Carpokes Porsche Forum

Hello All,

Last weekend I encountered something new (to me). I started my '87 944 na like normal. Left it idling while I went to bring in the trash bin from the curb. While walking back to the car, it stumbled, coughed, and died.

Since then, I haven't been able to restart it.

What I've checked...

• tach bouce - yup, like always
• fuel - about half a tank, but I added more anyway
• spark - removed #1 plug and grounded it against block while helpful girlfriend turned the key
• jumpered DME relay - yup, can hear the fuel pump running, can feel the supply line pulsing
• cam is turning - confirmed by removing rubber grommet and watching it turn while gf cranks
• speed/reference sensors - both test out fine at 900Ω (good YT Video)
• speed/reference sensor wiring/connectors - all fine, nothing cracked, pinched, pulled, etc
• no water intrusion from battery tray onto engine computer in passenger footwell
• no vacuum lines disturbed/broken/disconnected/etc
• no critter nest in air box

Car has 250k miles, engine rebuilt with lots of new stuff 40k miles ago including...

• speed/reference sensors - haven't been unplugged, moved, or touched since
• fuel pressure regulator and damper
• vacuum lines and rubber connectors
• distributor cap/rotor + plug wires

And in last ~1000 miles...

• new in tank fuel strainer (old one was fine - no crud in tank even)
• new timing belt (but cam is turning so it ain't that)
• new spark plugs

I have a fuel pressure test kit on the way - arrives tomorrow - but I don't think it's that. I'll check anyway.

The "spark plug grounded to block" test didn't give me a great looking spark. So that's something.

Distributor cap has one badly scored contact, one mildly-scored one, and the rest are fine. I would think I'd still get some attempt to fire the remaining two cylinders even if the badly/mildly scored contacts didn't work at all. Plus, car was literally running like normal when it coughed/sputtered/died.

Regardless, I'll be ordering a new cap and rotor.

When testing the speed/reference sensors, I could only occasionally get my various multi-meters to give me a reading across the outer pins. Those are supposed to be greater than 1MΩ. I'd briefly see a value of around 1.6MΩ (when set to the 2M scale on the non-auto-ranging DMMs) which would then go to "out of range". (I never got a reading on the auto-ranging ones.) Don't know what to make of that. I have a hard time thinking both would fail at exactly the same time. (These are genuine Porsche ones from about 10 years/40k miles ago)

I had NO problems with hard starts, rough idling, stalling, hesitation, etc until this. Car has been a champ.

So, what am I missing here?

The fuel injector connectors are the same failure-prone JT connectors as the speed and ref sensors. Since you have tach bounce and some spark, but no sign or combustion, I'd check for fuel in the cylinders. Pull some plugs and crank it -- you should see/smell fuel in the cylinders. Or just rattle the fuel injector harness as you crank it. All of the injectors are fired from the same driver inside the DME (despite what the harness might make you think), so if ONE injector is shorted out, then NONE of them will fire. Could be lots of things, of course, but that's the next thing I'd check. Usually you'd get some kind of pops and sputters if there is any spark and any fuel.

If everything checks out and you are still not convinced about the speed and ref sensors, I could pull resurrect one of my old Carmonikas if you're not an o-scope kind of guy.

Tach bounce is a big clue. This proves that the DME is receiving the speed/ref sensor inputs and is generating an ignition pulse, which is also fed to the tachometer. If either the speed or ref sensor inputs are missing, there will NOT be tach needle bounce, at least in the case of the later ('85.5+) 944.

"Coughed/sputtered/died" points to running out of fuel---loss of spark/speed sensor input results in instant stalling. Until you get your fuel pressure test kit, you can crack open the fuel rail test cap (be sure to use a 19mm wrench to counter-hold the fuel rail while loosening the cap) after cranking (wear safety glasses!) to get a sense if there is fuel pressure/flow.

If there isn't much fuel flow, could be a pump issue, but you indicated that the fuel pump runs when jumping terminals 30 to 87b at the relay board. If you have an ohm-meter, you can check if the DME is providing the ground signal to terminal 85b of the relay---a rare but not unheard of failure of the DME, though there could be another reason that signal is not present.

Let us know what you find.

Quick update,

The injector plugs/harness/connectors all look okay. Old and getting more brittle every year, but okay. Cylinders have the smell of fuel with spark plugs removed. Spark plugs are "wet" with fuel from non-ignited fuel.

I confirmed that the distributor rotor was spinning, in case the set screw had come loose for whatever reason.

However, after putting the cap back on and rigging up my new spark tester, it showed the dimmest faintest weakest indication you could hope to see. I set up my phone to shoot a video. It was incredibly hard to make out any flashes from the tester. Since I haven't used one before, I tested the tester on my buddy's truck. There it showed super obvious flashes.


So I pulled up an Article on Pelican about testing the ignition coil.

• I have proper voltage at the primary terminals - 12V on each side with the key on (the ignition system grounds the green wire to trigger the coil)
• However, across those terminals (disconnected) I have 1.5Ω, which exceeds the spec from the Pelican article (which is 0.4 - 0.6Ω)
• Still have 5.2KΩ from "green" terminal to coil output. So that's in-spec still

Regardless, the spark tester shows almost no indication of a spark. Plus my cap is worn. The coil input on the cap where it has a spring loaded "brush" to remain in contact with the rotor is worn down to almost nothing. A couple of the contacts are scored from making contact with the rotor.

