FM3808-70-T

Preliminary FM3808 256Kb Bytewide FRAM w/ Real-Time Clock Features 256K bit Ferroelectric Nonvolatile RAM • Organized as 32,752 x 8 bits • High Endurance 100 Billion (1011) Read/Writes • 10 year Data Retention • NoDelay™ Writes • 70 ns Access Time/ 130 ns Cycle Time • Built-in Low VDD Protection Real-Time Clock/Calendar Function • Clock Registers in Top 16 bytes of Address Space • Backup Power from External Capacitor or Battery • Tracks Seconds through Centuries in BCD Format • Tracks Leap Years through 2099 • Runs from a 32.768 kHz Timekeeping Crystal System Supervisor Function • Programmable Clock/Calendar Alarm • Programmable Watchdog Timer • Power Supply Monitor • Interrupt Output - Programmable active high/low • Control Settings Inherently Nonvolatile • Generates either Processor Reset or Interrupt Low Power Operation • 5V Operation for Memory and Clock Interface • Backup Voltage as low as 2.5V • 25 mA IDD Active Current • 1 µA IBAK Clock Backup Current Description The FM3808 combines a 256Kb FRAM array with a real-time clock and a system supervisor function. An external 32.768 kHz crystal drives the timekeeping function. It maintains time and date settings in the absence of system power through the user’s choice of backup power source – either a capacitor or a battery. In either case data in the memory array does not depend on the backup source, it remains nonvolatile in FRAM. In addition to timekeeping, the FM3808 includes a system supervisor to manage low VDD power conditions and a watchdog timer function. A programmable interrupt output pin allows the user to select the supervisor functions and the polarity of the signal. Both the FRAM array and the timekeeping function are accessed through the memory interface. The upper 16-address locations of the memory space are allocated to the timekeeping registers rather than to memory. The FRAM array provides data retention for 10 years in the absence of system power, and is not dependent on the backup power source for the clock. This eliminates system concerns over data loss in a traditional battery-backed RAM solution. In addition, clock and supervisor control settings are implemented in FRAM rather than battery-backed RAM, making them more dependable. The FM3808 offers guaranteed operation over an industrial temperature range of -40°C to +85°C. Pin Configuration A11 A9 A8 A13 WE VBAK INT VDD X1 X2 A14 A12 A7 A6 A5 A4 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 OE A10 CE DQ7 DQ6 DQ5 DQ4 DQ3 VSS DQ2 DQ1 DQ0 A0 A1 A2 A3 FM3808-70-T FM3808DK Ordering Information 70 ns access, 32-pin TSOP DIP module development kit Documentation for the DIP module development kit is available separately. This is a product in sampling or pre-production phase of development. Characteristic data and other specifications are subject to change without notice. Rev 1.1 May 2003 Ramtron International Corporation 1850 Ramtron Drive, Colorado Springs, CO 80921 (800) 545-FRAM, (719) 481-7000, FAX (719) 481-7058 www.ramtron.com Page 1 of 28 FM3808 Switched power System Supervisor Low VDD monitor/ Watchdog timer Watchdog timebase FRAM Array 32,752x8 VDD VBAK INT Interrupt Control Logic Data Alarm Address Decoder/ Bus Interface Address CE OE WE 32.768 kHz X1 Clock/Calendar 16 Clock/Calendar Registers X2 Figure 1. Block Diagram Pin Description Pin Name I/O A0-A14 Input DQ(7:0) /CE I/O Input /OE /WE X1, X2 INT Input Input Input Output VBAK Supply VDD VSS Rev 1.1 May 2003 Supply Supply Pin Description Address: The 15 address inputs select one of 32,752 bytes in the FRAM array or one of 16 bytes in the clock/calendar. The address is latched on the falling edge of /CE. Data: Bi-directional 8-bit data bus for accessing the FRAM array and clock. Chip Enable: The active low /CE input selects the device. The falling edge of /CE internally latches the address. Address changes that occur after /CE has transitioned low are ignored until the next falling edge occurs. Output Enable: The active low /OE input enables the data output buffers during read cycles. Deasserting /OE high causes the DQ pins to tri-state. Write Enable: The active low /WE low enables data on the DQ pins to be written to the address location latched by the falling edge of /CE. Connect 32.768 kHz crystal. Interrupt output: This output can be programmed to respond to the clock/calendar alarm, the watchdog timer, and the power monitor. It is programmable to either active high (push/pull) or active low (open-drain). Backup Supply Voltage: This supply is used to maintain power for the clock. It must remain between 2.5V and VDD-0.3V. VBAK is typically supplied by either a capacitor or a battery. Current is drawn from VBAK when VDD is below the VBAK voltage. Supply Voltage: 5V Ground. Page 2 of 28 FM3808 Functional Truth Table /CE /WE H X ! X L H L L /OE X X L X Function Standby/Precharge Latch Address Read Write combined memory map. The register map is described below, followed by a detailed description of each functional block. Overview The FM3808 integrates three complementary but distinct functions under a common interface in a single package. First, is the 32Kx8 FRAM memory block (minus 16 bytes), second is the real-time clock/calendar, and third is the system supervisor. The functions are integrated to enhance their individual performance, so that each provides better capability than three similar stand-alone devices. All functions use the same bytewide address/data interface and are memory mapped. Special functions, including the clock/calendar and supervisor system, are controlled via registers that reside in the top of the Register Map The top 16 FRAM address locations control the clock/calendar, alarm, and supervisor functions. The registers contain timekeeping data, control bits, or information flags. A short description of each register follows. Detailed descriptions of each function follow the register summary. Register Map Summary Table Data D6 D5 D4 D3 10 years 0 0 0 10 mo 0 0 10 date 0 0 0 0 0 0 0 10 hours 0 10 minutes 0 10 seconds /OSCEN reserved reserved CALS CAL3 WDS /WDW WDT5 WDT4 WDT3 WIE AIE PFE ABE H/L /Match 0 Alarm 10 date /Match 0 Alarm 10 hours /Match Alarm 10 minutes /Match Alarm 10 seconds D7 WDF AF PF CF TST Address 7FFFh 7FFEh 7FFDh 7FFCh 7FFBh 7FFAh 7FF9h 7FF8h 7FF7h 7FF6h 7FF5h 7FF4h 7FF3h 7FF2h 7FF1h 7FF0h D1 D0 Function years Years months Month date Date day Day hours Hours minutes Minutes seconds Seconds CAL2 CAL1 CAL0 Control-NV WDT2 WDT1 WDT0 Watchdog P/L reserved reserved Interrupts Alarm date Alarm Date hours Alarm Hours Alarm minutes Alarm Minutes Alarm seconds Alarm Seconds User-NV CAL W R Flags/Control D2 Range 00-99 1-12 1-31 1-7 0-23 0-59 0-59 1-31 0-23 0-59 0-59 Note that the shaded register bits are implemented in FRAM, therefore data at these locations is retained even without backup power. Rev 1.1 May 2003 Page 3 of 28 FM3808 Table 1. Register Map Address 7FFFh Description Timekeeping – Years D7 10 year.3 D6 10 year.2 D5 10 year.1 D4 10 year.0 D3 Year.3 D2 Year.2 D1 Year.1 D0 Year.0 Contains the lower two BCD digits of the year. Lower nibble contains the value for years; upper nibble contains the value for 10s of years. Each nibble operates from 0 to 9. The range for the register is 0-99. 