NVRAM getting scrambled when driving

[antus]

Hi amp rating thermos might kick out some RFI or power noise too. Maybe pull the fan relays out and keep an eye on your temperature when you drive around and see if that helps.

[BennVenn]

Is your earth strap fitted and in good condition? A bad strap will cause the charging circuit to use whatever ground it can find, including loom earths.

Odd though as the ecu has pretty good power filtering and noise rejection. I'd think it is more likely a bad NVRAM PCB, maybe a dry joint or crack in the /WE trace on the PCB or the WE pull up resistor. Hard to see a spike trashing the entire contents, maybe a few bytes here and there but an entirely corrupted NVRAM sounds more like the code has crashed and the ECU is writing garbage to it which suggests the ECU has control of the WE line.

Any chance of borrowing another NVRAM to rule it out?

[someguy360]

Is your earth strap fitted and in good condition? A bad strap will cause the charging circuit to use whatever ground it can find, including loom earths.

Any chance of borrowing another NVRAM to rule it out?

Thanks for the continued help guys,

Gareth, I am running resistor NGK plugs.

BennVenn yep earth strap is good, I put much thicker gauge earths in last time I had the engine out than the crap stuff Holden used from factory but I'll double check the connections tonight in case somethings loose. Unfortunately I don't know anyone else locally with another NVRAM board

I'd ruled out the Read/write wire solder joint and the wideband solder joint on the PCM PCB as I've replaced the PCM without change.

antus, I'll give the thermo fans a try too, thankfully it's about 5 degrees in these parts in the evening and highway conditions so I should be able to get away with a decent drive.

[BennVenn]

Is your VR a manual? I could loan you a homemade '808 style NVRAM for testing but I don't think it'll work in the Auto VR ECU

[someguy360]

Is your VR a manual? I could loan you a homemade '808 style NVRAM for testing but I don't think it'll work in the Auto VR ECU

Thanks so much for the offer but it's an auto 424 style unfortunately.

[antus]

nvrams are the same for 808 and 424 (the ecm nvram). but you've tried 2 boards, and 3 nvram ICs I believe (1 in the new kit and 1 from a distributor?) Gareth already replaced the kit? And corrupting with the RW jumper off is another reason it cant be the RW circuit.

[someguy360]

nvrams are the same for 808 and 424 (the ecm nvram). but you've tried 2 boards, and 3 nvram ICs I believe (1 in the new kit and 1 from a distributor?) Gareth already replaced the kit? And corrupting with the RW jumper off is another reason it cant be the RW circuit.

Hi Antus,

I haven't tried another NVRAM board, I've only had the NVRAM chip replaced not the board itself, it's still the original one I purchased at the start of the year.

I've replaced the NVRAM chip and the PCM but not the NVRAM board I haven't had the kit replaced.

BTW I purchased the replacement NVRAM chip from Gareth rather than a distributor so I knew it was a genuine one (I've heard horror stories of fake ones etc)

The NVRAM board itself is the only piece of the puzzle thats still original from before the problem started and hasn't been replaced.

[BennVenn]

If you haven't tried another NVRAM board, Logically that is what I'd be trying to rule out. I'm not too familiar with the official NVRAM design but from an electrical viewpoint, if the trace from the WE pin on the Dallas NVSRAM IC is broken, either completely or intermittent, anytime the CPU is running and that signal goes low, it'll trash the data instantly, regardless of the jumper position. This sounds to me like what is happening, possibly vibration induced as you say it doesn't happen idling at home. Just a guess and you'll need a known good full NVRAM module to confirm

[someguy360]

If you haven't tried another NVRAM board, Logically that is what I'd be trying to rule out. I'm not too familiar with the official NVRAM design but from an electrical viewpoint, if the trace from the WE pin on the Dallas NVSRAM IC is broken, either completely or intermittent, anytime the CPU is running and that signal goes low, it'll trash the data instantly, regardless of the jumper position. This sounds to me like what is happening, possibly vibration induced as you say it doesn't happen idling at home. Just a guess and you'll need a known good full NVRAM module to confirm

Thanks mate, I agree it's probably the next logical step.

@Gareth @antus do you have one lying around I might be able to try (happy to pay a deposit or something for security while it's with me) then if it proves to be the board I can either send back the faulty one or purchase that one etc, whatever works?

Thanks again for everyones help means a lot, hopefully I'll be able to get back to road tuning again soon

[charlay86]

Sounds like a bad connection.
Since we are dealing with the ECUs entire program and data memory storage, every signal line is critical...
Have you read back the scrambled contents of the nvram with an ic programmer after it has been corrupted? Id be interested to have a look.