LSB Address: 3Ch R/W Power Up Reset 00h A_AS0 A_AS1 A_AS2 A_AS3 A_AS4 A_AS5 A_AS6 A_AS7 Address: 3Dh R/W Power Up Reset 10h A_AS8 A_AS9 A_AS10 A_AS11 A_AS12 A_AS13 A_AS14 A_AS15 Address: 1Ch R/W Power Up Reset 00h L_AS0 L_AS1 L_AS2 L_AS3 L_AS4 L_AS5 L_AS6 L_AS7 Address: 1Dh Read Power Up Reset 08h L_AS8 L_AS9 L_AS10 L_AS11 L_AS12 L_AS13 L_AS14 L_AS15 MSB Acoustic Echo Canceller Adaptation Speed Register 1 MSB 7 A_AS 7 6 A_AS 6 5 A_AS 5 4 A_AS 4 3 A_AS 3 2 A_AS2 1 A_AS1 This register allows the user to program control the adaptation speed of the Acoustic Echo Canceller. This register value changes dynamically when the ’ASC-’ bit in the Acoustic Echo Canceller Control Register is low. The ’ASC-’ bit must be 1 when this register is under user control. The valid range is from 0000h to 7FFFh. The high byte is in Register 2 and the low byte is in Register 1. Smaller values correspond to slower adaptation speed. Acoustic Echo Canceller Adaptation Speed Register 2 MSB (A_AS2) 0 A_AS8 LSB 7 A_AS 6 A_AS 15 14 5 A_AS13 4 A_AS 12 3 A_AS11 2 A_AS 10 1 A_AS 9 See Above description Line Echo Canceller Adaptation Speed Register 1 MSB (L_AS1) L_AS1 7 L_AS 7 6 L_AS 6 5 L_AS 5 4 L_AS 4 3 L_AS 3 2 L_AS2 1 0 L_AS 0 LSB This register allows the user to program control the adaptation speed of the Line Echo Canceller. This register value changes dynamically when the ’ASC-’ bit in the Acoustic Echo Canceller Control Register is low. The ’ASC-’ bit must be 1 when this register is under user control. The valid range is from 0000h to 7FFFh. The high byte is in Register 2 and the low byte is in Register 1. Smaller values correspond to slower adaptation speed. Line Echo Canceller Adaptation Speed Register 2 7 L_AS 6 L_AS 15 14 5 L_AS13 (L_AS2) 0 L_AS8 LSB 4 L_AS 12 3 L_AS11 2 L_AS 10 1 L_AS 9 See Above description 23 MT93L16 Address: 24h R/W Power Up Reset 80h L0 This bit is used in conjunction with Rout Limiter Register 2. (See description below.) RESERVED Preliminary Information Rout Limiter Register 1 (RL1) 7 MSB L0 6 - 5 - 4 - 3 - 2 - 1 - 0 LSB Address: 25h R/W Power Up Reset 3Eh L1 L2 L3 L4 L5 L6 L7 L8 Rout Limiter Register 2 (RL2) 7 MSB L8 6 L7 5 L6 4 L5 3 L4 2 L3 1 L2 0 L1 LSB In conjunction with bit 7 (L0) of the above (RL1) register, this register (RL2) allows the user to program the output Limiter threshold value in the Rout path. Default value is (1f40)h which is equal to 3.14dBmo Maximum value is (7FC0 )h = 15 dBmo Minimum value is (0040)h = -38 dBmo Address: 26h R/W Power Up Reset 3Dh L0 L1 L2 L3 L4 RESERVED Sout Limiter Register (SL) 7 MSB L4 6 L3 5 L2 4 L1 3 L0 2 - 1 - 0 LSB This register allows the user to program the output Limiter threshold value in the Rout path Default value is (1f40)h which is equal to 3.14dBmo Maximum value is (7F40 )h 24 Preliminary Information Address: 03h Read Power Up Reset 00h - MT93L16 Firmware Revision Code Register (FRC) 7 MSB FRC2 6 FRC1 5 FRC0 4 - 3 - 2 - 1 - 0 - LSB RESERVED FRC0 FRC1 FRC2 Revision code of the firmware program currently being run (default=rom=00). Address: 3fh R / W Power Up Reset 00h C0 C1 C2 C3 - Bootload RAM Control Register (BRC) 7 MSB - 6 - 5 - 4 - 3 RAM_ROMb 2 BOOT 1 - 0 - LSB RESERVED. Must be set to zero. RESERVED. Must be set to zero. BOOT bit. When high, puts device in bootload mode. When low, bootload is disabled. RAM_ROMb bit. When high, device executes from RAM. When low, device executes from ROM. RESERVED Address: 07h Read Power Up Reset FFh SIG7 SIG6 SIG5 SIG4 SIG3 SIG2 SIG1 SIG0 Bootload RAM Signature Register (SIG) MSB 7 SIG 7 6 SIG 6 5 SIG5 4 SIG 4 3 SIG3 2 SIG 2 1 SIG 1 0 SIG0 LSB This register provides the signature of the bootloaded data to verify error-free delivery into the device. Note: this register is only accessible if BOOT bit is high (bootload mode enabled) in the above BRC register. While bootload is disabled, the register value is held constant at its reset seed value of FFh. 25 Package Outlines D e ZD R E H