7FFEh Timekeeping – Months D7 0 D6 0 D5 0 D4 10 Month D3 Month.3 D2 Month.2 D1 Month.1 D0 Month.0 Contains the BCD digits for the month. Lower nibble contains the lower digit and operates from 0 to 9; upper nibble (one bit) contains the upper digit and operates from 0 to 1. The range for the register is 112. 7FFDh Timekeeping – Date of the month D7 0 D6 0 D5 10 date.1 D4 10 date.0 D3 Date.3 D2 Date.2 D1 Date.1 D0 Date.0 Contains the BCD digits for the date of the month. Lower nibble contains the lower digit and operates from 0 to 9; upper nibble contains the upper digit and operates from 0 to 3. The range for the register is 1-31. 7FFCh Timekeeping – Day of the week D7 0 D6 0 D5 0 D4 0 D3 0 D2 Day.2 D1 Day.1 D0 Day.0 Lower nibble contains a value that correlates to day of the week. Day of the week is a ring counter that counts from 1 to 7 then returns to 1. The user must assign meaning to the day value, as the day is not integrated with the date. 7FFBh Timekeeping – Hours D7 0 D6 0 D5 10 hours.1 D4 10 hours.0 D3 Hours.3 D2 Hours2 D1 Hours.1 D0 Hours.0 Contains the BCD value of hours in 24-hour format. Lower nibble contains the lower digit and operates from 0 to 9; upper nibble (two bits) contains the upper digit and operates from 0 to 2. The range for the register is 0-23. 7FFAh Timekeeping – Minutes D7 0 D6 10 min.2 D5 10 min.1 D4 10 min.0 D3 Min.3 D2 Min.2 D1 Min.1 D0 Min.0 Contains the BCD value of minutes. Lower nibble contains the lower digit and operates from 0 to 9; upper nibble contains the upper minutes digit and operates from 0 to 5. The range for the register is 0-59. 7FF9h Timekeeping – Seconds D7 0 D6 10 sec.2 D5 10 sec.1 D4 10 sec.0 D3 Seconds.3 D2 Seconds.2 D1 Seconds.1 D0 Seconds.0 Contains the BCD value of seconds. Lower nibble contains the lower digit and operates from 0 to 9; upper nibble contains the upper digit and operates from 0 to 5. The range for the register is 0-59. Rev 1.1 May 2003 Page 4 of 28 FM3808 Address 7FF8h Description Control-Nonvolatile D7 OSCEN D6 Reserved D5 Reserved D4 CALS D3 CAL.3 D2 CAL.2 D1 CAL.1 D0 CAL.0 /OSCEN Reserved CALS CAL.3-0 /Oscillator Enable. When set to 1, the oscillator is halted. When set to 0, the oscillator runs. Disabling the oscillator saves battery power during storage. On a no-battery power up, this bit is set to 1. The RTC will not run until the oscillator is enabled. Set this bit to 0 to activate the RTC. Do not use. Should remain set to 0. Calibration sign. Determines if the calibration adjustment is applied as an addition to or as a subtraction from the time-base. This bit is implemented in FRAM. Calibration is explained below These four bits control the calibration of the clock. These bits are implemented in FRAM. 7FF7h Watchdog Timer D7 WDS D6 WDW D5 WDT.5 D4 WDT.4 D3 WDT.3 D2 WDT.2 D1 WDT.1 D0 WDT.0 WDS /WDW WDT.5-0 Watchdog Strobe. Setting this bit to 1 reloads and restarts the watchdog timer. Setting the bit to 0 has no affect. The bit is cleared automatically once the watchdog timer is reset. The WDS bit is write only. Reading it always will return a 0. Watchdog Write Enable. Setting this bit to 1 masks the watchdog timeout value (WDT.5-0) so it cannot be written. This allows the user to strobe the watchdog without disturbing the timeout value. Setting this bit to 0 allows bits 5-0 to be written on the next write to the Watchdog register. The new value will be loaded on the next internal watchdog clock after the write cycle is complete. This function is explained in more detail in the watchdog Timer section below. Watchdog Timeout selection. The watchdog timer interval is selected by the 6-bit value in this register. It represents a multiplier of the 32 Hz count (31.25 ms). The minimum range or timeout value is 31.25 ms (a setting of 1) and the maximum timeout is 2 seconds (setting of 3Fh). Setting the watchdog timer register to 0 disables the timer. These bits can be written only if the /WDW bit was cleared to 0 on a previous cycle. 7FF6h Interrupts D7 WIE D6 AIE D5 PFE D4 ABE D3 H/L D2 P/L D1 Reserved D0 Reserved WIE AIE PFE ABE H/L P/L Watchdog Interrupt Enable. When set to 1 and a watchdog timeout occurs, the watchdog timer drives the INT pin as well as the WDF flag. When set to 0, the watchdog timeout affects only the WDF flag. Alarm Interrupt Enable. When set to 1, the alarm match drives the INT pin as well as the AF flag. When set to 0, the alarm match only affects the AF flag. Power-Fail Enable. When set to 1, the power-fail monitor drives the pin as well as the PF flag. When set to 0, the power-fail monitor affects only the PF flag. Alarm Battery-backup Enable. When set to 1, the alarm interrupt (as controlled by AIE) will function even in battery backup mode. When set to 0, the alarm will occur only when VDD>VLO. High/Low. When set to a 1, the INT pin is push/pull active high. When set to a 0, the INT pin is open drain, active low. Pulse/Level. When set to a 1, the INT pin is driven active (determined by H/L) by an interrupt source for approximately 200 ms. When set to a 0, the INT pin is driven to an active level (as set by H/L) until the Flags/Control register is read. 7FF5h Alarm – Date of the month D7 M D6 0 D5 10 date.1 D4 10 date.0 D3 Date.3 D2 Date.2 D1 Date.1 D0 Date.0 /M Contains the alarm value for the date of the month and the mask bit to select or deselect the date value. Match. Setting this bit to 0 causes the date value to be used in the alarm match. Setting this bit to 1 causes the match circuit to ignore the date value. Rev 1.1 May 2003 Page 5 of 28 FM3808 Address 7FF4h Description Alarm – Hours D7 M D6 0 D5 10 hours.1 D4 10 hours.0 D3 Hours.3 D2 Hours2 D1 Hours.1 D0 Hours.0 /M Contains the alarm value for the hours and the mask bit to select or deselect the hours value. Match: Setting this bit to 0 causes the Hours value to be used in the alarm match. Setting this bit to 1 causes the match circuit to ignore the Hours value. 