FlyingShawn

Ok, ready to enter my nightmare?

Bought my first P-car a few months ago: a 2002 C4 Cab with 98.5k miles. Car came with an intermittent #5 misfire which the previous owner and indy mechanic were highly confident was the result of a bad DME (they’d already tried replacing the injectors and testing the harness). Previous owner had purchased a used DME, but wasn’t able to get it programmed because the nearest dealer was ~4 hours away and decided to sell the car instead. Knowing I’d have to replace the computer also resulted in a pretty good discount when negotiating the vehicle price.

Shipped the car home and as expected, #5 cylinder fuel injector electrical fault (P205) and misfire (P305) intermittent codes that got worse as the engine warmed up.

Step one: pulled the DME and immobilizer and shipped them down to Specialized ECU Repair in Florida for diagnostics. They insisted my ECU was perfectly fine based on their bench tests and claimed it ran fine in their test car (although I don’t know how extensively they tested it considering my experience was that the problem was much less frequent when the engine was cold). They offered to program and ship me one of their “rebuilt” ECUs for $1000, but if it wasn’t a fix there was a 30% restocking fee (are you kidding me?!). I opted to have them send my original DME back.

Next, thinking Specialized ECU might be right and my DME was fine after all, I took the car to the local indy P-car expert and we dove deep on trying to test everything else we could think of:
-Swap #5 and #6 fuel injectors? Check, problem stayed at cylinder 5 (injectors 4, 5, and 6 were all replaced in December of 2017, injectors 1-3 are older).
-Resistance on the fuel injectors? Check, 12.5ohms on all six injectors.
-Voltage at the injectors remains constant, no voltage drop to the DME on either the ground or positive sides
-Wiring harness? Electrically checked at the DME connector and at the components themselves to eliminate the harness.
-Connectors between the harness and the fuel injectors? Checked.
-Clean the Mass Air Flow sensor (a possible fix we discovered on a Rennlist thread where a user had weird misfires and fixed them by replacing his MAF, but he noted an improvement simply by cleaning it)? No change
-Bypass the usual wire and power the injector directly from the battery? Check.
-Check for AC current leaking out of the alternator? Check.
...All with no fix.

At this point we decided to go back to the “bad ECU” theory. I ordered a used DME off eBay from a car that’d been in a front-end collision (figured the donor ECU should be fine considering its placement behind the rear seats) for $120 (definitely beats $3k for a DME directly from Porsche!) and had our local Porsche dealership program it. Here’s where it gets weirder…

On ECU “B” (for the sake of clarity I’ll call the original one “A” and this one “B”) the problems related to cylinder #5 went away! Except… now cylinder #1, which had never previously given us any problems at all, was throwing P201 (Injector Fault) and P301 (#1 cylinder misfire) codes worse than we’d ever had with ECU A and cylinder 5! They were so bad that as the engine warmed up the computer gave up on cylinder #1 altogether and stopped sending signals to that fuel injector (we hooked up noid lights to the #1 and #5 fuel injector circuits to monitor the signals and the #1 eventually just went out altogether)!!

We checked the resistance in the wiring harness to the #1 injector: no problems there. Then we swapped back to ECU A: sure enough, cylinder #5 started misfiring and cylinder #1 worked great. At this point we were thinking that, unlikely as it seems, we’d replaced a DME with a bad #5 injector driver with a DME with a bad #1 injector driver.

We returned ECU B to eBay, ordered ECU "C" from another donor car, and once again had it programmed by the local dealership. Installed it in the car and… Exactly the same behavior as ECU B: P201 and P301 codes eventually resulting in a full shutdown of the #1 injector and no problems at all in cylinder #5. What are the odds that, on top of getting two bad computers in a row to replace our original one, both ECUs B and C would have exactly the same problem on the same cylinder???

We’re stumped. The only thing we can think of that we haven’t tried is to swap the #1 and #2 injectors to definitively rule out the #1 injector itself (which is a long shot at best considering it runs fine with ECU A).

What else could be causing this? We’ve scoured the forums here and over at Pelican and we can’t find anything remotely similar or that we haven’t already tried, especially not anything that would make sense of the misbehaving cylinders reliably swapping between #1 and #5 depending on which computer is installed.

Any ideas or advice would be greatly appreciated!!!

DBJoe996

And this is why the technicians get paid the big bucks. What's your fuel pressure at the rail? How old are your coils and plugs? Maybe start with ECU "A" and this time swap the coil packs around.

