RAMTRON 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
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.
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
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
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
FM3808-70-T
FM3808DK
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
Ordering Information
70 ns access, 32-pin TSOP
DIP module development kit
Documentation for the DIP module development kit is
available separately.
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
VDD
VBAK
System Supervisor
Low VDD monitor/
Watchdog timer
Watchdog
timebase
INT
FRAM Array
32,752x8
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
Input
/WE
Input
X1, X2
INT
Input
Output
VBAK
Supply
VDD
VSS
Supply
Supply
Rev 1.1
May 2003
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
Address
7FFFh
7FFEh
7FFDh
7FFCh
7FFBh
7FFAh
7FF9h
7FF8h
7FF7h
7FF6h
7FF5h
7FF4h
7FF3h
7FF2h
7FF1h
7FF0h
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
D2
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
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
Description
7FFFh
Timekeeping – Years
D7
D6
D5
D4
D3
D2
D1
D0
10 year.3
10 year.2
10 year.1
10 year.0
Year.3
Year.2
Year.1
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
D6
D5
D4
D3
D2
D1
D0
0
0
0
10 Month
Month.3
Month.2
Month.1
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
D6
D5
D4
D3
D2
D1
D0
0
0
10 date.1
10 date.0
Date.3
Date.2
Date.1
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
D6
D5
D4
D3
D2
D1
D0
0
0
0
0
0
Day.2
Day.1
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
D6
D5
D4
D3
D2
D1
D0
0
0
10 hours.1
10 hours.0
Hours.3
Hours2
Hours.1
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
D6
D5
D4
D3
D2
D1
D0
0
10 min.2
10 min.1
10 min.0
Min.3
Min.2
Min.1
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
D6
D5
D4
D3
D2
D1
D0
0
10 sec.2
10 sec.1
10 sec.0
Seconds.3
Seconds.2
Seconds.1
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
Description
7FF8h
Control-Nonvolatile
/OSCEN
D7
D6
D5
D4
D3
D2
D1
D0
OSCEN
Reserved
Reserved
CALS
CAL.3
CAL.2
CAL.1
CAL.0
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
Reserved
CALS
WDS
/WDW
WDT.5-0
7FF6h
WIE
AIE
PFE
ABE
H/L
P/L
7FF5h
/M
Rev 1.1
May 2003
D7
D6
D5
D4
D3
D2
D1
D0
WDS
WDW
WDT.5
WDT.4
WDT.3
WDT.2
WDT.1
WDT.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.
Interrupts
D7
D6
D5
D4
D3
D2
D1
D0
WIE
AIE
PFE
ABE
H/L
P/L
Reserved
Reserved
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.
Alarm – Date of the month
D7
D6
D5
D4
D3
D2
D1
D0
M
0
10 date.1
10 date.0
Date.3
Date.2
Date.1
Date.0
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.
Page 5 of 28
FM3808
Address
Description
7FF4h
Alarm – Hours
D7
D6
D5
D4
D3
D2
D1
D0
M
0
10 hours.1
10 hours.0
Hours.3
Hours2
Hours.1
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
D6
D5
D4
D3
D2
D1
D0
M
10 min.2
10 min.1
10 min.0
Min.3
Min.2
Min.1
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
D6
D5
D4
D3
D2
D1
D0
M
10 sec.2
10 sec.1
10 sec.0
Seconds.3
Seconds.2
Seconds.1
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
WDF
AF
PF
CF
TST
CAL
W
R
Rev 1.1
May 2003
Flags/Control
D7
D6
D5
D4
D3
D2
D1
D0
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.
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
Rev 1.1
May 2003
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
Page 7 of 28
FM3808
32 Hz
512 Hz
Clock
Divider
Oscillator
1 Hz
W
Update
Logic
32.768 kHz
crystal
CF
Years
8 bits
Months
5 bits
Date
6 bits
Hours
6 bits
Minutes
7 bits
Seconds
7 bits
Days
3 bits
User Interface Registers
R
Figure 2. Real-Time Clock Core Block Diagram
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
± 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.
Page 8 of 28
FM3808
Table 2. Calibration Adjustments
Rev 1.1
May 2003
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
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
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
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
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
Program Calibration D4-D0
00000b
00001b
00010b
00011b
00100b
00101b
00110b
00111b
01000b
01001b
01010b
01011b
01100b
01101b
01110b
01111b
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
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 Match Bit Examples
Seconds
1
0
0
0
0
Rev 1.1
May 2003
Minutes
1
1
0
0
0
Hours
1
1
1
0
0
Date
1
1
1
1
0
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
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).
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 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.
32.768 kHz
Oscillator
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.
Clock
Divider
1 Hz
32 Hz
Zero
Compare
Counter
WDS
WDW
7FF0.7
WDF
Load Register
D
Q
Q
write to
Watchdog
register
7FF7.5-0
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
Rev 1.1
May 2003
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.
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
VDD
WIE
P/L
PF
Power
Monitor
Pin
Driver
PFE
VINT
INT
H/L
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
Rev 1.1
May 2003
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.
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.
FFFFh
Block 31
FC00h
FBFFh
Block 30
F800h
Memory Architecture
F7FFh
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.
F400h
F3FFh
Block 28
F000h
0FFFh
Block 3
0C00h
0BFFh
Block 2
0800h
07FFh
Block 1
0400h
03FFh
Block 0
0000h
Figure 6. Address Blocks
A9-A8
11b
Row 255
Row 254
Row 253
Row 252
Row 3
Row 2
01b
Row 1
10b
Block 4
A14-A10
00100b
Row 0
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
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
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.
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.
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.
Users who are modifying existing designs to use
FRAM should examine the memory controller for
timing compatibility of address and control pins.
