This is for an I35, so the manual is slightly different. I attached those pages below for reference.
Here's a video of the original issue from last month: I am holding it in 2nd gear and as soon as it touches 2,500 RPM, it cuts fuel and bangs the RPM down 200-300.

NOTE: It drives just fine under 2500 RPM. Doesn't seem down on power or stuttering, but I can only give it about 10-15% throttle to keep it under 2500. The P0650 has been there for almost 2 years since I shipped the ECU off to be repaired by Foreign Auto ECU Repair in TX. I had a MAF die in the meantime, but didn't throw a code. I found in the EC manual that P0650 paired with any other code will cause this "limp mode" of 2500RPM

Following the manual, I tested for spec voltage hitting pin 66 on the gauge cluster connector. .5V in OFF and 13V+ when running. CHECK
Proceed to 3: Probe pin 45 on gauge connector and 33 on ECM. There is continuity, but resistance is something like 120,000 ohms CHECK, GO TO 6, intermittent issue
Checked all the grounds I could see and everything looks fine. No broken/damaged wires by the timing chain cover ground point. I am no expert digging deeper into this, so ¯\_(ツ)_/¯

So I go back to step 5 and I run into this when testing:
I am supposed to check continuity on 45 and 66 of the gauge cluster. See below pictures for the result:


probing 45 positive 800,000 ohms

probing 66 positive 140,000 ohms

closeup of gauge cluster test says 66 probed positive is supposed to not be conductive

full manual section for I35


The I35 manual has this extra step for testing the gauge cluster. This section 5 is not present for the 02-03 Maxima.
Anyways, does this reading tell me that the gauge cluster is the cause of the problem? It says that 66 probed positive should not have continuity.
Thanks for any help

The evap solenoid is, I think, the same solenoid type as the VIAS solenoids. They're easy to pull apart and test and clean/reuse. If you have a sticky valve, should be an easy fix. Check all hoses and harness of course, download the FSM and pull up the workflow for the codes.

I updated the original post to show what I found. The limp mode is caused by having a P0650 and ANY OTHER code thrown in tandem. Since the ECU cannot illuminate the CEL properly, it puts you in limp mode since you are seemingly unaware of any other codes being present.

I still need to address the evap when I can, but this 0650 code will forever cause limp mode if I don't fix it first.

How do you know your ECU is malfunctioning this way?

​​​​​​​That's the point of the DTC P0650, is it not? That it cannot detect or has a circuit issue with the Check Engine light? Outside of testing for this intermittent issue, the only conclusion I can come to is that the ECU was improperly repaired or that the instrument cluster has a circuit issue. I tested the instrument cluster as the manual states, but I'm not sure what to make of the result. You can see that there is clearly continuity between those contacts. Albeit extremely high resistance measurements.

And it doesn't seem intermittent. That code has been there since day 1 that I got the ECU back from repair. I've cleared it multiple times and it comes back very quickly, if not instantly.

I can clear the codes and rev it out to 6,000, so this P0650 certainly looks to be my culprit. Any other write-ups around here that discuss how to diagnose this code? Preferable with pictures of these junction boxes and electrical connectors.

Bingo. You just solved your own problem. I have a few ECU's for my '03. Dime a dozen on ebay (think I paid $35 shipped for my last one?). Toss that ECU (or disassemble it and see if anything obvious can be reversed/repaired) and get a new one. Program it with NDSII and you're gtg. And I'd make whoever "repaired" your ECU pay for it.

That's been my thought, too, but I had to double check all my other possibilities. What about the readings for continuity on the CEL circuit? That is for-sure not a fault of the gauge cluster? When the ECU fried, I replaced the throttle body to be safe. Unplugged the motor mounts. When I got the ECU back, my MAF had also failed. I thought it odd that the ECU surge could take out a MAF. Could the same be possible for the CEL circuit?

Is the I35 ECU the same as the Max, though? I know the pins and primary harness are the same, but the secondary harnesses are often different.

Although it is certainly within the realm of possibility, I just think it would be highly unlikely that there was an electrical malfunction in the cluster. I guess the only way to rule it out would be to inspect/repair the cluster or get a spare from a yard or ebay and test it. That's what's great about these cars, the abundance and availability of cheap, cross-compatible parts among models.

What are you using to pull and clear codes?

Cheapo OBD2 puller from Amazon. Maybe $30.

