Those voltages look normal, 0.1 of a difference won't matter in this case.

Personally I would replace/cut/splice the wire for that coil pack to the ECM before looking into a replacement ECM. Hot insulation gets squishy, might just be a conductor issue. Just because it has continuity doesn't mean the insulation isn't nicked somewhere and starts making contact against chassis when it's hot and squishy. If it's grounding itself out it's going to zero out the voltage.

You could do a temporary wire by just running it up and out of the cabin air filter. Just leave the cover off and run the wire up through that giant opening into the battery box, leave the battery box lid off.

 

I ran some more tests today.

I hooked up an Arduino to the ECM harness connector to simulate signals that the ECM may be providing to the coil packs. Basically, it just confirms if the wiring is providing the correct voltages and not shorting out anywhere. (See Pics Below)




For the most part, the signals checked out...

I can very confidently say that I do not think there are any obvious shorts and/or conductivity issues.

I sent both constant voltage signals (5V) and PWM signals to coil packs 1,3, & 5 and saw that the measured voltages from the connectors were more or less the same.
I tugged on the wires pretty decently for coilpack 3 and did not see anything change.
The signal for coilpack 3 was also never shorted to ground during the tests.

This suggests ECM issues, assuming everything checked out as I said...

However, what I did notice was that sometimes, the signal for coilpack 3 would fluctuate between 0.00-0.03V lower than the other two coils. This is not a very significant difference, however I believe this signal for the coil pack is pretty sensitive to changes in nominal voltage. In fact, when starting the car after I could see that the live signal for coilpack 3 was between 0.17-0.20V (so not a dead misfire after ECM reset, like usual) which is precisely 0.00-0.03V below the expected nominal voltage of 0.2V that I measure on the other coils.

By the time I had conducted my test, it was already starting to cool off in the afternoon and I have noted that high temperatures seem to make the car run worse with misfires appearing much quicker and seeing the voltage of coilpack 3 go down to 0.00V at times. So, I think I would like to run this test once more at high noon when it is very hot just to be super sure that the heat isn't affecting the harness in such a way to knock down that voltage. I am not quite sure how it would do this if it is not being shorted out besides the potential of bad contacts or bad ground (but I would also assume a bad ground to affect more than just one coilpack...)

If the heat is not making significant influence on the measured simulated signal, I think I can finally for sure say that this issue resides 100% in the ECM. However, if the measured signal drops even further, I will be replacing the wire in the harness to see if it can take care of the issue.

 

Use an infrared thermometer and a can of compressed air. You can turn the can upside down and shoot liquid CO2 compressed air at the ECM cover until it's frosted to cool it off and check it's temp with the IR thermometer. Then see if the problem magically disappears.

I would definitely leave the cover on the ECM for this though, freezing the board with compressed air causes too much expansion/contraction issues and you can peel components off a board pretty quick, that cover on the ECM though has quite a bit of an air gap between it and the PCB.

You're probably going to need to get a custom license plate "CEL P0303" though, or replace the ECM.

OR BOTH!

 

Hahaha I like that license plate idea, might have to steal it.

I actually think you were on to something with the wire conductivity issues when it is hot. If I wiggle the wires enough, it does seem to make a difference. I think a new spliced in wire might make the difference that allows for consistent quality conductivity. The voltage seems to only be dropping intermittently and not much at that, but it is definitely making a difference in the misfires. I just would rather not run a janky looking wire through my battery lid opening or cut my harness, but it looks like that is where I am headed sadly

Also just FYI I looked into getting the ECM tested and the quote I got was more than I would have to spend for a brand new ECM from an aftermarket supplier so that's out for me. If I was sure it was the ECM I would replace it, but I am now starting to believe its the intermittent conductivity issues of that single wire like we discussed. I'll try and update when I splice the harness and drive a decent amount of miles without problems so there is some closure.

 

Yeah that janky wire is just for testing purposes, if that actually fixes the problem then just add a wire to the factory harness and cable tie/tape it in to look good.

You can even use that spiral wire loom that's for home stereos and stuff, it holds up surprisingly well in automotive use, just get the black colored one.

Don't use the corrugated wire loom, I hate that stuff...

Don't use the braided wire loom that has the velcro to hold it together, doesn't hold up in automotive use.

Use the stuff that looks like this, if you use a black GPT wire it will blend in perfectly. It takes a little longer to install the loom compared to other options but it's worth it imo. Plus it's REALLY cheap and all the brands seem to be the same quality, doesn't matter if you get the stuff from Home Depot, Harbor Freight, online, Walmart, or high end brand stuff like Gardner Bender. It's all made from polyethylene which is fuel and oil resistant.

 

I was thinking to just use some TXL wire (I think 18 gauge is what's on the stock harness?) https://www.painlessperformance.com/wc/lv.php?sl=71858. Probably a temporary solution and test that thoroughly (driving and all for several miles), then if it seems fine redo it and make it nice. It's an intermittent issue so I need to be SURE that it is fixing the problem before putting in all the work to make it look nice.

Either way this is going to be a decent amount of work. I am still deciding whether to try and remove the current wire outright from the connectors, or pull-back the bundling/ heat shielding material, splice it, then solder or crimp a new wire in. Not looking forward to this... but a new harness is far more expensive.

 

TXL is fine, it's just a little more expensive.

 

I was looking into replacing the wire today and realized it would actually be 2 wires, since the signal wire for coils 1 & 3 both run through the condenser. I was trying to see if I could get the misfire to appear and then test each wire individually to hopefully isolate a bad one.

When I had the car running sure enough a misfire appeared, but it was cyl 4!!! Confirmed misfire on cylinder 4. This is beyond me at this point... unless that was just a complete coincidence and now there is a separate issue with cylinder 4, something bigger is going on. Maybe the ECM after all?

I can't believe this its becoming such a headache to work on. I am losing motivation fast.

 

Have you replaced that condenser? It's like $10, Nissan part number 28351-89902

 

I think I've actually been getting mixed up on what the condenser was whoops. The signal wires for coil pack harness connectors 1 & 3 actually just pass through another connector, not the condenser. I am not really sure replacing it will solve anything but for 10$ might be worth a shot. If there was an issue with it, all of the coilpacks should have their power supply effected.

I also checked the signals at coil 4 harness connector and sure enough, there was adequate power on terminal 3, continuity to ground on terminal 2, and no signal on terminal 1. Same issue as cylinder 3. What was interesting is that for one, terminal 2 actual measured a constant voltage of 0.1V, but still had continuity to ground. And terminal 1 had 0 volts most of the time so I assumed it was permanently grounded, but it actually had resistance to ground, not continuity. Even cylinder 3 would have continuity to ground on the signal wire as the ECM pulsed it. I am still not quite sure what this implies.