Gonna order a cap/rotor/coil. And hood struts because dammit if those didn't decide to start crapping out right now, too.

Still want to check fuel pressure at the rail - tester won't get here until later. Thank you for reminding me to counter-hold the fuel rail while removing the end-plug. That's a thing I could have easily messed up.

With most multimeters and dirty car parts, it is hard to measure down to fractions of ohms. Crud on the terminals, etc. often adds an ohm or four. If you measured 1.5ohms, that could still be good (I'd guess probably is). You can also measure from the positive post to the center post. I don't have a spec handy, but it should not be an open circuit or shorted -- something in the 5k to 25k ohm range. [Edit: sorry, looks like you already checked that.] The ignition driver/transistor inside the DME is one of the few parts that are known to go bad -- the transistor can overheat and get fried. It can also melt its own solder joints and lead to cracks. (In later variants of the 944, they took this circuit out of the DME, presumably to better manage the heat flowing through it...). It is a big T03 transistor mounted vertically on its own little board on one side of the DME board -- looks like the one below. The transistor is no longer available, but there are modern replacements that can be used. Worth opening the DME and looking for any obvious signs of problems with or around that transistor.

Tom, those are good points. I didn't use a great multi meter when I tested it in the drizzle this morning. The difference between 0.5 and 1.5 ohms may indeed be within the margin of error for the DMM I used. That and the terminal corrosion, etc.

I've been hesitant to crack open my DME, in case I should need to send it in to someone who rebuilds them (and they balk at it being opened previously). Though I think I remember seeing someone now selling a drop in replacement built with modern components. F9 Tech perhaps?

I'm gonna pull the DME from the car and take a look this afternoon.

usury wrote: Mon Oct 21, 2024 1:19 pm Tom, those are good points. I didn't use a great multi meter when I tested it in the drizzle this morning. The difference between 0.5 and 1.5 ohms may indeed be within the margin of error for the DMM I used. That and the terminal corrosion, etc.

I've been hesitant to crack open my DME, in case I should need to send it in to someone who rebuilds them (and they balk at it being opened previously). Though I think I remember seeing someone now selling a drop in replacement built with modern components. F9 Tech perhaps?

I'm gonna pull the DME from the car and take a look this afternoon.

ECU Doctors in Florida won't care if you cracked open the DME. You should see the stuff I've sent them. Very much worth opening yours though. If the ignition transistor burned out, it may (or may not) be obvious. I've also fixed more then one DME (ok, technically 1 DME and 1 KLR) with nothing more than a soldering iron and magnifying glass -- just reflowing cracked solder joints. Lots of nerdy ways to test the DME, but the quickest is just to try a spare DME in your car and/or put your DME in a different car.

And yes @FTECH9 makes a brand new replica of the DME with virtually all new modern components.

Well that was easy! Damn. Car started on the first crank. Literally, one crank and it started.

I opened the DME, which just takes a little patience, a small flat screwdriver to bend the tabs up, and a small pliers to straighten the tabs so they clear the base plate. The screws are straight forward.

Upon opening, I noticed no smell of burned electronics, so that was good. Except for a single spider web, it was a pristine 80's time capsule in there.


The large driver transistor is super easy to identify. So are the three solder joints where it meets the main board. I needed my lighted magnifying glass on a weighted stand to really see anything wrong at all. (Get one if you do any of this kind of stuff yourself - you won't regret it.)

It was tough to photograph, and even tougher to see. I could barely feel the crack with the tip of my pinky fingernail. Here I've done some high-dynamic-range color processing on the close up for it to be noticeable.

There were a couple other power transistors on the other heat sink that had crusty/questionable solder joints, so I removed the solder from those pads (with a cheap-o spring loaded solder sucker - also get one, you won't regret it) and applied fresh.

I think it was the "big" crack pictured above though. Even with the worn distributor, car runs fine. Still gonna replace the cap/rotor. And still gonna test the fuel pressure at the rail, since the test kit arrived while I was installing the DME back onto the floor of the car.

Seems fixed. Cost zero dollars. Well, zero plus a fuel pressure test kit I haven't used yet (wanted one anyway), plus a spark tester, plus realizing its time to do cap/rotor.

At least the ignition coil works.

Thanks everyone!

Awesome! So glad you found it! No better feeling than finding a needle in a haystack like that and fixing it for free!

What wires are you running? Any weird aftermarket ignition parts? Although cracked solder joints are getting more and more common simply due to age, that particular joint often cracks due to excess heat. With all stock parts, it shouldn't overheat, but they can overheat if the down stream ignition parts are drawing too much current. A very long time ago I fell for the hype around Nology wires, which have a built in capacitor. I discovered by accident that they just smoked that transistor. I did a little test back to back, and those wires in particular made the transistor super hot -- way too hot to touch. With the stock wires and no other changes, the transistor was much much cooler. Just food for thought...

Tom wrote: Mon Oct 21, 2024 6:15 pm Awesome! So glad you found it! No better feeling than finding a needle in a haystack like that and fixing it for free!

It totally does! I feel pretty good tonight.

Tom wrote: Mon Oct 21, 2024 6:15 pm What wires are you running? Any weird aftermarket ignition parts?

As far as I know, no special aftermarket ignition bits at all. I've never tried to get "moar power" out of the car in the 25 years I've had it. Not even a K&N air filter. When I had a quality Porsche shop in Wisconsin rebuild the engine a decade ago, I asked them to keep it as stock as practical given parts availability/cost. I certainly wouldn't have gone for high performance plug wires.