pgmfi.org

ECU is P36 rev0 (Beat), after replacing caps engine runs great, except for eacv keeping idle at 1500rpm and now after changing the caps and making car run better ECU light is on, even in service mode, which according to factory service manual means that ecu is broken.

How find out what the fault is, when now there are no obviously damaged components?

As far as I looked at the circuit it seem that C3 (+C5), which was worst and biggest "filters" gnd for eacv(!) and all injectors. I'm not sure if ecu issue is related to the eacv thought.

update: crazy, the leak form cap eaten the ground trace, but so clean i did not know it was supposed to be there

the ECU was trying to tell you that something was wrong internally, but unfortunately it only has one way to communicate a very long list of potential issues.
Solid CELs are always interesting to troubleshoot, especially after major work has been done.

so repairing the missing ground trace resolved the solid CEL?


any chance you'd be able to take pics of your ECU and post them on here? P36 is not a common ECU on here, some pics may help someone else in the future.

No, even replacing eaten traces, did not help.

Can I use CN2 to read data of it? There is clearly some data sent visible on oscilloscope, but if i connect rx and tx to the pc using usb virtual com port and watch terminal with 9600 or 38400. I see no incoming data. I did not send anything to the ecu or modify j12 or j4 - i do not really understand how that works.

Also QM2 sta413a transistor set seems to be broken.

No idea what pictures could help - pictures of the ecu with eaten trace?

eaten ground traces
https://i.imgur.com/fsTtQwJ.png

not eaten ground
https://i.imgur.com/W2FQRDV.png

if I send ecu some random gibberish, it actually responds with

If i send 'a' it responds with 'b' enter
I was not able to find how to communicate with stock ecu using uart

I found only this but it is not using CN2
https://github.com/kerpz/ArduinoHondaOB ... bd_uni.ino