7FF3h Alarm – Minutes D7 M D6 10 min.2 D5 10 min.1 D4 10 min.0 D3 Min.3 D2 Min.2 D1 Min.1 D0 Min.0 /M Contains the alarm value for the minutes and the mask bit to select or deselect the minutes value Match: Setting this bit to 0 causes the Minutes value to be used in the alarm match. Setting this bit to 1 causes the match circuit to ignore the Minutes value. 7FF2h Alarm – Seconds D7 M D6 10 sec.2 D5 10 sec.1 D4 10 sec.0 D3 Seconds.3 D2 Seconds.2 D1 Seconds.1 D0 Seconds.0 /M Contains the alarm value for the seconds and the mask bit to select or deselect the minutes value. Match: Setting this bit to 0 causes the Seconds value to be used in the alarm match. Setting this bit to1 causes the match circuit to ignore the Seconds value. 7FF1h User-Nonvolatile D7 D6 D5 D4 D3 D2 D1 D0 This register is an uncommitted nonvolatile register. 7FF0h Flags/Control D7 WDF D6 AF D5 PF D4 CF D3 TST D2 CAL D1 W D0 R WDF AF PF CF TST CAL W R Watchdog Timer Flag. This read-only bit is set to 1 when the watchdog timer is allowed to reach 0 without being reset by the user. It is cleared to 0 when the Flags/Control register is read. Alarm Flag. This read-only bit is set to 1 when the time and date match the values stored in the alarm registers with the match bit(s) = 0. It is cleared when the Flags/Control register is read. Power-fail Flag. This read-only bit is set to 1 when power falls below the power-fail interrupt threshold VTP. It is cleared to 0 when the Flags/Control register is read. Century Overflow Flag. This read-only bit is set to a 1 when the values in the years register overflows from 99 to 00. This indicates a new century, such as going from 1999 to 2000 or 2099 to 2100. The user should record the new century information as needed. This bit is cleared to 0 when the Flags/Control register is read. Invokes factory test mode. Users should always set this bit to 0. Calibration Mode. When set to 1, the clock enters calibration mode. When CAL is set to 0, the clock operates normally. Write Time. Setting the W bit to 1 freezes updates of the timekeeping registers. The user can then write them with updated values. Setting the W bit to 0 causes the contents of the time registers to be transferred to the timekeeping counters. This bit affects registers 7FF9h - 7FFFh. Read Time. Setting the R bit to 1 copies a static image of the timekeeping registers and places them in a holding register. The user can then read them without concerns over changing values causing system errors. The R bit going from 0 to 1 causes the timekeeping capture, so the bit must be returned to 0 prior to reading again. This bit affects registers 7FF9h - 7FFFh. Rev 1.1 May 2003 Page 6 of 28 FM3808 Real-Time Clock Operation The real-time clock (RTC) consists of an oscillator, a clock divider, and a register system to access the information. It divides down the 32.768 kHz timebase to provide the user timekeeping resolution of one second (1Hz). The RTC will not run until the oscillator is enabled. The ocillator enable bit is bit 7 of register 7FF8h and is automatically set to a one (disabled) when the device powers up without a backup supply. Static registers provide the user with read/write access to the time values. The synchronization of these registers with the timekeeper core is performed using R and W bits in register 7FF0h. Setting the R bit from 0 to 1 causes a transfer of the timekeeping information to holding registers that can be read by the user. If a timekeeper update is in progress when the R is set, the update will be completed prior to loading the registers. Another update cannot be performed unless the R bit is first cleared to 0 again. Setting the W bit causes the timekeeper to freeze updates. Clearing it to 0 causes the values in the time registers to be written into the timekeeper core. Users should be sure not to load invalid values, such as FFh to any of the timekeeping registers. Updates to the timekeeping core occur continuously except when frozen. A diagram of the timekeeping core is shown below. Backup Power The real-time clock/calendar is intended for permanently powered operation. When primary system power fails, the VDD voltage will drop. When it crosses the voltage on the VBAK supply pin, the clock/calendar power will switch to the backup power supply VBAK. The supervisor function, described below, controls the switchover process as part of a more complete power management circuit. The clock uses very little current, which maximizes battery life. Although a backup battery may be used with the FM3808, the key advantage to combining a clock function with FRAM is that the configuration data (shaded regsiter bits in Table 2) is nonvolatile and does not require a battery backup power source. Therefore, it is more practical to use a capacitor as a backup energy source than a battery-backed RAM/clock combo. With the FM3808, the user has the choice of using a battery or a capacitor as the backup source. Some of the considerations in the backup source decision include: The expected duration of power outages, the difficulty of resetting the time if lost, and the cost tradeoff of using a small battery versus a capacitor. The following functions are powered from the backup power source when VDD < VBAK (backup mode) : • Clock/calendar core • Alarm interrupt/comparator • INT pin (determined by ABE & AIE bits); active low only • Flags connected to related functions The following functions are not powered and are disabled when VDD < VLO : • User interface • Watchdog timer • Power monitor & band-gap (VDD < ≈ 2.0V) • Flags connected to related functions • All FRAM access & updates • Calibration operation • INT pin if programmed as active-high Rev 1.1 May 2003 Page 7 of 28 FM3808 32 Hz 512 Hz W Oscillator 32.768 kHz crystal Years 8 bits Months 5 bits Date 6 bits Hours 6 bits Clock Divider 1 Hz Update Logic CF Minutes 7 bits Seconds 7 bits Days 3 bits User Interface Registers R Figure 2. Real-Time Clock Core Block Diagram ± 4.34 ppm or ± 0.19 minutes per month at the calibrated temperature. The calibration setting is nonvolatile and is stored in 7FF8h (bits D4-D0). This value only can be written when the CAL bit is set to a 1. To exit calibration mode, the user must clear the CAL bit to a 0. When the calibration mode is entered, the user can measure the frequency error on the INT pin. This error expressed in ppm translates directly into timekeeping error. An offsetting calibration adjustment corrects this error. However, the correction is applied by adding or removing pulses on a periodic basis. Therefore, the correction will not appear on the 512 Hz output. The calibration correction must be applied using the values shown in Table 2. The timekeeping accuracy can be verified by comparing the FM3808 time to a reference source. Calibration When the CAL bit