A A1 Pin #1 70 B 0.51 x 45° (.020) ±.008 (.014) 0.335 7° ±0.20 0.63 ±0.10 (.025) ±.004 GAGE PLANE C DETAIL - A Notes: 1. Lead Coplanarity should be 0 to 0.10mm (.004") max 2. Package surface finishing (2.1) Top Matte: (Charmilles #18-30) (2.2) All Sides: (Charmilles #18-30) (2.3) Bottom Matte: (Charmilles #18-30) 3. All dimensions excluding mold flashes 4. Max. deviation of center of package and center of leadrame to be 0.10mm (.004") 5. Max. misalignment between top and bottom center of package to 0.10mm (.004") 6. End flash from the package body shall not exceed 0.152 (.006") per side (D) 7. Dimension B shall not include dambar protrusion/intrusion and solder coverage. 8. Not to scale 9. Dimension in inches 10.Dimensions in (millimeters) L Q A QSOP - Quad Shrink Outline Package 36-Pin Dim Dim 36-Pin Min e H L Q R ZD Min A A1 B C D E .096 (2.44) .004 (0.10) .011 (0.28) .0091 (0.23) .598 (15.20) .291 (7.40) Max .104 (2.64) .012 (0.30) .020 (0.51) .0125 (0.32) .606 (15.40) .299 (7.60) Max .0315 inches (ref) 0.80mm .398 (10.11) 0.16 (0.40) 0° .025 (0.63) .414 (10.51) .050 (1.27) 8° .035 (0.89) .0335 inches (ref) 0.85 http://www.mitelsemi.com World Headquarters - Canada Tel: +1 (613) 592 2122 Fax: +1 (613) 592 6909 North America Tel: +1 (770) 486 0194 Fax: +1 (770) 631 8213 Asia/Pacific Tel: +65 333 6193 Fax: +65 333 6192 Europe, Middle East, and Africa (EMEA) Tel: +44 (0) 1793 518528 Fax: +44 (0) 1793 518581 Information relating to products and services furnished herein by Mitel Corporation or its subsidiaries (collectively “Mitel”) is believed to be reliable. However, Mitel assumes no liability for errors that may appear in this publication, or for liability otherwise arising from the application or use of any such information, product or service or for any infringement of patents or other intellectual property rights owned by third parties which may result from such application or use. Neither the supply of such information or purchase of product or service conveys any license, either express or implied, under patents or other intellectual property rights owned by Mitel or licensed from third parties by Mitel, whatsoever. Purchasers of products are also hereby notified that the use of product in certain ways or in combination with Mitel, or non-Mitel furnished goods or services may infringe patents or other intellectual property rights owned by Mitel. This publication is issued to provide information only and (unless agreed by Mitel in writing) may not be used, applied or reproduced for any purpose nor form part of any order or contract nor to be regarded as a representation relating to the products or services concerned. The products, their specifications, services and other information appearing in this publication are subject to change by Mitel without notice. No warranty or guarantee express or implied is made regarding the capability, performance or suitability of any product or service. Information concerning possible methods of use is provided as a guide only and does not constitute any guarantee that such methods of use will be satisfactory in a specific piece of equipment. It is the user’s responsibility to fully determine the performance and suitability of any equipment using such information and to ensure that any publication or data used is up to date and has not been superseded. Manufacturing does not necessarily include testing of all functions or parameters. These products are not suitable for use in any medical products whose failure to perform may result in significant injury or death to the user. All products and materials are sold and services provided subject to Mitel’s conditions of sale which are available on request. M Mitel (design) and ST-BUS are registered trademarks of MITEL Corporation Mitel Semiconductor is an ISO 9001 Registered Company Copyright 1999 MITEL Corporation All Rights Reserved Printed in CANADA TECHNICAL DOCUMENTATION - NOT FOR RESALE