SRAM based NVRAM

[Staggie]

Hi All,

This may interest some, have been a long time lurker of the forum.

Had some leftover AS7C1024 based SRAM from a project so I thought I would see if I could get it to function as a NVRAM correctly since the Dallas NVRAM's are certainly thin on the ground these days.
Anyway, I finally had success last night getting it to connect (on the bench anyway) and happily read/write from TunerPro and Flash Tool.

Worst part was getting the GQ-4x4 to behave and the WE line and the bank switching logic causing a bit of grief.

the twin CR1225 batteries just clear the case, I need to see if can track down a lower profile SMD header socket for the main connector, currently sitting at 9mm from base of PCB, so should be able to lose a couple mm without putting too much stress on the ECM's pins.

[vlad01]

Excellent! Been waiting for something like this for years.

Although my NVRAM is still working fine (2011), I know it's days are limited and in the few times I bricked it, it's a real pain to reprogram due to how fragile and fickle the pins are.

[Staggie]

I hear you there mate, am glad I drew up a programming adapter at the same time so it just drops straight into the programmer, bricked it more times than I care to admit getting the thing working in the little ol' 808.

The lifespan on the Dallas chip was part of my reasoning for the dual battery, also the vertical batteries allows for a far smaller PCB while keeping the ability for knock board and original netres chips if wanted, yeah the CR1225's aren't going to last anywhere near as long, but as long as one battery has charge can easily swap them out while keeping the data intact.

[Holden202T]

i think the issues i have been having with mine is the same thing you mention programming it ....

could to make the board slightly longer on the side you normally plug a memcal into on the nvram boards and then use the cage battery holders on the edge of the board ?

im using 1632 batteries for size ....

also i use SSM-135-F-DV-LC for the ecu header, its not overly cheap and needs to be ordered, but its 7.3mm insulkation height.

[Staggie]

Happy to share what i found with the programming issues apart from many hrs of swearing at it. Do you have a pullup on the A16 pin? Also anything other than a pullup on the WE pin?

Plenty of room to possibly bring the board out and use smd CR1632 holders, I was just trying keep it tight for this test, plus can change the batteries with memcal still in the ecu, did test in a 808 with the plastic end plate and the battery holders clear by about 1mm, no end plate she is a bit close. A piece of kapton tape will be easy insurance.

Those headers look the goods, just did a digikey order too

[Holden202T]

yeah i chose that header as it has locating holes in the pcb to make assembly easy.

yes pullup on A16, A15 and WE# ... ive tried removing them but it doesnt seem to make a difference. CS2 is also held high.

as for digikey, i try and get enough stuff to order so postage is free

[vlad01]

Which exact model of sram are you using?

On the data sheet, there are faster and slower models as well as a 5v and 3.3v variants.

What interests me is how the faster and lower operating voltage model is more efficient in standby than the other variants.

But I could not find a min standby voltage spec listed anywhere, only the operating voltage ranges, so assuming you've used the 5v one to be compatible with the 808, how does the batteries keep them alive without going to the 3.3v chip? From what I gather from the data sheet, the standby is not listed as it's probably only the one spec for all which is the operating voltage ranges, they do list standby current, but not voltage.

So assuming the min voltage is 4.5 and 3 for the two types are the hard limits in the data sheet, I am confused how the batteries keep them going? Or am I missing something obvious?

I looked into this to see if there was an ultra low power equiv to see if battery life could be improved and looked at the data sheet to find those specs. So I'm interested in what you ended up doing on your setup and how it works regarding all of this.

[Holden202T]

festy put my on to R1LP0108ESN-5SI#B1 as a lower power usage part than what i first had.

does this answer your questions about sleep power usage ?

2026-04-26_22-37-47.jpg

[Staggie]

AS7C1024B-12JIN is my current SRAM, they are now EOL, but the D series is a drop in replacement.
Annoyingly there isnt figures in the data sheet for data retention currently, i will need to test it the old fashioned way now the circuit is functioning correctly.
But you have raised a very valid point Vlad

[vlad01]

Thanks! Odd how the data sheet is missing that info.

It does answer my questions, but looks like the part festy used is a lot lower power than this chip, going by the fact it uses far less than 10 uA in typical scenarios.

Going by the mAh only, a CR1225 is around 48mAh, and if we take the typical value of the 0.6uA. It's around 9 years life (80000 h)

But it's never that simple, the usage might be more than that given the max parameters at different temps. Also, I could not find any calculators or formulas online that did battery life that also accounts for voltage output over time. So to maintaining a min of 2v going by the data sheet, the life would be be obviously a lot less than 9 years. But I'd say it would still be significantly long enough, maybe 1-2 years or so is my guess on a single battery of that size. This is all assuming the specs of what chip Festy provided to 202.

The AS7C1024 appears to be nowhere near as power efficient from what I gather. It does state in the data sheet a max standby power consumption of 55mW for the 5v part. A CR1225 in mWh is around 150, so yeah...

So having the right chip selection should allow you to design in a good battery life.