in register 7FF0h is set to 1, the clock enters calibration mode. Interrupts are disabled in CAL mode. Calibration operates by applying a digital correction to the counter based on the frequency error. In CAL mode, the INT pin is driven with a 512 Hz nominal square wave. Any measured deviation from 512 Hz is converted into an error in ppm. This error corresponds to a correction value that must then be written by the user into the calibration register 7FF8h. The correction factors are listed in the Table 2. Positive ppm errors require a negative adjustment that removes pulses. Negative ppm errors require a positive correction that adds pulses. Positive ppm adjustments have the CALS bit set to 1, where as negative ppm adjustments have CALS = 0. After calibration, the clock will have a maximum error of Rev 1.1 May 2003 Page 8 of 28 FM3808 Table 2. Calibration Adjustments Measured Frequency Range Min Max 512.0000 511.9978 511.9978 511.9933 511.9933 511.9889 511.9889 511.9844 511.9844 511.9800 511.9800 511.9756 511.9756 511.9711 511.9711 511.9667 511.9667 511.9622 511.9622 511.9578 511.9578 511.9533 511.9533 511.9489 511.9489 511.9444 511.9444 511.9400 511.9400 511.9356 511.9356 511.9311 Measured Frequency Range Min Max 512.0000 512.0022 512.0022 512.0067 512.0067 512.0111 512.0111 512.0156 512.0156 512.0200 512.0200 512.0244 512.0244 512.0289 512.0289 512.0333 512.0333 512.0378 512.0378 512.0422 512.0422 512.0467 512.0467 512.0511 512.0511 512.0556 512.0556 512.0600 512.0600 512.0644 512.0644 512.0689 Error Range (ppm) Min Max 0 4.34 4.35 13.02 13.03 21.70 21.71 30.38 30.39 39.06 39.07 47.74 47.75 56.42 56.43 65.10 65.11 73.78 73.79 82.46 82.47 91.14 91.15 99.82 99.83 108.50 108.51 117.18 117.19 125.86 125.87 134.54 Error Range (ppm) Min Max 0 4.34 4.35 13.02 13.03 21.70 21.71 30.38 30.39 39.06 39.07 47.74 47.75 56.42 56.43 65.10 65.11 73.78 73.79 82.46 82.47 91.14 91.15 99.82 99.83 108.50 108.51 117.18 117.19 125.86 125.87 134.54 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Program Calibration D4-D0 00000b 10001b 10010b 10011b 10100b 10101b 10110b 10111b 11000b 11001b 11010b 11011b 11100b 11101b 11110b 11111b 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Program Calibration D4-D0 00000b 00001b 00010b 00011b 00100b 00101b 00110b 00111b 01000b 01001b 01010b 01011b 01100b 01101b 01110b 01111b Rev 1.1 May 2003 Page 9 of 28 FM3808 Supervisor Operation The Supervisor function includes a clock/calendar alarm, a watchdog timer, and a power monitor. A programmable interrupt pin (INT) provides maximum flexibility to the system designer. The INT pin is designed to allow either reset or interrupt capability to the host processor. Alarm The alarm function compares user-programmed values to the corresponding time-of-day values. When a match occurs, the alarm event occurs. The alarm drives an internal flag AF (7FF0 bit D6) and may drive the INT pin if desired. There are four alarm match fields. They are Date, Hours, Minutes, and Seconds. Each of these fields also has a Match bit that is used to determine if the field is used in the alarm match logic. Setting the Match bit to ‘0’ indicates that the corresponding field will be used in the match process. Depending on the Match bits, the alarm can occur as specifically as one particular second on one day of the month, or as frequently as once per second continuously. The MSB of each Alarm register is a Alarm Match Bit Examples Seconds 1 0 0 0 0 Minutes 1 1 0 0 0 Hours 1 1 1 0 0 Date 1 1 1 1 0 Match bit. Examples of the Match bit settings are shown in the table below. Selecting none of the match bits (all ‘1’s) indicates that no match is required. The alarm occurs every second. Setting the match select bit for seconds to ‘0’ causes the logic to match the seconds alarm value to the current time of day. Since a match will occur for only one value per minute, the alarm occurs once per minute. Likewise setting the seconds and minutes match select bits causes an exact match of these values. Thus, an alarm will occur once per hour. Setting seconds, minutes, and hours causes a match once per day. Lastly, selecting all matchvalues causes an exact time and date match. Selecting other bit combinations will not produce meaningful results, however the alarm circuit should follow the functions described. There are two ways a user can detect an alarm event, by reading the AF flag or monitoring the INT pin. The AF flag in the register 7FF0h (bit D6) will indicate that a date/time match has occurred. The AF bit will be set to 1 when a match occurs. Reading the Flags/Control register clears the alarm flag bit (and all others). A hardware interrupt pin may be used to detect an alarm event. The AIE bit in the register 7FF6h The interrupt function is described below. Alarm condition No match required = alarm 1/second Alarm when seconds match, = alarm 1/minute Alarm when seconds, minutes match, = alarm 1/hour Alarm when seconds, minutes, hours match, = alarm 1/day Alarm when seconds, minutes, hours, date match, = alarm 1/month Rev 1.1 May 2003 Page 10 of 28 FM3808 Watchdog Timer The Watchdog timer is a free running down counter that uses the 32 Hz clock (31.25 ms) derived from the crystal oscillator. The oscillator must be running (/OSCEN=0) for the watchdog to function. It begins counting down from the value loaded in the Watchdog Timer register (7FF7h). The counter consists of a loadable register and a free running counter. On power up, the watchdog timeout value in 7FF7h is loaded into the counter load register. Counting begins on power up and restarts from the loadable value any time the Watchdog Strobe WDS bit (7FF7h bit D7) is set to 1. The counter is compared to terminal value of 0. If the counter reaches this value, it causes an internal flag and an optional interrupt output (see interrupts below). The user can prevent the timeout interrupt by setting WDS bit to 1 prior to the counter reaching 0. This causes the counter to be reloaded with the watchdog timeout value and to be restarted. As long as the user sets the WDS bit prior to the counter reaching the terminal value, the interrupt and flag never occurs. New timeout values can be written by setting the watchdog write bit (7FF7h bit D6) to 0. When the /WDW bit is 0 (from a previous operation), new writes to the watchdog timeout value 7FF7h bits D5D0 allow the timeout value to be modified. When /WDW is a 1, then writes to bits 7FF7h bits D4-D0 will be ignored. The /WDW function allows a user to set the WDS bit without concern that the watchdog timer value will be modified. A logical diagram of the watchdog timer is shown below. Note that setting the watchdog timeout value