Why not just monitor the supply voltage, trigger an interrupt when the crit ical voltage is met during a dropout and save state to flash. It only takes a ms or so to do that. When you detect the drop of the supply voltage, set all peripherals and external circuits to idle, and use only power for the flash storage task
Cheers
Klaus
Didn't find your answer? Ask the community — no account required.
M
Michael A. Terrell
Something like the General Instruments 1400? It was used in a lot of Unisys terminals to store their configuration.
W
WangoTango
We use a lot of FRAM. You damn near can't wear it out and you can write at almost any sane SPI speed.
T
Tim Williams
I don't know that it actually has a limit. It's like asking whether type 2 ceramic capacitors have a wear life. (Which they do, in a sense, but that's about aging: time, not cycles.)
I haven't heard if there's actually weird physics that does cause a finite cycle life on very small bits of dielectric.
Well, they do specify a write limit in their data sheets, but it is an insanely large number. The FM24C256, which we use a tone of, specifies
10^10 read/write endurance. I don't know how they arrived at that, but there it is.
J
John Larkin
At a mere 100K writes per second, that can be done in less than a month.
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
W
WangoTango
Well, obviously, you pick a part appropriate for the job. I don't write more than 1/S on average. If you are just using it to store calibration data, it would last "forever". The newer FRAMs have endurances much, much higher, I just don't need it.
The OP said,
Which would make the part ideal, which is why I mentioned it.
R
rickman
I can't tell which way John went on this, but if anyone else has a need for nonvolatile storage with unlimited read/write capability here is a pretty good solution. Looks like they have a 4 kbit and a 16 kbit part.
formatting link
Infinite read/write cycles to the SRAM array with EEPROM to back up the SRAM on power fail. 1 millions write cycles on the EEPROM and >200 year retention. "The EERAM works from -40C to 125C and is automotive grade qualified." Beats the hell out of anything using a battery.
Looks like they have some block protect logic for parts of the SRAM array and the chip has three ways to trigger a backup operation. It senses power fail itself and uses a cap to power the backup operation. It can be triggered by an external signal or it can be commanded by software. Backup takes 50 ms.
In a SOIC-8 or TSSOP package it is not very large and is even available in an 8 pin PDIP. The only down side I see it a 5 volt power requirement. Opps, they have 3 volt versions as well! I'm impressed. I wonder if anyone else makes these?
Rick C
P
Phil Hobbs
Potentially nice part, thanks.
Cheers
Phil Hobbs
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics
160 North State Road #203
Briarcliff Manor NY 10510
hobbs at electrooptical dot net
http://electrooptical.net
L
Lasse Langwadt Christensen
speaking of nice parts,
formatting link
J
John Larkin
The board is still in layout (only about 500 more BGA balls to connect) and I have an SO8 SPI serial flash on the schematic.
We have a microSD card for the main storage, and we found some "high endurance" SD cards, intended for video storage, so we might use the SD as the current-setting backup, with some algorithm that further helps the high-endurance+wear leveling. Or I can stuff a FRAM into the SPI location.
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
Join the Discussion
Have something to add? Share your thoughts — no account required.
Didn't find your answer?
Ask the community — no account required
Report Content
You are reporting this content to the moderators. They will look at it
ASAP.