808Bill

Coils would have been my first thought. I'm surprised no one has even mentioned this...?
Make sure your coil wire on #5 is fully seated with a click. Check the wiring for damage, mice love to chew on wires.

Ahsai

Do you know if the injectors are genuine Porsche parts and the right parts for your MY?

Sounds like an electrical problem to me. The P0201 and P0205 are complaining that the DME has detected OPEN circuits on those injectors.
The injectors get +12v from the same junction as the coils and there are no other connectors involves (except the injector connectors). However, the injectors gets the ground from the DME via the x59/1 connector 0- one of the two big round connectors on the right side of the engine bay. I believe it's the top one but please verify yourself..
You can spray the pins on both sides with Deoxit and check again. The pins are 5, 6, 7, 19, 20, 21 for cyl #1-6, respectively.

Imo000

If you can do most of the diagnosis yourself by start swapping out existing parts then you could save a bunch of money. At this point everything points to all 3 DMEs being bad but as you stated, the chances are low that that is actually what's happening. I would start a new systematic diagnosis. Do what you did first and then keep going by swapping out and testing with other injectors and/or the DME. Also would have a close look at the coils and plugs but only do it in a systematic way or else you'll loose site of what the results are. Also, if the spark plugs are put might as well put a borescope down the cylinders and have quick look.

FlyingShawn

I'm away from home for work at the moment, so I won't be able to check the service records until tomorrow or Thursday, but I know all new plugs were installed this December and I believe the coils were also replaced within the last couple of years. I'm also going to check with my mechanic to see what swapping around he's already tried that I may have forgotten to write earlier. The thing we keep coming back to is "if the problem was _____component, why would it stop/start misbehaving depending on which DME is installed?"

In the meantime, please keep the ideas flowing!

FlyingShawn

Just heard back from my mechanic on one of my questions. According to his research, there are three conditions that'll trigger a P020_ fault:
1. Incorrect resistance in the injector (all six are identical at 12.5ohm in our testing).
2. No voltage signal when the DME reads it through the injector coil winding
3. A short-to-ground causing the injector to stay open (which we should have detected in all of our harness testing!)

I'll update all of you on what happens when we swap the #1 and #2 injectors. If the problem doesn't follow the injector, I'm wondering if we should try swapping the #1 and #2 coils (Idea being that the #5 cylinder problems might be the result of ECU A being bad and the #1 cylinder problems might be legitimate issues that the bad ECU A wasn't detecting but ECUs B and C did detect. It's a stretch, but not any more of a stretch than the odds of three bad DMEs and 2/3 of them having the same fault).

jumpy chunky

This may be a long shot, but might be easier than swapping injectors.
The shade tree version of trouble shooting would be installing any one of the three ECUs with a same M-Y
996 and checking against a working ECU and car.
Would your Indy have a source?
Anything from eBay would be suspect. Could be where previous owner picked his up.

FlyingShawn

Testing an ECU in another car (even same model year and version) would require either reprogramming the ECU to match the other car ($$$ and can only be done by a dealer or specialist like that outfit in FL) or would require swapping all three of the ECU, immobilizer, and ignition switch between cars to have a matched set. Even considering that, it'd be something we'd consider if we had access to another car and permission to test it, but my indy doesn't have any 996 clients that he thinks would be willing to help like that.

ECU A is original to the car, B and C are the ones I obtained from eBay (from a different seller than the one the previous owner obtained the one he couldn't get programmed).

FlyingShawn

Just got back from visiting the shop. He got bogged down on another car today, so he hasn't had a chance to swap the #1 and #2 injectors yet, but we spent a little time talking about what to try next if that doesn't work.

One of the things I kept circling back to as a result of this thread is the possibility that there might be an issue with the coils. Despite the many, many problems he's fixed with coils over time, he can't see any way how a bad coil would cause these faults because of the three P020_ trigger conditions I mentioned earlier. As @Ahsai mentioned the coils and injectors all share the same +12v source, but the P020_ fault is being sensed by the DME on the ground side of the circuit and during our original testing with ECU A we powered the #5 injector straight from the battery to bypass everything on that +12v circuit.

Nevertheless, despite everything seeming to point to three bad DMEs in a row, I just can't wrap my head around the odds of that. So, after we try swapping the #1 and #2 injectors on ECU C, the next thing we're going to try is swapping the #1 and #2 coils to see if the problem chases a particular coil.