to 0 would be otherwise meaningless and therefore disables the watchdog function. The output of the watchdog timer is a flag bit WDF (7FF0h bit D7) that is set if the watchdog is allowed to timeout. The flag is set upon a watchdog timeout and cleared when the Flags/Control register is read by the user. The user can also enable an optional interrupt source to drive the INT pin if the watchdog timeout occurs. The interrupt function is described on page 13. Oscillator 32.768 kHz Clock Divider 1 Hz 32 Hz Counter Zero Compare 7FF0.7 WDF WDS WDW D Q Load Register Q 7FF7.5-0 Watchdog register write to Watchdog register Figure 3. Watchdog Timer Block Diagram Rev 1.1 May 2003 Page 11 of 28 FM3808 Power Monitor The FM3808 provides a power management scheme with either power-fail interrupt or processor-reset capability. It also controls the internal switch to backup power for the timekeeper and protects the memory from low-VDD access. The power monitor is based on an internal band-gap reference circuit that compares the VDD voltage to various thresholds. The power monitor compares VDD to three thresholds. The first is an interrupt threshold (VTP). When VDD drops below the VTP level, the event will set the power fail flag PF (7FF0h bit D5). It also can drive the INT pin as described in the Interrupts section. The second threshold is the low VDD memory lockout voltage VLO. This level prevents low voltage writes to the FRAM array, which may otherwise result in corrupted data. At this point, access to the memory array and clock registers will be blocked until VDD rises above VLO. The lockout voltage VLO always trips below VTP. When VDD drops below VLO, all inputs will be ignored. On power up, the chip enable input will be ignored while VDD is below VLO, but must be pulled high prior to VDD reaching VLO. At the third threshold, the internal supply switches from VDD to VBAK for the timekeeper. This switchover will occur at the level when VDD is less than VBAK. When switchover occurs, the clock will begin to draw power from VBAK rather than VDD. This event may be above or below the VTP or VLO level depending on whether a battery or capacitor backup is used. To conserve the life of the backup source, the power monitor circuit is only operated from VDD. When VDD has dropped too low for the monitor to work, it ceases operation. However, the power monitor will reenergize as VDD rises on power up. On power-up, after the band-gap energizes, the reverse sequence will occur. As soon as the band gap is functional, it will re-assert both selections for switch over and power fail. As the VDD rises further, the device will revert to the primary power source VDD, allowing memory access and clock operation. As the VDD rises above VTP, the power-fail condition will be removed. Note that the PF flag will not be cleared until the Flags/Control register is read. The following figure illustrates the various events tracked by the power monitor. VDD VTP VLO VBAK VRST VTP VLO VBAK VRST Figure 4. Power Monitor Events In the diagram, VRST is the voltage at which an activelow interrupt will have sufficient drive strength to pull the INT pin low. Rev 1.1 May 2003 Page 12 of 28 FM3808 Interrupts The supervisor was designed to serve diverse applications. Its sophistication and programmability make the supervisor function highly configurable for the host system. The interrupt block is capable of providing system interrupt or reset conditions, and can even power up a system at a preprogrammed time. Although the INT pin is described as an interrupt, the pin may be used as a reset source as well. The supervisor provides three potential interrupt sources. They include the watchdog timer, the power monitor, and the clock/calendar alarm. Each can be individually enabled and assigned to drive the INT pin. In addition, each has an associated flag bit that the host processor can use to determine the cause of the interrupt. Some of the sources have additional control bits that determine functional behavior. In addition, the pin driver has three bits that specify its behavior when an interrupt occurs. A functional diagram of the interrupt logic is shown below. The three interrupts each have a source and an enable. Both the source and the enable must be active (true high) in order to generate an interrupt output. Only one source is necessary to drive the pin. The user can identify the source by reading the Flags/Control register, which contains the flags associated with each source. All flags are cleared to 0 when the register is read. The cycle must be a complete read cycle (/WE high), otherwise the flags will not be cleared. The power monitor has two programmable settings that are explained above in the power monitor section. Once an interrupt source is active, the pin driver determines the behavior of the output. It has two programmable settings as shown below. Pin driver control bits are located in the Interrupts register 7FF6h bits D3-D2. WDF Watchdog Timer WIE P/L PF Power Monitor VINT PFE Pin Driver H/L VDD INT ABE AF Clock Alarm AIE Figure 5. Interrupt Block Diagram According to the programming selections, the pin can be driven in the backup mode for an alarm interrupt. In addition, the pin can be an active low (open-drain) or active high (push-pull) driver. If programmed for operation during backup mode, it can only be active low. Lastly, the pin can provide a one-shot function so that the active condition is a pulse, or a level condition. In One-Shot mode, the pulse width is internally fixed at approximately 200 ms. This mode is intended to reset a host microcontroller. In Level Rev 1.1 May 2003 mode, the pin goes to its active polarity until the Flags/Control register is read by the user. This mode is intended to be used as an interrupt to a host microcontroller. The control bits are summarized as follows. Watchdog Interrupt Enable - WIE. When set to 1, the watchdog timer drives the INT pin as well as an internal flag when a watchdog timeout occurs. When Page 13 of 28 FM3808 WIE is set to 0, the watchdog timer affects only the internal flag. Alarm Interrupt Enable – AIE. When set to 1, the alarm match drives the INT pin as well as an internal flag. When set to 0, the alarm match only affects the internal flag. Power-fail Interrupt Enable - PFE. When set to 1, the power-fail monitor drives the pin as well as an internal flag. When set to 0, the power-fail monitor affects only the internal flag. Alarm Battery-backup Enable - ABE. When set to 1, the clock alarm interrupt (as controlled by AIE) will function even in battery backup mode. When set to 0, the alarm will occur only when VDD > VLO. AIE should