From memory the 10uF on the large board is right next to the transistors that switch the idle control valve. I would therefore suspect it is in as decoupling for these surges, and will mean it is seeing transients of 1 - 2 A, way outside normal rating for a 10u Electrolytic.As above, I will probably be changing to surface mount ceramic instead for future repairs. Back when these ecu's were designed, a 10u ceramic was a large item, now they're down to 0603 and I think even 0402 size (AKA microscope tiddlywinks).

_________________
If at first you don't suck seed, try drier grain.

For reference I'm now fitting through hole 10uF ceramic - RS 502-0201. Not cheap but being through hole is a better technical solution than a surface mount.

_________________
If at first you don't suck seed, try drier grain.

Another update to this.



I've now had a total of 3 of the ECU's I previously repaired come back again. 2 of these were early ones I repaired and both have had another capacitor fail. A different capacitor in each case.



I'm now changing policy to changing all electrolytics on any ECU's I work on since it seems others are failing. sadly, the others are larger values so not viable to use ceramics. I'm just using good quality high temperature electrolytics which will hopefully give another 20+ years.



Why they are failing I'm not sure. Electrolytics are fragile in a few ways - High ripple currents, especially high frequency - high temperature and leaving unpowered for long periods of time. The ones which failed previously were used for decoupling without a small ceramic in parallel to take the high frequencies - a recipe for disaster.

_________________
If at first you don't suck seed, try drier grain.

Well thanks 'nibbles', I cannot even remember which car it was that you did the ECU on but I am in contact with both new owners and could pass on your details in case they have problems, if that is ok with you?
I have a 1971 triumph 13/60 convertible now ...no electronics

_________________
Previously a UK WRC & a CS
Now...a 13/60 cv herald ,a Honda CR-Z and an Amica

Happy to help if needed. This information would also enable local electronic repair places to carry out repairs as well.

If the file sharing works, this is a picture showing locations and values of all electrolytics. Replacements should have voltage rating at least as high as original, and temperature rating at least 105C.

https://drive.google.com/open?id=19C19y ... SfDezjOoDB

Pictures of 2 ecu's with further failed capacitors (different cap. in each case).

https://drive.google.com/open?id=135OOb ... HYg1PAcB_I

https://drive.google.com/open?id=13AuDZ ... CYPYlIasD3

_________________
If at first you don't suck seed, try drier grain.

Cheers mate. I will save that and pass it to anyone who may need it.
Thanks

_________________
Previously a UK WRC & a CS
Now...a 13/60 cv herald ,a Honda CR-Z and an Amica

That's interesting Chris - I believe the theory in the past was a dubious batch of components, do you think that is still likely or could it be something else?

Cheers for sharing.

Mike

_________________
GT4DC Events Co-ordinator
1996 ST205 UK

I stopped thinking that a while back. Knowing the one on the bottom board was decoupling for the idle solenoid with quite high current ripple I concluded is was ripple, particularly the absence of a ceramic in paralell.

I haven't managed to trace the circuits round the 2 that have failed subsequently. One being only 10V rated would have probably been on a 5V rail, or could have been in a low current scenario in a signal circuit. Either way, probably not such high ripple current. I'm begining to think temperature may be a factor - I've never actually checked how hot it gets, but it is down near blower outlets.

_________________
If at first you don't suck seed, try drier grain.

Probably mostly unrelated but when my Link G4 PNP failed last year I asked about temperature and they said it could have been a factor, the Link has the capacity to monitor its own temperature, when I get the car running again I'll see if I can get some numbers.

_________________
GT4DC Events Co-ordinator
1996 ST205 UK

A further update.

Of the 3 ecus:

1. This was a fairly late repair which I used the ceramic 10uF which worked for a while then failed again. Another capacitor had failed which was replaced locally but still failed to work. Subsequently as a last ditch attempt, the ceramic 10uF capacitors were tested and one had failed. It was replaced with an electrolytic and this ECU now works.
2. This was an MR2 turbo ECU which I originally replaced with electrolytics. In this case, it apparently didn't work. It was returned to me and worked fine in an RC. When it returned again, it apparently worked for 5 minutes then failed again. Although I suspect another issue with the car, he sent the ECU to a specialist repairer and nearly £600 later it now apparently works fine. Supposedly chips were blown etc.
3. This was the second ECU I did on an RC. Done but not yet tested.

On the strength of all this, I'm changing away from using ceramics back to electrolytics. Very surprising as ceramics are usually very reliable. I'm wondering if there are some very large but short voltage surges as electrolytics tend to be a bit more robust to infrequent large but short overvoltages.

_________________
If at first you don't suck seed, try drier grain.