What would cause corruption of EPROM data?

[brogenville]

Following various strange issues with the running of my car, and with things getting progressivly worse to the point that the engine virtually refused to run at all, I removed my MBE ecu and sent it off to the OEM to see whether any it had any faults present (I'd pretty much tried everything else at this point). It turns out that the ecu itself works fine, but that was data corruption on the EPROM chip itself. Quite the extent of this corrupton I cant say, but it was sufficient anyway to stop there being any lambda correction possible.

All in all, I view this as being good news, as despite the cost of getting a new chip burned (which at £160 seems pretty steep), at least I now why whats stopping the engine from running. The only thing I'm worried about now is putting the ecu back in, and having the new chip cooked as soon as I start it up. So, my question is: is this EPROM degredation likley to be a random event, or is there likley to be something else lurking that could cause the exact same thing to reoccur straight away?

Anyone else suffered from EPROM corruption able to comment?

The only thing I can think of is that for a short period of time I had the ecu earth bolted onto the block at the starter motor bolts- perhaps some stray current from the starter earthed through the ecu?

Thanks as always in advance guys.

[annodomini2]

1. EMI (Electro magnetic Interference), this would be a design flaw.

2. Shorting somewhere

3. Too many erase/write cycles

4. Duff chip.

ETA: 5. Duff Software

Edited by annodomini2 on Wednesday 28th July 13:10

[brogenville]

But for 3 above (too many erase/write cycles), the acceptable number is likley to be pretty high isn't it (as in, in the thousands)?

I only say that because I did reset the adaptive map values quite a few times during the diagnosis of the problems. Not enough to "wear out" the chip though I would think.

Thnaks for your answer by the way- gives me food for thought.

[Le TVR]

Is it an EPROM, EEPROM or other?

Does it have a clear 'window' on the top?

Usually EEPROMs are more common as they are more convenient to write and erase. They can also suffer from data loss or corruption due to poor power supply regulation and decoupling especially on the reset line (ie during switch on/off).

[stevieturbo]

Jump starting, any sort of power surges...messing with wiring, or wiring faults..

Who knows really

But given you did mess with some wiring....that wont have helped matters, like I said at the time.

[annodomini2]

brogenville said:

But for 3 above (too many erase/write cycles), the acceptable number is likley to be pretty high isn't it (as in, in the thousands)?

I only say that because I did reset the adaptive map values quite a few times during the diagnosis of the problems. Not enough to "wear out" the chip though I would think.

Thnaks for your answer by the way- gives me food for thought.

Depends on the chip, old designs have very few write cycles, can be in the 10's to 100's

As LeTVR said, also depends on the technology type.

[Mr2Mike]

If it really is an EPROM (which is extremely unlikely unless it's an ancient ECU) then corruption could be caused by age, exposure to radiation or long term exposure to sunlight if the erase window has not been covered (all rather unlikely).

If it's an EEPROM then modern versions of these tend to have very high endurance; 100k to 1Million writes or so. Corruption can be caused by the device being powered down during a write - this should be allowed for in the design.