only be set when the INT pin is programmed for active low operation. In addition, it only functions with the clock alarm, not the watchdog. If enabled, the power monitor will drive the interrupt during all normal VDD conditions regardless of the ABE bit. The application for ABE is intended for power control, where a system powers up at a predetermined time. Depending on the application, it may require dedicating the INT pin to this function. High/Low – H/L. When set to a 1, the INT pin is active high and the driver mode is push-pull. The INT pin can drive high only when VDD>VLO. When set to a 0, the INT pin is active low and the driver mode is open-drain. Active low (open drain) is operational even in battery backup mode. Pulse/Level – P/L. When set to a 1 and an interrupt occurs, the INT pin is driven for approximately 200 ms. When P/L is set to a 0, the INT pin is driven high or low (determined by H/L) until the Flags/Control register is read. When an enabled interrupt source activates the INT pin, an external host can read the Flags/Control register to determine the cause. Remember that all flags will be cleared when the register is read. If the INT pin is programmed for Level mode, then the condition will clear and the INT pin will return to its inactive state. If the pin is programmed for Pulse mode, then reading the flag also will clear the flag and the pin. The pulse will not complete its specified duration if the Flags/Control register is read. If the INT pin is used as a host reset, then the Flags/Control register cannot be read during a reset. Care should be taken in reading the flags as a new source may occur after the pin goes active but before the register is read. During a power-on reset with no battery, the interrupt register is automatically loaded with the value 24h. This causes power-fail interrupt to be enabled with an active-low pulse. Rev 1.1 May 2003 Page 14 of 28 FM3808 FRAM Memory Operation The memory array is logically organized as 32,768 x 8 with the upper 16 bytes disabled and allocated to the RTC and supervisor control settings. It is accessed using an industry standard SRAM-type parallel interface. It is virtually identical to the 32Kx8 FM1808 in function. The memory array in the FM3808 is inherently nonvolatile via its unique ferroelectric process. All data written to the part is immediately nonvolatile with no delay. Functional operation of the FRAM memory is similar to SRAM type devices. The major operating difference between the FRAM array and an SRAM (besides nonvolatile storage) is that the FM3808 latches the address on the falling edge of /CE. Users access 32,752 memory locations each with 8 data bits through a parallel interface. The complete 15-bit address specifies each of 32,768 bytes uniquely, with the upper 16 locations allocated to timekeeping functions. Internally, the memory array is organized into 32 blocks of 8Kb each. The 5 mostsignificant address lines decode one of 32 blocks. This block segmentation has no effect on operation, however the user may wish to group data into blocks by its endurance requirements as explained in a later section. The access and cycle time are the same for read and write memory operations. Writes occur immediately at the end of the access with no delay. A precharge operation, where /CE goes inactive, is a part of every memory cycle. Thus unlike SRAM, the access and cycle times are not equal. The FM3808 is designed to operate in a manner very similar to other bytewide memory products. For users familiar with BBSRAM, the performance is comparable but the bytewide interface operates in a slightly different manner as described below. For users familiar with EEPROM, the obvious differences result from the higher write-performance of FRAM technology including NoDelay writes and much higher write-endurance. Read Operation A read operation begins on the falling edge of /CE. At this time, the address bits are latched and a memory cycle is initiated. Once started, a full memory cycle will be completed internally even if the /CE is taken inactive. Data becomes available on the bus after the access time has been satisfied. After the address has been latched, the address value may be changed upon satisfying the hold time parameter. Unlike an SRAM, changing address values Rev 1.1 May 2003 will have no effect on the memory operation after the address is latched. The FM3808 will drive the data bus when /OE is asserted low. If /OE is asserted after the memory access time has been satisfied, the data bus will be driven with valid data. If /OE is asserted prior to completion of the memory access, the data bus will not be driven until valid data is available. This feature minimizes supply current in the system by eliminating transients due to invalid data. When /OE is inactive, the data bus will remain tri-stated. Write Operation Writes occur in the FM3808 in the same time interval as reads. The FM3808 supports both /CE and /WE controlled write cycles. In all cases, the address is latched on the falling edge of /CE. In a /CE controlled write, the /WE signal is asserted prior to beginning the memory cycle. That is, /WE is low when /CE falls. In this case, the part begins the memory cycle as a write. The FM3808 will not drive the data bus regardless of the state of /OE. In a /WE controlled write, the memory cycle begins on the falling edge of /CE. The /WE signal falls after the falling edge of /CE. Therefore, the memory cycle begins as a read. The data bus will be driven according to the state of /OE until /WE falls. The timing of both /CE and /WE controlled write cycles is shown in the electrical specifications. Write access to the array begins asynchronously after the memory cycle is initiated. The write access terminates on the rising edge of /WE or /CE, whichever is first. Data set-up time, as shown in the electrical specifications, indicates the interval during which data cannot change prior to the end of the write access. Unlike other truly nonvolatile memory technologies, there is no write delay with FRAM. Since the read and write access times of the underlying memory are the same, the user experiences no delay through the bus. The entire memory operation occurs in a single bus cycle. Therefore, any operation including read or write can occur immediately following a write. Data polling, a technique used with EEPROMs to determine if a write is complete, is unnecessary. Precharge Operation The precharge operation is an internal condition where the state of the memory is prepared for a new access. All memory cycles consist of a memory access and a precharge. The precharge is user initiated by taking