944 S2 Stalling/Hesitating when wet

[neilschelly]

I have been having an issue for some time now with my relatively-new-to-me 1989 944 S2 with ~215k miles. The car was undriveable didn't run at 208k when I acquired the car, so a lot has been refreshed, and it runs great on dry days now. When it's wet out, I have the following symptoms, and I'd love the community help troubleshooting.

* The car will lose power suddenly when driving. It won't happen right away, but it will happen during a rainy drive within 20-25 minutes. It will also linger with the car the next day for a bit, even if it's sunny and warm.
* .It can stall out, but I can usually save it if I clutch in and slam on the accelerator to rev the motor. It will also generally always start right up, and it will idle fine. The power only really cuts out when the engine is under load. Sometimes that happens immediately if you haven't revved the engine high enough before lifting the clutch pedal. When I've had to get home and not be stuck, I've come home with a burnt clutch smell because clutch drops are reliable ways to get going again.
* It will be less likely to happen when the the tachometer is kept higher, like dropping a gear to keep the tach over 4-5k, but that is not a given.
* I have the Focus9 DME relay with the diagnostic LEDs and the pre-prime fuel pump timer. That might explain why it always starts up right away because fuel is always available at startup.
* I have witnessed the Focus9 DME relay lights when this is happening. The lights for fuel pump power and fuel pump control signal from the ECU blink when the engine hesitates.
* I have replaced the DME with a jumper to force the fuel pump to stay on. This keeps the car from stalling out, but it will still hesitate. occasionally. If I just keep my foot on the gas, it will continue again.

Things I'm confident of...
* I have reliable power at 14V everywhere I can test.
* I have cleaned every ground I can find, including the ones behind the headlights, the ones all around the engine, the ones back by the fuel pump, the ones near the fuse box (above and below), near the battery, etc.
* The cam position sensor is new, as well as the connector that disintegrated when I first touched it.
* All ignition wires, distributor, the ignition coil, and the "other ignition coil" that is behind the left headlight are new/good.
* All fuel lines and filters (outside and inside the tank) and the pump are recently replaced.
* The battery is good and reasonably new.
* All the ignition components under the hood are staying dry, even when I drive without the engine trays installed. The same for the cam position sensor, engine reference sensor, and the fuel injector harnesses.
* I idled the car in my driveway for 30-40 minutes and just sprayed everything under the hood with a hose for that hole time. It never hesitated or skipped or anything. Every connector and ground connection got a straight blast of hose water jetted at it with no impact, but it wouldn't drive well after that. Whatever water had gotten _somewhere_ needed to be gone before the car drove well enough again.
* The dashboard under the driver's side is staying dry. There's no water getting into the wiring under the fuse box, and the driver's side floor is staying dry (except for whatever comes from my shoes). The floor and ECU under the passenger side is also staying dry.
* The distributor is staying dry. The S2 doesn't have that extra seal behind it that could allow water in.

Top Theories:
* Water is getting _somewhere_. It sounds obvious, but that explains why it doesn't happen right away, and it would explain why the issue may continue the next day with nice weather for a little while.
* When it happens, the ECU is interrupted. That would explain the ECU signal for the fuel pump cutting out. That would explain the engine still hesitating when the fuel pump jumper is forcing the fuel pump on permanently.
* The tray under the windshield over the HVAC fans has a few cracks in it. Maybe water is getting under that tray and grounding something out?
* Either the engine reference sensor or the cam position sensor (or the signal from them) is getting interrupted. That would cause the ECU (probably?) to cut the fuel pump signal and ignition.

I'm running out of ideas. Has anyone else ever run into anything like this?

[neilschelly]

Originally Posted by gb944

Obviously, the parts cannon approach is not working..

You can lose the unhelpful attitude. This is a car that sat for 6 years outside and parts rotted. Parts that I listed as new/replaced have been replaced needed replacing to solve one problem or another.

I have no idea what a Bosch Hammer or Durametric are. I'll look into them.
-N

[djnolan]

You may be on the right track with water getting in the dash. Check to see if the battery tray is corroded and letting water in on the passenger side where the dme and harness is located. This is a common problem, and there are other ways for water to get in on that side too.

[neilschelly]

Originally Posted by djnolan

You may be on the right track with water getting in the dash. Check to see if the battery tray is corroded and letting water in on the passenger side where the dme and harness is located. This is a common problem, and there are other ways for water to get in on that side too.

I meant to include that all drainage pathways are clear. All 4 sunroof drains and the ones by the battery and fusebox are clear and flow easily. There's no corrosion under the battery, and there's no obvious way I can see for water to get from there to the dash.

Further, there's no sign of moisture under the dash on the driver's or passenger's side when the behavior is occurring. I've got all the under-dash covers off on both sides so I could literally drive around until it starts happening and confirm that everything is dry.
-N

[djnolan]

if the battery tray has a leak it will drip right on to the DME. One thing to look for is a rusty dme housing. Also water can get in around the windshield seal. The hvac fan housing can also leak where it seals to the body. Are the crank sensor connectors good and tight? Was the alternator belt slipping when it got wet? did you try misting with a squirt bottle on the ignition cables when it was running to try to detect a miss?

regarding the alternator if it was raining and you had all of the accessories running and the alternator belt was slipping, that might feel like a miss. one way to test the alternator belt is to idle the engine and gradually turn on accessories (lights, A/C, rr defogger, etc) to see if the belt squeals. you can even spray a little water on it.

