'12 Impala - Condeming Bad ECM

Customer brought in a 2012 Chevy Impala. Complaint was limp mode and service stabilitrack. Customer has installed carquest throttle body, and pedal.

Pulled codes - P0697 - 5v ref Circuit 3 symptom 00 - MAP and APP 2 sensor are on this circuit. Backprobed both sensors on this circuit KOEO, and had 5v. Checked scan data and also have 5v on that circuit.

Clear codes - went for test drive - about 5 miles in - Limp mode and stabilitrack returned - Pulled codes P2138 - backprobed APP 1 & 2 @ pedal & at EC. I never went full throttle just drove leisurely. 

• APP 1 @ 0% 0.98 @ 100% 4.18v (both @ pedal and @ ecm)
• APP 2 @ 0% 0.48v @ 100% 2.1v (both @ pedal and @ ecm)
• Looking at Data pids on scanner - APP 1 @ 0% 0.98v @ 100% 3.61v (this is to me is where the discrepancy is coming from on P2138)
• APP 2 - 0% 0.49v @ 100% 2.1v

with the error only coming on at 100% I assumed if i dont go 100% throttle then the code shouldnt return. I went another test drive, this time the vehicle went a few miles further than last time, at one point my scanner randomly lost comms with ECM. Nothing happened to the vehicle it was driving normally. I returned vehicle to shop, and it never went into limp mode.

I type up report about intermittent issue, I need mileage, I do KOEO, and the vehicle went back into limp mode - pulled code and have P0697 but scan data and sensors all have 5v.

I believe I am looking at an internal fault on the ECM, but am not 1400 dollars confident that its the issue. This will be the first PCM I have condemned being a year into the business. Something tells me the dead give away is the data pid not match what im seeing coming out of the pedal. 

I was advised to do a heat soak test, hood closed, and let it idle and see if I lose comms with the vehicle again. If I lose coms again then its most likely a bad PCM.

Any advice here on what test I could perform on the PCM?

Definitely understand your hesitation to throw an ECM at this one, especially with the problem being intermittent.

I think I'd tend more towards a harness issue than an ECM issue, based solely on prior experience with these vehicles. These ECM's do go bad in several ways, but I haven't seen one have an internal 5V ref issue.

Next time you take a crack at this one, don't even drive it. Key on and backprobe the 5V reference at the MAP sensor. Get the multimeter somewhere you can see it. Then gently rock the vehicle forwards and backwards on the parking pawl. The idea is to get the engine to torque back and forth in its mounts, (hopefully) getting the harness to rub and show a loss of 5V.

If you can reproduce the loss of 5V, you're onto something. Like SD says, heat and vibration. Keeping the engine cold will allow you to gently wiggle any suspect harness areas without getting burned.

I grabbed a harness diagram that includes X111, which is the one of the engine to instrument panel connectors. X111 carries the APP wires, among others. X111 is labeled as '9' in the picture:

 

The diagram can give you a rough idea of how the harness routes and possible suspect areas.

Tyler thanks for the advice. I have already checked the 5v ref circuit for a short to ground and didnt see anything. I just spent another hour checking it again and found nothing. I rocked the vehicle on parking pawl as you described to rock the engine and I did not see any drop outs. I was pushing on the harness to try and ground it out and got nothing.

I pulled the X11 connector apart to check for water intrusion and the connector was clean.

I did use my scope on this as well over a 1s time base so if it dropped out I would have seen it.

I have already tried heat soaking the engine for an hour, and got nothing.

One weird thing I noticed today, I cleared the codes yesterday, parked it, I got in it at lunch KOEO, and the engine went back into limp mode, and the same code, P0697.

I think I am going to scope the App 2 5v ref signal and drive around the vehicle and see if I see anything out of ordinary.

My problem is I literally shook down the entire harness already so I am not even sure how I will handle the situation if I see a blip on the scope.

Here is the only blip I saw in the 5v ref circuit 3 was at startup. That hash is at startup.