the /CE signal high or inactive. It Page 15 of 28 FM3808 must remain high for at least the minimum precharge timing specification. The user dictates the beginning of this operation since a precharge will not begin until /CE rises. However, the device has a maximum /CE low time specification that must be satisfied. Block 31 Block 30 FFFFh FC00h FBFFh F800h F7FFh Memory Architecture FRAM memory internally operates with a read and restore mechanism. Therefore, each read and write cycle involves a change of state. The memory architecture is based on an array of rows and columns. Each access causes an endurance cycle for an entire 32-bit row (4 bytes). The memory array is divided into 32 blocks, each 1Kx8. The 5-upper address lines decode the block selection as shown in Figure 6. Data targeted for significantly different numbers of cycles should be located in separate blocks since memory rows do not extend across block boundaries. Each block of 1Kx8 consists of 256 rows and 4 column address locations. The address lines A0-A7 decode row selection and A8-A9 lines decode column selection. This scheme facilitates a relatively uniform distribution of cycles across the rows of a block. By allowing the address LSBs to decode row selection, the user avoids applying multiple cycles to the same row when accessing sequential data. For example, 256 bytes can be accessed sequentially without accessing the same row twice. In this example, one cycle would be applied to each row. An entire block of 1Kx8 can be read or written with only four cycles applied to each row. Figure 7 illustrates the organization within a memory block. Block 29 Block 28 F400h F3FFh F000h Block Block 3 Block 2 Block 1 Block 0 0FFFh 0C00h 0BFFh 0800h 07FFh 0400h 03FFh 0000h Figure 6. Address Blocks A 9- A 8 11b Row 252 Row 253 Row 254 Row 255 Block 4 A 14- A 10 00100b Row 0 Row 1 Row 2 Row 3 10b 01b 00b A0-A7 00h 01h 02h 03h FCh FDh FEh FFh Figure 7. Row and Column Organization Rev 1.1 May 2003 Page 16 of 28 FM3808 FRAM Design Considerations SRAM and FRAM alike begin each read/write cycle with a new address being driven prior to the chip enable transition low. The falling edge of chip enable latches the address and a memory access starts. For subsequent memory accesses, SRAMs allow /CE to remain low while the address bus changes. FRAM devices do not allow this signalling. Every FRAM access requires a falling edge of /CE, therefore users cannot ground this pin as you might with SRAM. Users who are modifying existing designs to use FRAM should examine the memory controller for timing compatibility of address and control pins. Each memory access must be qualified with a low transition of /CE. In many cases, this is the only change required. An example of the signal relationships is shown in Figure 8 below. Also shown is a common SRAM signal relationship that will not work for the FRAM devices. The reason for /CE to strobe for each address is twofold: it latches the new address and creates the necessary precharge period while /CE is high. Valid Memory Signaling Relationship CE FRAM signaling Address Address 1 Address 2 Data Data 1 Data 2 Invalid Memory Signaling Relationship CE SRAM signaling Address Address 1 Address 2 Data Data 1 Data 2 Figure 8. Memory Address and /CE Relationships Rev 1.1 May 2003 Page 17 of 28 FM3808 Real-time Clock Design Considerations The principal design issues in using the real time clock are selection and specification of backup energy source and the selection of the timekeeping crystal. Selection of the backup source is primarily a choice between a capacitor and a battery, and the specifications needed for each. Selection of the crystal is based on mechanical (surface mount versus through-hole) considerations and the characteristic capacitance. Each topic is discussed briefly. Backup Power Source The FM3808 is designed to accommodate either a battery or a capacitor as a backup power source. Unlike SRAM-based timekeepers that depend on the battery to make data nonvolatile, the FM3808 is unrestricted. Data stored in FRAM is not dependent on the backup battery in any way. This means that capacitor backup, which should be less expensive, is a option. Selection of a capacitor is determined by the expected duration of power outage where timekeeping must be maintained, and the practical difficulty in resetting the time should it be lost. If the time is relatively easy to reset, or a typical power loss is only a brownout, the capacitor may be a good, cost effective choice. In addition, portable systems that use a battery for primary power are good candidates for capacitor backup. If the time is very difficult to reset, or the power outage may be longer than a capacitor can supply, then a small battery is best. Each system and application can be evaluated for the difficulty in setting the time. However, the expected backup times for several capacitor choices are illustrated below. These figures cannot be used as guarantees due to the unknown leakage characteristics in external components, but they provide guidelines for realistic expectations about capacitor use. In the scenario using capacitor backup, the charging circuit must also be considered. A typical representation is shown below. The backup times are based on a starting backup voltage (fully charged) of 4.7V and minimum backup voltage of 2.5V. A 0.3V forward drop from 5.0V VDD might be expected from a schottky-diode. Note the graph, which shows approximate backup current as a function of backup voltage. Thus, the higher voltages at the beginning of discharge provide less incremental backup time than the lower voltages near the end of discharge. However, the total backup time depends on the capacitor size and the maximum, fully charged voltage. One important note about capacitor backup is that the times are incremental. Each time power is restored the capacitor is fully recharged. Rather than examining the cumulative time without power in the system over a 10-year period, the capacitor design is only concerned with the maximum time without power for one outage. If the times available for a capacitor are not sufficient, then a battery is the best selection. Most users opt for a 3V lithium coin. Note that with nonrechargeable batteries, the reservoir is not replenished so the critical parameter is the total time without power during the useful life of a system. For 1 year without power (total) during a 10-year system life, the battery capacity must be at least 9.25 mAhr. For 5 years without power during a 10-year period, it becomes 46 mA*hr. 5V Supply x1 VDD Schottky Diode Capacitor Size 100 µF 1000 µF 10000 µF 0.47 F (super