Flakey ECU?

[Bicyclops]

So I'm breaking in my engine on a basemap tune and it's running rough and afterfiring. I pulled coil connectors and one of the trigger wires is shorted to ground. The rest of them ~33ohm. I unplugged the connector at the ECU and plugged it back in and low and behold it's fixed. Will it stay fixed? Runs good, afterfiring gone. I put a couple of hundred miles on it and all of a sudden it's doing it again. Same wire shorts to ground. Vexing.

Ideas? Find another 02 ECU and try that?

Ed

[lpmagruder]

I'd look for shorts before replacing the ECU. If you didn't redo the engine harness wiring on an 02 donor that's 21+ year old wiring that's been sitting in a really hot spot. Even on my 07 donor it was brittle, I have an 03 motor where the wire insulation was flaking off when bending that section of harness to remove from intake manifold during a rebuild.

Clamping the volt meter on the setting where it beeps when shorted, then walking around the car giving all the wiring a good shake is usually my first step.

Those coil connectors go bad so often that there's readily available splice-in sections. I've had one intermittently fail open, though a little more has to go wrong for it to short to ground.

[FFRWRX]

I have found this to be a miracle cure for bad connections. I use it every time I remove and re-plug a connector. I've used it on flashlight batteries, light bulbs, various other things. Not cheap, but always worth a try.

Screenshot 2023-12-18 at 9.34.15 AM.png

[Bicyclops]

More info:
The first time it happened, I verified that it wasn't firing by trying to spark a plug to the valve cover while cranking. Switched plugs and coils and still no joy. The suspect plug and coil worked fine on the adjacent coil connector so I concluded that the shorted wire was the proximate cause of the non firing plug. I cut the wire where I have the harness going forward through the firewall and the coil connector was immediately unshorted. The wire going to the ECU was still shorted. Pulled the seats and firewall to get at the ECU which I had mounted sandwiched between the fuel tank cover/firewall and a sheetmetal firewall I put behind the frame. Referencing my wiring diagram to find the correct connector on the ECU, pulled it and verified the wire running aft was no longer shorted. That tells me that the short had to be within the ECU itself. Short of pulling the cover off that and diddling around with something I had no expertise with, I was kind of stumped so I plugged it back in for no particularly good reason and was surprised that it was not longer shorted. Needless to say I didn't then have a real good feeling about stuff that fixes itself without revealing a cause, but I thought why not give it a try. I spliced the wire back together and checked that it would spark a plug while cranking. It also measured the same ~33 ohms as the others. Started and ran smooth so I reassembled everything and drove it for about 200 miles continuing the engine break in until it suddenly started running rough and afterfiring unburned gas in the exhaust again. I haven't bothered with the spark plug test because the shorted wire at the connector tells the story. I really don't think I can attribute this to a bad wire or connector outside of the ECU.

I found a used ECU on ebay for a couple of hundred bucks that claims to be tested with a video of an engine starting and running (conclusive proof of goodness, right?) and a 30 day warranty. I'm looking for ideas before I pop for it. There are also other ECUs on ebay for "parts only". Is it even feasible to repair one of these by, say soldering in a relay or coil or something? Another choice is a $1700 Haltech plug and play which I would have to have shipped somewhere else because I'm in California.

Ed

[RPGs818SNA]

Just to be sure, I think you are saying that you measured the subject pin on the ECU with the connector removed and found it grounded, thus eliminating the wire to the coil as the culprit. If so, I would carefully inspect the ECU connector for any debris or bent pins, and then would be tempted to take the covers off, blow any dust or debris off, and measure one last time before ordering another ECU. A 30 day warranty will give you plenty of time to be sure the coil pin is fixed and check every other function you can think of to be sure the ECU is working. Good luck Ed.

RPG

[Bicyclops]

Thanks RPG,

Where I've got the ECU mounted it's hard to get a probe onto the right pin. What I measured was actually the connector pin at the spark plug. When I unplugged it from the ECU, open to ground as you would expect. Plugged in, ~33 ohms where it had been closed to ground before the unplug and replug. Just did it again and now 477kohm. Not sure what that means. I did inspect the pins on the ECU via borescope and everything looks fine. At this point, you're right - I probably don't have much to lose by opening the box and taking a look see.