Valid Memory Signaling Relationship
CE
FRAM
signaling
Address
Data
Address 1
Address 2
Data 1
Data 2
Invalid Memory Signaling Relationship
CE
SRAM
signaling
Address
Data
Address 1
Address 2
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
x2
VDD
Schottky
Diode
VBAK
FM3808
Backup Capacitor or
Super/Golden Cap
Capacitor Size
100 µF
1000 µF
10000 µF
0.47 F (super cap)
Backup Time
3 minutes
30 minutes
5.1 hours
240 hours
Figure 9. Capacitor Backup Circuit
Rev 1.1
May 2003
Page 18 of 28
FM3808
IBAK vs VBAK
2.5E-06
IBAK
2.0E-06
1.5E-06
1.0E-06
5.0E-07
0.0E+00
4.7
4.5
4.25
4
3.75
3.5
3.25
3
2.75
2.5
VBAK
Figure 10. Backup Current vs. Voltage
IBAK may go up or down from the specified value as
a function of the capacitive load.
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
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
Error ppm
-60
-80
-100
-120
-140
-160
85
75
65
55
45
35
25
15
5
-5
-15
-25
-35
-200
-45
-180
Ambient Temperature
C
Figure 11. Typical Crystal Error vs. Temperature
Rev 1.1
May 2003
Page 19 of 28
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 (VDD<VBAK)
Notes
1. VDD = 5.5V, /CE cycling at minimum cycle time. All inputs at CMOS levels, all outputs unloaded.
2. VDD = 5.5V, /CE at VIH, All other inputs and DQ pins at TTL levels.
3. VDD = 5.5V, /CE at VIH, All other inputs and DQ pins at CMOS levels.
4. VBAK = 3.0V, VDD < VBAK; oscillator running. This parameter is characterized, not 100% tested.
5. VSW occurs when VDD drops below VBAK. VSW is also the point at which the timekeeper draws current from the VBAK pin,
6.
7.
8.
9.
rather than from VDD. VSW is not otherwise used for control signals or functions.
INT pin conditions are IOL = 80 uA and VOL = 0.4V.
VIN, VOUT between VDD and VSS.
Memory and register access is blocked when VDD < VLO.
VDD=0, VBAK = 3.0V, IOL = 4.2 mA.
Rev 1.1
May 2003
Page 20 of 28
FM3808
Read Cycle AC Parameters (TA = -40° C to + 85° C, VDD = 4.5V to 5.5V unless otherwise specified)
Symbol
Parameter
Min
Max
Units
tCE
Chip Enable Access Time ( to data valid)
70
ns
tCA
Chip Enable Active Time
70
2,000
ns
tRC
Read Cycle Time
130
ns
tPC
Precharge Time
60
ns
tAS
Address Setup Time
0
ns
tAH
Address Hold Time
10
ns
tOE
Output Enable Access Time
10
ns
tHZ
Chip Enable to Output High-Z
15
ns
tOHZ
Output Enable to Output High-Z
15
ns
Write Cycle AC Parameters (TA = -40° C to + 85° C, VDD = 4.5V to 5.5V unless otherwise specified)
Symbol
Parameter
Min
Max
Units
tCA
Chip Enable Active Time
70
2,000
ns
tCW
Chip Enable to Write High
70
ns
tWC
Write Cycle Time
130
ns
tPC
Precharge Time
60
ns
tAS
Address Setup Time
0
ns
tAH
Address Hold Time
10
ns
tWP
Write Enable Pulse Width
30
ns
tDS
Data Setup
30
ns
tDH
Data Hold
5
ns
tWZ
Write Enable Low to Output High Z
15
ns
tWX
Write Enable High to Output Driven
10
ns
tHZ
Chip Enable to Output High-Z
15
ns
tWS
Write Setup
0
ns
tWH
Write Hold
0
ns
Notes
1
2
4
5
1
1
Notes
1
1
1
2
2
This parameter is periodically sampled and not 100% tested.
The relationship between /CE and /WE determines if a /CE- or /WE-controlled write occurs. There is no timing
specification associated with this relationship.
Power Cycle Timing (TA = -40° C to + 85° C)
Symbol
Parameter
tINT
INT signal active after VTP
tPD
Last Access Complete to VLO
tRI
VLO to inputs recognized on power-up
tR
Rise time of VDD from VBG to VLO
tF
Fall time of VDD from VLO to VBG
Notes
1
2
3
Notes
Min
0
1
100
100
Max
100
Units
ns
ns
µs
µs
µs
Notes
1,2
1,3
1,4
1,5
1,5
This parameter is periodically sampled and not 100% tested.
If power monitor is programmed to generate INT.
Access is blocked at VLO. The last access should be complete prior to reaching VLO. The early warning power fail interrupt
may be useful in accomplishing this.
Failing to satisfy tRI may result in the first access being ignored. Failure to raise /CE to a logic high prior to VDD>VLO 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
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.
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
tWH
tWS
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
VDD
tR
VINT
VLO
VINT
VLO
VSW
VSW
VRST
VRST
tFCO
tIPU
tINT
INT
tPD
tRI
Inputs
CE
INT Pin Timing
INT
source
occurs
INT
source
occurs
P/L=1
P/L=0
tIPU
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
Register
7FF6
VBAK
24h
New value written (register is battery backed)
tIPU
INT
hi-Z
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
1.20
max
13.30-13.55
11.70-11.90
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
Rev 1.1
May 2003
0°- 5°
Page 27 of 28
FM3808
Revision History
Revision
0.1
0.2
Date
Dec 19, 2000
Sept 19, 2001
1.0
July 31, 2002
1.1
May 2, 2003
Rev 1.1
May 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.
Page 28 of 28