cap) Backup Time 3 minutes 30 minutes 5.1 hours 240 hours x2 VBAK FM3808 Backup Capacitor or Super/Golden Cap Figure 9. Capacitor Backup Circuit Rev 1.1 May 2003 Page 18 of 28 FM3808 IBAK vs VBAK 2.5E-06 2.0E-06 IBAK 1.5E-06 1.0E-06 5.0E-07 0.0E+00 4.7 4.5 4.25 4 3.75 VBAK 3.5 3.25 3 2.75 2.5 Figure 10. Backup Current vs. Voltage Crystal Selection The second passive component needed for the RTC function is the timekeeping crystal. A 32.768 kHz time-base is required, and the FM3808 is designed to accept a low cost crystal. The major parameters associated with the crystal are timekeeping accuracy and backup current. The FM3808 is designed to accept a crystal with a characteristic capacitance of 6 pF. Deviations from this specification will lead to different accuracy and IBAK from the specified values. Though accuracy is unlikely to improve, the IBAK may go up or down from the specified value as a function of the capacitive load. The timekeeping accuracy is also a strong function of the operating temperature due to errors in crystal frequency. Temperature behavior of timekeeping crystals is well known and it follows a curve like the one shown below. The specific crystal manufacturer should be consulted for the behavior of their specific device. Note the error in frequency ppm. One ppm is roughly 2.6 seconds per month in timekeeping error. 0 -20 -40 -60 Error ppm -80 -100 -120 -140 -160 -180 -45 -35 -25 -15 -5 15 25 35 45 55 65 75 85 Page 19 of 28 5 -200 Ambient Temperature C Figure 11. Typical Crystal Error vs. Temperature Rev 1.1 May 2003 FM3808 Electrical Specifications Absolute Maximum Ratings Symbol Description VDD Power Supply Voltage with respect to VSS VIN Voltage on any signal pin with respect to VSS TSTG TLEAD Storage temperature Lead temperature (Soldering, 10 seconds) Ratings -1.0V to +7.0V -1.0V to +7.0V and VIN < VDD+1.0V -55°C to + 125°C 300° C Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only, and the functional operation of the device at these or any other conditions above those listed in the operational section of this specification is not implied. Exposure to absolute maximum ratings conditions for extended periods may affect device reliability. DC Operating Conditions (TA = -40° C to + 85° C, VDD = 4.5V to 5.5V unless otherwise specified) Symbol Parameter Min Typ Max Units Notes VDD Power Supply Voltage 4.5 5.0 5.5 V IDD VDD Supply Current – Active 10 25 mA 1 ISB1 Standby Current – TTL 500 2 µA ISB2 Standby Current – CMOS 150 3 µA VBAK Clock Backup Voltage 2.5 3.0 VDD V IBAK Clock backup current 1 4 µA VTP VDD trip point voltage that activates INT pin 4.35 4.65 V VLO VDD Lockout Voltage 4.2 4.49 V 8 VSW VDD Voltage that causes switch to VBAK VBAK V 5 VRST VDD Voltage for Active INT pin 1.2 V 6 ILI Input Leakage Current 10 7 µA ILO Output Leakage Current 10 7 µA VIH Input High Voltage 2.0 VDD + 0.5 V VIL Input Low Voltage -0.5 0.8 V VOH Output High Voltage (IOH = -2.0mA) 2.4 V VOL Output Low Voltage (IOL = 4.2mA) 0.4 V VOLB Output Low Voltage (INT pin) 0.7 V 9 Device in backup mode (VDDVLO may result in improper operation. Slew rate for proper transition between the locked-out condition and normal operation. Rev 1.1 May 2003 Page 21 of 28 FM3808 Supervisor AC Parameters (TA = -40° C to + 85° C, VDD = 4.5V to 5.5V unless otherwise specified) Symbol Parameter Min Typ Max Units tIPU INT output pulse width 150 200 300 ms tFCO Flags/Control register access to INT pin clear 100 ns Notes 1 2 P/L = 1; pulse mode. P/L= 0; level mode. From the end of the access where the Flags/Control register is read and the flag cleared. Notes 1 2 Data Retention (VDD = 4.5V to 5.5V unless otherwise specified) Parameter Min Units Notes Data Retention 10 Years 1 Notes 1. The relationship between retention, temperature, and the associated reliability level will be characterized separately. Capacitance (TA = 25° C, f=1.0 MHz, VDD = 5V) Symbol Parameter CIO Input/output capacitance (DQ) CI Input capacitance CXTAL X1, X2 Crystal pin capacitance Notes 1 2 Max 8 6 12 Units pF pF pF Notes 1 1 1, 2 This parameter is periodically sampled and not 100% tested. The crystal attached to the X1/X2 pins must be rated as 6pF max. AC Test Conditions Input Pulse Levels Input rise and fall times Input and output timing levels Equivalent AC Load Circuit 0 to 3V 10 ns 1.5V 1.3V 3300Ω Output 50 pF Rev 1.1 May 2003 Page 22 of 28 FM3808 Timing Diagrams Read Cycle Timing tRC tCA tPC CE tAH tAS A0-14 tOE OE tOHZ DQ0-7 tCE tHZ /CE-Controlled Write Cycle Timing tWC tCA tPC CE tAH tAS A0-14 tWS tWH WE OE tDS tDH DQ0-7 Rev 1.1 May 2003 Page 23 of 28 FM3808 /WE-Controlled Write Cycle Timing tWC tCA tPC CE tAH tAS A0-14 tWS tWH tWP WE OE tWZ tWX DQ0-7 out tDH tDS DQ0-7 in Rev 1.1 May 2003 Page 24 of 28 FM3808 Power Cycle Timing Picture assumes VSW < VINT tF tR VINT VLO VSW VRST VDD VINT VLO VSW VRST tFCO tINT tIPU INT tPD tRI Inputs CE INT Pin Timing INT source occurs INT source occurs P/L=1 tIPU P/L=0 INT tFCO CE INT source flag cleared Rev 1.1 May 2003 Page 25 of 28 FM3808 INT pin Timing at Power Up VTP VTP VDD VRST VBAK Register 7FF6 INT hi-Z 24h New value written (register is battery backed) tIPU driven low tIPU (~ 200 ms) This drawing shows the behavior of the INT at power-up first without a backup supply VBAK then with a backup supply. On power-up without a backup power supply, the device will initialize with the Power Fail bit enabled (PFE=1) and the INT pin will be active low (open drain) and pulse output mode (P/L=1). The Interrupt register 7FF6h defaults to the value 24h. Once VDD is established, the register 7FF6h may be written to a different value to change the behavior of the INT pin. When a battery is used as a backup source, VDD must be applied prior to inserting the battery to prevent battery drain. Once VDD is applied and a battery is inserted, the current drain on the battery is guaranteed to be IBAK(max). Rev 1.1 May 2003 Page 26 of 28 FM3808 Mechanical Drawing 32-pin TSOP (JEDEC MO-142 variation BA) All dimensions in millimeters 13.30-13.55 11.70-11.90 1.20 max 1 2 3 0.50 typ 0.17-0.27 typ 7.90-8.10 0.95-1.05 0.05-0.15 13.30-13.55 R 0.08 min R 0.08-0.20 0.50-0.70 ° 0 - 5° Rev 1.1 May 2003 Page 27 of 28 FM3808 Revision History Revision 0.1 0.2 1.0 1.1 Date Dec 19, 2000 Sept 19, 2001 July 31, 2002 May 2, 2003 Summary of Changes Initial Release Changed ISB spec, redefined crystal capacitance specs, data retention temperature condition. General cleanup. Fixed shading in Table 1, registers 7FF1h–7FF4h. Increased storage temperature range. Changed input data hold time tDH to 5ns. Changed tCA (max) spec. Reworded notes 2 and 3 in DC Operating table. Rev 1.1 May 2003 Page 28 of 28