Ed

[RPGs818SNA]

Once you dismount the ECU and can get to the pins, I recommend measuring the resistance to ground for all 4 pins, C21 through C24, to see if one is different. If not, it could still be in the cable connector, and only shows up when the connector is plugged in. If one IS different, well, it's in the ECU. The fact that you are getting varying readings, 0, 33, 477k ohms, suggests that it is a mechanical fault in the ECU connector or the mating cable connector. It should be pretty easy to tell with the ECU dismounted. Hope this is easy to find and resolve. Good luck.
RPG

[Bicyclops]

Once you dismount the ECU and can get to the pins, I recommend measuring the resistance to ground for all 4 pins, C21 through C24, to see if one is different. If not, it could still be in the cable connector, and only shows up when the connector is plugged in. If one IS different, well, it's in the ECU. The fact that you are getting varying readings, 0, 33, 477k ohms, suggests that it is a mechanical fault in the ECU connector or the mating cable connector. It should be pretty easy to tell with the ECU dismounted. Hope this is easy to find and resolve. Good luck.
RPG

I'm more confused than ever. I've got the board on the bench and not a single pin on any connector is electrically connected to the big ground block on the left side of the board that the cover screws down to. Pin C18 would be the one you'd expect the ignition coils to ground to. I read no continuity between that pin, the ground block, or any other pin listed as ground on the wiring diagram. I also read wide open between any listed ground pin and the four pins for the ignition coils. Not sure what to think about that.

Ed

[RPGs818SNA]

Having the ECM on the bench is good progress, Ed.

The cover of the ECM provides shielding to protect the ECM from electromagnetic noise and is grounded to the frame, I believe. The signals to and from the engine need to be relative to the engine ground, so all those ECM grounds are connected to the engine. Since you've disconnected the ECM from the frame ground and the engine grounds, there's no longer a path between them, so they read open.

Are you able to read the resistance between C21 to C24 and C18?

RPG

[Bicyclops]

They are all wide open. No connection at all. What really surprises me is that no pin has any connection to the frame of the ECU. Not one. It is a big square circuit board with one large aluminum block on the left side. The cover has a screw holding it to that block and the cover would be grounded to the frame of the unit by the four screws on the corners. The frame is bolted onto the car so I expected that block to be the central ground point for the whole board. There doesn't seem to be such a thing. There is no continuity between any of the ECU pins which connect to ground on the car. If you look at GE-1 which is the two ground tabs that bolt to the intake manifold, Those hit multiple pins on the ECU - C18, A7,A15, and E1, also D8, D9, A22, C17, and C8 as well as other points in the harness. I've got no continuity between any of these within the disconnected and uncovered ECU. I can't test C21 -C24 to ground because I can't find a ground point to test to. I understand a little about electricity, but I don't think I understand this PCB.

Ed

[RPGs818SNA]

You can use C18 as ground to test C21-C24. It wouldn't hurt to measure the same pins on the wire connector, again using C18 as ground. If you're interested, I can go into more detail as to why the frame ground isn't connected to the engine ground pins. Maybe after we get your engine up and running.
RPG

[Bicyclops]

All four are open to C18. I'll try putting it back in the car and test again tomorrow. Even if it fixes itself and runs well again how could I trust it to not do this again only further from home next time? I've limped home on 3 cylinders twice now.

Ed

[RPGs818SNA]

That's good data to have. Be sure to measure those same pins on the connector looking back toward the coils before you reconnect the ECM. I agree that even if it runs well again, the problem isn't really fixed and will likely reappear when you are as far as possible from home.

[Bicyclops]

Well - she starts and sounds fine. I'm not in a hurry to put her all back together and go for a canyon road jaunt tho. Gonna rain for the next few days. I think I'll take this opportunity to chase that pesky coolant leak and wait for the used ECU to arrive.

Ed

Thanks for advice.

[Bicyclops]

I've got the E-bay ECU installed and loaded my basemap tune on it. Seems to be running fine. Did a canyon road loop with it this morning and I'm happy with the way she pulls. I'm reving all the way up to 4250 now. ;-) Just trying to break her in easy.

Ed