Before I throw money at a new ECM, perhaps someone can confirm my troubleshooting steps on my 95 M42 Ti.
Symptoms: car began to lose power and RPMs on the freeway. Pressing gas pedal would not raise RPMs. I was just able to get it home and park it in my garage. Engine was rough even at idle. No CEL. I disconnected the battery afterwards, so any stored codes were lost.
The next day, I got it to start once, and it ran for maybe 1-2 seconds. All further attempts failed - the starter turned over fine, but none of the cylinders were firing. No CEL, and no new codes, although I don't know how quickly they get stored after a reset happens.
After some troubleshooting, I found that the fuel pump was dead. There was no noise when applying 12V directly to the pump. The new pump buzzes when 12V is applied and it runs when the fuel pump relay pins are jumpered, so the pump and the wiring from fuse box is OK. I can also hear gurgling from around the intake area - presumably it's the fuel pressure regulator doing its job.
The problem is that the car still wouldn't start. Same problem as before. It turns over, but none of the cylinders are firing.
Further tests on the main DME relay show that the ground socket (85) is not being set by the ECM when the key is in the On (second) position and the relay is inserted. Instead, I'm getting about 0.8 volts difference between the relay socket and the grounding point on the front wheel well. Sockets 30 and 86 are showing 12V, as expected, and the relay clicks when energized. I also tried opening the relay and gently cleaning the soot from the contacts with a strip of paper.
Next, I removed and unplugged the ECM. (No signs of any water damage, BTW, just lots of dust and bugs in the "ECM cave") The wiring between socket 85 for the main DME relay and its corresponding pin on the ECM connector is fine. I can also measure the new cam and crank sensors' resistance from their pins on the ECM connector.
So, I believe this leaves me with only one final possibility - the ECM may be bad. I don't understand how the worn out fuel pump problem could be related to the ECM problem - the fuel pump relay and its fuse should have isolated the pump from the rest of the system, no?
I also opened the ECM case, and checked the solder joints and traces with a jeweller's loupe but saw nothing suspicious. I reflowed the pins on a few of the bigger components, the ones hidden under the heatsinks, but it didn't solve the relay problem.
Other things I checked/replaced, which might help eliminate other possibilities:
- Battery was new last year, and shows 12.5V.
- I found and cleaned about 4 or 5 grounding points in the engine bay. I may have missed some, but the ones I saw were all quite clean.
- Coil pack is getting voltage from its plug and the coil resistance values are within specs. I do see some fine cracks in the outer casing around the mounting bolts. I pried the top cover off, and the epoxy stuff inside looks solid. Any arcing problems here would affect the plugs, not the ECM, right?
- I also replaced the cam and crank sensors while I was troubleshooting - both the old ones were OK though.
- AFM is getting power from its cable, and the resistance varies when the door is swung open.
- Ignition wires and plugs are new.
- Idle control valve was surprisingly clean and dry. PO had replaced it not too long ago.
- I spotted and fixed only one vaccuum leak while replacing cam sensor - the intake boot had a big rip.
- Car has some sort of alarm/keyless entry system installed. It seems to be responding OK. Can these fail and disable or otherwise mess up the ECM somehow?
Also, can someone confirm whether or not the fuel pump should be priming the system for 2 seconds with the key is at the On (second) position in a Motronic 1.7 system? I can't hear anything on mine. I saw this mentioned somewhere in reference to other E36s, but the bentley manual doesn't say anything about it.
Thanks for any suggestions. If anyone in the Seattle, WA area has a spare 282 ECM they would like to sell, feel free to PM me.
Michael