Crazy Adaptives

[Hawkeye1922]

Has anyone seen adaptive maps looking like this?

This appeared after an hour or so of driving and, unsurprisingly, the car is undriveable. The laptop can seemingly read the MBE data normally, and everything else looks normal (engine will idle). The ECU won't take the reset-adaptives command either.



Is there any explanation/resolution other that "ECU F , buy a new one"??

[notaping]

Hi, Very colourful, but obviously not correct. If you are able - I would open up the MBE and check for any water damage, dry joints etc. Even if it was o2 sensor failure you would get a more meaningful chart. One of my sensors failed last year, but it still recorded this. . .



Other than that - the picture looks to regular to be anything other than a MBE fault.

Anyone else - suggestions?

[Englishman]

Yes, looks like an ECU fault, but worth checking the map on the odd cylinders before you start spending.

[ukkid35]

My MBE can suffer from heat soak, causing the car to be almost undriveable, however, once it cools down again it's fine

But since you can't reset the map, that sounds more serious

Even so, I would try reseating the chips, and looking for water damage as already suggested

[notaping]

Hawkeye1922 - Are you using the correct cable to connect the MBE to your computer? The MBE has 2 data modes - Broadcast mode, & Byte mode. All varieties of diagnostic software available will only run in Byte mode. This is set when connecting the recommended TVR configured cable.

I only ask because I don't know what the software would display if it queried the adaptive memory range in Broadcast mode. Also the reset functions won't work in Broadcast mode.

Just a thought.

[Hawkeye1922]

ukkid35 said:

My MBE can suffer from heat soak, causing the car to be almost undriveable, however, once it cools down again it's fine

But since you can't reset the map, that sounds more serious

Even so, I would try reseating the chips, and looking for water damage as already suggested

My symptoms seem to seem to appear after a period of driving, and then go away again after several hours (at least 3) stopped. Is that what you experience when you say "heat soak"?

Have you ever tried plugging diagnostics in whilst its undriveable?

[Hawkeye1922]

Englishman said:

Yes, looks like an ECU fault, but worth checking the map on the odd cylinders before you start spending.

Well, yesterday, the odd cylinders map was identical
This morning I have switched on and plugged in again. Haven't started the engine. The odd bank seems to have recovered its original adaptives. The even bank is all zeros

My current hypotheses are:
1. UKKid's heat soak phenomenon
2. The NVRAM chip is on its way out, either because its internal battery is dying, or it's just old, causing it to give erroneous values SOMETIMES
3. joints/chip seating as suggested
4. Other difficult-to-diagnose ECU fault, causing the NVRAM chip to be incorrectly read SOMETIMES

So I'm going to invest £25 in a replacement NVRAM chip, and while i'm changing it, check for other obvious dry joints. My issue is then testing it - I think I've run out of my 5 breakdown recoveries this year...

Anyone got a spare ECU they can lend while I'm testing it??

Edited by Hawkeye1922 on Sunday 30th June 14:58

[Hawkeye1922]

notaping said:

Hawkeye1922 - Are you using the correct cable to connect the MBE to your computer? The MBE has 2 data modes - Broadcast mode, & Byte mode. All varieties of diagnostic software available will only run in Byte mode. This is set when connecting the recommended TVR configured cable.

I only ask because I don't know what the software would display if it queried the adaptive memory range in Broadcast mode. Also the reset functions won't work in Broadcast mode.

Just a thought.

Thanks - good thought. My cable is well over 10 years old, probably came from China in the first place, worked well with the original TVR diagnostic software for years. I've only recently started using it with the new software, but it has been dependable so far.

I think what's leaning me toward ECU fault is:
- when the adaptive maps are bonkers, the engine behaves exactly as I would expect according to those bonkers maps (complete fuel cut in the blue areas), so I think they are "real" in RAM, and the ECU onboard software is operating the engine correctly
- when the maps aren't bonkers, I can successfully reset maps and fault codes
- even when the maps are bonkers, the diag software is still able to read all other ECU parameters, and they look normal

Would be good to know how to determine whether the unit is in broadcast or byte mode

[notaping]

Hi, If you've had and used the cable for a while then I think we can rule that one out, and upon reflection It must be in Byte mode to accept the "Adaptive" command in the first place, also it wouldn't read the other values if it wasn't - so all ok there.

I think you're correct in that the values must be stored in the Dallas chip - however if the chip battery was failing I would expect the values to drop to zero and not contain a repeating sequence.

Incidentally - while the numbers are wrong - they are not random. The adaptive values are stored column by column starting from the top left - ending up bottom right. Your numbers go from negative max to positive max in 256 steps (repeating). I don't know what that tells us, but I suspect it's the part of the MBE that generates the values that's at fault and not the part that stores them.

As for heat soak - get yourself some freeze spray from Amazon and when it happens again start spraying components on the board to see if you can recover it.

[Juddder]

My gut feel, and from my experience with 1980's arcade machines, is that the RAM chips can often fail over time and cause this kind of random behaviour on heat / time turned on etc. - All Williams boards with the mixed voltage RAM are a b*gger for doing this

The Dallas chips are now pretty scarce but I believe the DS1230Y-120IND+ from Analog is a direct pin for pin replacement and includes a Lithium power source with 10 years minimum lifetime so would be a perhaps better modern replacement

https://www.analog.com/media/en/technical-document...

One of the guys with a Tamora did this and it was recommended to Jules (I believe that's his handle that asked the question in this thread) so should be an easier part to source from Mouser, or CPC - example links below

https://www.pistonheads.com/gassing/topic.asp?h=0&...

On borrowing an ECU - I have a spare full-functioning Speed 6 ECU but sadly these refuse to boot when they are swapped in with V8 EPROMs (*there must be something different with the revisions on code start jump point or something like that) so I'm not sure if it would be much use for you to test, but you are of course more than welcome to borrow it.

Finally, I'm pretty sure I have a few of the Dallas NVRAM chips as backups which I am more than happy to post one out to you and you can try it to see if it fixes / changes your problem before having to spend out cash on a new one

HTH Alex

https://cpc.farnell.com/maxim-integrated-products/...

https://eu.mouser.com/ProductDetail/Analog-Devices...

[Hawkeye1922]

Juddder said:

On borrowing an ECU - I have a spare full-functioning Speed 6 ECU but sadly these refuse to boot when they are swapped in with V8 EPROMs (*there must be something different with the revisions on code start jump point or something like that) so I'm not sure if it would be much use for you to test, but you are of course more than welcome to borrow it.

Finally, I'm pretty sure I have a few of the Dallas NVRAM chips as backups which I am more than happy to post one out to you and you can try it to see if it fixes / changes your problem before having to spend out cash on a new one

Greatly appreciated - thanks. These goodies arrived today





So the patient is coming into surgery in about half an hour...

[Hawkeye1922]

An update: The first Dallas chip, from a well-known online retailer that sounds a bit like Rodeo Sack, arrived with bent pins - useless

So I ordered a couple more from Farnell. One's now in - performing fine so far. Just need to take it for a really long drive and get it properly warm.

One very useful discovery - if you take the Dallas chip out altogether the car runs fine. The adaptive maps are all zeros at first, but they still populate with values as the learning takes place. Of course as soon as you key off its back to zero. So this will at least get you home if you experience the same symptoms I did.