[neilschelly]

The battery tray has no leaks I can find. The DME is dry, and there's no rust on it. The windshield was replaced about a year ago, so I doubt the seal is failing already, but I guess it might have been replaced badly.

The HVAC fan housing is a strong theory right now. I will probably remove that plastic cowl cover soon and look for signs of water intrusion under there.

The crank reference sensor and the cam sensor are both tight. The cam sensor has been recently replaced, and the connector was replaced at that time because the old one crumbled apart when I replaced it.

I have seen no indication of a leaking belt, and there's no other signs of a low voltage situation like dimmin lights, HVAC fans slowing, voltage meter being extra erratic, etc. I sprayed the whole engine bay, including the ignition cables with a hose and never got a stumble.
-N

[neilschelly]

I found my oil pressure sender was disconnected from both wires. One wire broke from it's ring terminal, and the other one just pulled itself off the spade connector. I'm not sure if those were supposed to be two different connection mechanisms originally or if that's been repaired before. Either way, both wires are connected again. I really doubt that this is the electrical ground issue I'm looking for because I don't see how either of these could have been causing issues only when wet, but it's fixed now.

I also found the oxygen sensor wire wasn't properly supported. The metal clip wasn't attached to the stud over the torque tube, but it is now. Again, that isn't necessarily a source of the problem, but it does suggest I should take a closer inspection at the cable to the O2 sensor when I get a chance. It's clearly been replaced before, and it's got a lot of wiring splices along its path wrapped in just electrical tape. Again, no obvious culprits, but it certainly doesn't look great.

I have another theory to investigate too. I have the alarm control unit near the DME (928.637.104.05). I don't have specific reasons to suspect this, but I'm pretty sure interrupting the voltage signal to the DME relay is the way this implements it's immobilizer. Again, this unit has stayed dry when this problem is occurring, but I don't really know all the things it's connected to. I wonder if something is convincing the alarm unit to cut power to the DME relay.

[Bulldog9]

I would also check the fuel pump wiring and connections, especially in the rear and under the car.

You'll find it!

Does this happen at all when dry?

[neilschelly]

I did confirm all the wiring for the fuel pump is clean/good, and that ground has also been cleaned.

This _never_ happens when dry. It _might_ happen in the dry if the water is still present wherever it is getting to after a rain storm, but it will "fix" itself quickly enough that I can drive normally within a few minutes.

[walfreyydo]

One thing you can check is in a dark garage, engine running, spray a fine mist of water over the spark plug wires, ignition coil, etc and look for electrical arching. If so, youll need to replace whatever component is causing it. Main suspects would be plug wires (including the one running from the coil), ignition coil or cap and rotor. I know you think its related to fuel and the DME relay, but this is a very easy and quick thing you can check.

I had this issue quite commonly on my old GTI VR6, where the ignition coil plastic housing actually had developed small cracks and would arch in wet weather, causing a misfire. You could see it in the dark and a fine mist of water.

Another thing you could try is swapping DME computers, just to eliminate that as an issue - this would make more sense in relation to the blinking fuel pump lights on the DME relay (remember that per the clarks garage DME relay diagram, the relay gets a ground signal from the DME to complete the relay switch and send 12V from battery to the pump!). An issue with the DME computer could hypothetically interrupt that ground signal and cut power to your pump...

You can also try putting a multimeter on that ground signal (pin 85b) and see if its switching on/off when you are having the issue. That will help you isolate whether the ground signal from the DME computer is the issue or not.

I have a known, good working DME computer from my 89 S2 available.

[neilschelly]

I pulled the O2 sensor, and I decided it was worth replacing after seeing it too. There was definitely some peeling electrical tape from at least two spots on the wiring. Now, I'm just waiting for some rain to really test it.

I _did_ go for a drive-thru car wash, which has definitely triggered this before. There were no issues. There's no rain in the forecast for at least the next week, and I've already turned off the hoses outside for winter, so I was impatient. I'm hesitant to call victory yet, but I'm cautiously optimistic.
-N

[neilschelly]

Finally, some rain...

It's literally been this long since I had an opportunity to drive in the rain. Today, I was out in heavy-ish, consistent rain, on back roads and highways, with stop and go driving, restarting after parking for a bit, etc. There wasn't a single hesitation or misfire, and this weather definitely would have caused a problem every time before.

The O2 sensor was it! Woohoo! Thanks all for your help/ideas!
​​​​​​​-N

[azbanks]

Congrats on finding/fixing your issue. I hate intermittent problems like that.

[Monkey Wrench]

Nice to hear the story had a happy ending! Intermittent issues can definitely cause sleep loss ;-) Congratulations !