I35 parts are unfortunately not as abundant as Maxima, but it's nothing like my old 95 Mitsubishi Diamante. Finding parts for that was a needle in a haystack. The nearest yard that I can look for inventory is 2 hours away. I'm going to a fancy dinner at Bern's steakhouse a few miles from a big yard with 3 of these cars. Maybe if I wear gloves and bring a change of clothes, I can slip over there before dinner. Otherwise, $130 is about the best price I can find and the mileage is only off by 4K without having to reprogram it. I can get a cluster for testing for $25 if I can find a yard nearby.

Let us know. I'm in Tampa if I can help. There are four LKQ yards in the Tampa area, all of which I know quite well.

I just did a quick check at Tampa, Clearwater and Largo LKQ yards and found six 2002-2003 I35's in nice shape. I could run parts for you if you needed. There's also six I35 clusters on ebay at the top of the search results I'm sure you could bargain down to $100 or so.

I ended up going to Tampa LKQ and grabbed 2 clusters just in case one had a problem. Both clusters did not solve the CEL issue. They also all showed similar resistance levels before I installed them. I wanted to be sure to check that before testing them.

The EVAP problem fixed itself and is no longer throwing a code. So no more limp mode for now. Small victory there. With either of these clusters plugged in, I can sit there in the car with it running or just "on", clear the codes, and the CEL never even shuts off. That stupid light is ALWAYS ON. I cleared the codes, drove 2-3 miles and was constantly unplugging and plugging in the OBD2 to scan for codes. It would come up clear several times with no stored or pending codes, but that dumb light just stays on.

Does anyone have an idea why?

I'll end up tossing the clusters on eBay to recoup some of the almost 4 hours round trip. The yard was HOT today. 96 degrees and humid. Nasty

Here's the clusters installed and the light is still on. One picture is 5 seconds after I cleared the codes and the other is after I installed the other cluster and took it for a drive. Code came back, but CEL stayed illuminated.

You need a Better Scan Tool.. Your issue could be else where.. Other modules in this car can request the CEL, that You cannot see. Has there been any recent engine work done to this car?

If you suspect a wiring issue, run a overlay wire from that circuit on the back of the cluster straight to the ecu and retest. Also thoroughly inspect and chcek your maf reading at the maf with the dmm.

Time for a new ECU.

Do you know what system could send a CEL? I was unaware of that. I thought the other systems would illuminate diffent light EX: (TCS light=Transmission, ABS light=Brakes, Battery light=Alternator) Could you explain this overlay wire test you mention?
The MAF tested fine. The voltage climb was steady for the signal wire. I'll double check this afternoon.

This is a possibility. I mentioned to the ECU repair shop that this CEL has been persistent since repaired over a year ago. It never goes out, as I said above. Even resetting the codes, it stays illuminated. I told them I would do my due diligence to experiment with the instrument cluster since the EC manual describes the cluster as a possible fault. Just proved that was fine yesterday, so it's either another module like Reality mentioned above or the ECU is at fault. The shop said they would source another ECU, I send mine in, they swap the security chip to donor ECU and send back. Shouldn't have to program the key. Would I need to do a re-learn in this situation, though?


When the ECU (partially) fried last year, we could smell it. (here is the original post)The car drove fine except for some intermittent surging/lack of power (related to MAF that failed later) and CEL did not illuminate. No other codes. I unplugged the motor mounts at that time. Car drove fine for several weeks/months until it started running like total crap. Sent out the ECU thinking the short had finally taken out a critical module, replaced throttle body to be safe, and left motor mounts unplugged. ECU gets back. CEL P0650 is now present as soon as plugged in from repair. Still runs like absolute garbage. Found the MAF sensor at fault. Replaced it and car ran fine, but still that P0650 MIL circuit code hangs around. Drive it fine for a year until knock sensor code pops up. Car goes into limp mode. Replace knock sensor and crappy brittle harness. Everything fine for about 2 months until evap code pops. Car is in limp mode again. This is when I find in the manual that P0650 plus any code causes limp mode since ECU can't tell the driver there is an issue.

To update this:
I sent the ECU for repair a seond time when it burned up. They swapped my security chip over to a donor board so no reprogramming.

I get the ECU back and it instantly fries again. The car still runs fine, but it smoked part of the ECU and the CEL fizzles on/off. Same P0650.

So I void the warranty and take a look at what fried. Followed the trace back to the pinout


this pin trace was burnt on the board


PIN 4 ENGINE MOUNTS





So I go back and check the motor mounts I thought I unplugged. I unplugged the front, but unplugged god-only-knows what else right next to it. The rear motor mount is still plugged in. I'll check it for continuity tomorrow and see if there is a short.

Hopefully give a resolution to anyone else that visits this years from now.