Two other things of note:
1. Remember how I said I was pretty sure the coils had all been replaced recently? Well.... we're not so sure. I've got an invoice from the shop in Montana that'd been taking care of the car for the previous two owners that claims all six coils were replaced in July of 2016, but the coils on the car don't match the part numbers on the invoice! The invoice is for 99760210402, the coils on car are Berus but a different part number (sorry, don't have it written). My mechanic thinks that either these are the original coils and they were never replaced or these coils were old stock and the MT shop made a mistake when writing the invoice.
2. On a related note, I discovered that I'd omitted something from the original diagnostic regimen: Before we started replacing ECUs we replaced the #5 coil and that new coil is still on #5 even though it hasn't seemed to make any difference.

Ahsai

Exactly. When they swap the injectors, they should also note the p/n and cross check. Your coils may be due but I highly doubt they will fix your injector error codes.
Also, don't forget my comments about the x59 connectors.

FlyingShawn

Excellent idea to check the injector part numbers, especially after this business with the coils not matching the paperwork!

And, forgive me, but I'm not enough of a car guy to recognize the acronyms: FRA, RKAT, AFR?

Ahsai

Never mind those FRA comment. i got your case mixed up with another thread

Porschetech3

Sounds like you and your Tech are now going in circles. Y'all seem to be competent, but may be too emotionally attached to the problem to be objective. Might want to step back, take a breath a fresh air, and try a new approach

Instead of running until you get the code, then doing the usual fault tree diagnosis. Look at the overall picture.

1. How intermittent is it 0% to 100%?
2..When you say it gets worse when hot, how much worse scale of 0- 10 ?
3. Can you manipulate it ? Make it better or worse by wiggling things, tapping on things?
4. When the problem starts, does it stay? If not , what makes it go away?
5. Has the engine ever been out of the car?

Think of the DME as Big Smart Grounding Device, it controls injectors and coils and actuators by grounding them. Yes it has power to it, but does all the high current through ground. Make sure the ground points are good, all of them. Check them for tightness, then loosen and clean them and re-torque. GP9 and GP8 are the main ones for the DME, but also check the ground-strap to engine, and the battery to chassis cable.

It will all make sense once you find the problem, and the really tough ones usually end up being something simple, but hard to put your finger on.

FlyingShawn

Haha, yes, going in circles is definitely how it feels!! A code like P020_ is supposed to be related to one of three components (DME, Injector, and the wiring between them, particularly the ground), so when we've independently eliminated all three it's definitely vexing. There have been a number of ~week long pauses in the work for that very reason or because we've been waiting on something (parts, programming, etc), but no such clarity has emerged.

As for your questions... (for the sake of discussion let's assume all of these are answered in relation to ECU A and cylinder #5 unless noted because that's where I have more data)

-On a cold day, say<~30F, it is much more rare, sometimes seeming to be close to 0% until the engine really warms up. After that it's a noticeable, fairly constant low-grade misfire.
-On a warmer day, say 50-65F, a few occasional (20%) misfires are noted within the first few minutes after startup. Once the engine warms up it gets to the point of seeming to misfire on nearly every rotation and each misfire is more noticeable (not huge, but not something that can be mistaken for road imperfections like the infrequent misfires on a lukewarm engine on a cold day).
-On ECUs B and C where the problem occurs with cylinder #1 instead of #5, the misfires start within a couple of minutes of startup even on a cool day and by the time the engine warms up power is cut to the #1 injector altogether and it becomes a five cylinder engine (this is evidenced by the ride and by the noid light extinguishing altogether instead of flickering with DME signals to the injector).

The problem gets worse at higher engine speeds (as in: when the engine is still warming up and it's in the 20% misfire range at 2000-3000rpm, it can reliably be bumped up to 70%+ by throttling up to 3500-4500+rpm).
Beyond that, except for moving the problem by switching ECUs, we have not noted any variation even when changing the power source to the injector or bypassing the wire from the DME to the injector.
As a caveat on the above answer: since the problem has required a warmer engine, nearly all of our testing has been road testing rather than running the car stationary in the shop, so there has been very little wiggling and tapping tried.

Once it starts, it can be somewhat calmed briefly by coasting the car (fuel cutout to all injectors) and then running at very low load, but that doesn't last long and comes back quickly when even moderate power is applied.

IMS was done at 93k miles (August 2016), so it would have been dropped then, right? Sorry, I'm new to P-cars and I'm still learning what requires a drop and what doesn't. Beyond that, no, I'm not seeing anything in the service records I have going back to 2013/66k miles that might imply an engine drop.

Sounds like a handful of worthwhile and relatively-easy items to test! Thanks!