Problem with Intermittent Misfire, Worsens at WOT

#3

Bad afm can do this too. Do a search for afm with my username and you’ll find the thread from a couple yrs ago. “bad” includes an afm that has had it’s spring tension screwed with.

#4

If you do a smoke test you may get some good info.

#5

Google diy smoke generator.

#6

Thanks guys, I’ll do the smoke test tomorrow and let you know.

Ranger, I doubt it’s the AFM because it’s the same one I had since purchase and it ran great with both the old Stickley engine and the used junkyard engine. I had an AFM problem with the ETA car, which had similar (but not identical) symptoms (would always clear up at WOT). But I’ll check your thread.

Also, for future reference, this is a great resource for checking AFM’s:
http://www.the944.com/afm.htm

Note that the “early 944” is the same as the ETA AFM, and the “late 944” style AFM is the b25 and super eta AFM.

#7

What you’ll see in the thread I mentioned is that I spent the better part of a year struggling with symptoms like yours. Early in that fight I installed a fuel pressure gauge and a F/A meter on my dash, and connected both to the data logger. Sadly, it took me many months before I believed what the F/A meter was trying to tell me.

Gauges and data like that are worth their weight in gold when engine management is kicking one’s butt.

#8

Crankshaft damper bolts could be loose. Faulty wiring harness to the injectors or a bad connection at the injector harness. Those are two easy things to check that have given me high rpm misfire in the past.

RP

#9

When the engine would loose power did the tach take a big drop relative to road speed? If it did I’d say that the CPS and/or the engine harness is bad. If the tach didn’t react I’d be looking at C191 and considering the possibility that damage within the engine harness is interfering with the signals to the injectors. Oh yeah, a dodgy WOT switch is the TPS can also do this.

#10

Replace the coil…its breaking down at higher temps.

#11

Chuck, I did replace the coil with a used one I had on hand which was known good (pulled while troubleshooting the ETA car). I have another new coil I could try but doubt it’ll do the trick. The used coil never got hot because I had the hood up the entire time. Depends on how desperate I get.

Jim, I saw posts from you earlier on whether the tach dropped. I thought that the tach dropped while on track and having issues, but subsequent reproduction of the problem has yielded no tach signal dropouts. This would seem to indicate not a CPS or CPS wiring problem, right? I pulled the injector connector board and probed each conductor from the injector harness connector to the injector and sensor connectors, and they seemed ok, but that doesn’t rule out an intermittent connection issue within the assembly. The injector harness connector terminals looked fine. Is that C191? I didn’t get a chance to check my schematic yet.

Patton, I’m going to check on the crank damper bolts now.

Ranger, I’m working on fabricating a smoke machine to see what we find. No matter what it looks like a great tool that I was previously unaware of. Looks like I need to make one, then! :wink: Unfortunately, progress was slowed due to loss of my daily driver in a flood last Wednesday and dealing with corresponding issues. Hopefully I’ll have it built and in use by this weekend. Nichrome wire is on order today.

Thanks again guys and I’ll let you know what I find!

#12

If the tachometer doesn’t react you can rule out the CPS or it’s data as a suspect. C191 is the round connector under the intake between the engine harness and the fuel injector harness.

A diesel glow plug is an excellent heat source for a smoke machine and is pretty close to what the commercial machines use.

#13

So a quick round of examination yielded tight damper bolts and tight ground wire connections (figured why not, right?). Per BEdmonds, I listened for leaks. It seems to me that there was a slight hissing at idle, near the #2 cylinder intake port. Though I don’t have the smoke machine assembled yet, I tried spraying brake cleaner here and there, and thought that it might have sped the engine up after about 3 seconds when I hit the vicinity of the #2 intake. It’s far from conclusive, but I feel like it’s a reasonable indication. If so, the smoke machine should give us conclusive proof.

For those of you who’ve built your own smoke generators, what did you use for an enclosure, and what did you use for your smoke source? I see baby oil/mineral oil is common, but wasn’t sure if some had a preference for something else.

Thanks again!
Matt

#14

I have built one and used an old tank I had from some sort of sprayer or low pressure vessel (about 4" in diameter and 6" tall). The heater was a glow plug. I brazed fittings in for the glow plug and smoke output and used the existing fitting on the top of the tank for air injection. I had to hunt around to find a small regulator that could reliably work in the 2-4psi range needed. Smoke oil works best but baby oil will work.

You will also need a way to plug the exhaust and make up something to replace the AFM that has a port for smoke injection.

#15

Case of beer says it’s what i said: intake manifold gaskets. Any takers?

#16

Great learning thread for all of us.
When the problem is resolved, I’ll go old school, hit the print button, and put this in my “engine” file folder.

RP

#17

Mine is very similar to Jim’s. I bought an empty 1 qt paint can from Home Depot/Lowe’s, and drilled three holes in the top lid for the glow plug, air in, and air out. I used mineral oil.

#18

Should the smoke test not show problems, replace the fuel pump. I spent 12 dyno hours trying to tune out a high speed miss only to find the fuel pump was not holding pressure/volume properly.

#19

Ok, so I built my smoke machine and sealed my intake and exhaust. I found no observable pressure at the tailpipe, despite using a fairly substantial pressure (based on the sound and the expansion of my intake seal – a rubber glove) to supply the smoke. I applied smoke to the intake via the inlet to the air box and via the fuel pressure regulator manifold vacuum signal. Strangely, when applying smoke to the inlet to the air box, I could not observe smoke from the manifold vacuum port. Perhaps this is because the smoke appears to be heavier than air. However, in both cases, I observed no leaks from the intake manifold side (excepting the air box moisture drain hole, which I plugged before continuing my inspection). I discovered that the “bitch tube” (per Ranger’s description) leaked pretty severely at the block side (left side of engine, above the left side engine mount bracket, the vent tube between the intake manifold and the block). Other than that, I could not locate any leaks of any kind. When I pulled the dipstick, I also observed smoke escaping, so there was plenty of smoke in the crankcase. Not sure how the smoke got into the crankcase in such quantity, as I wouldn’t expect it to go past the rings so easily. Thoughts?

#20

The crankcase is connected to the intake manifold via the breather tube between the intake and valve cover. To do this test properly you must have a low range pressure regulator and a pressure gauge. A regulator & gauge rated for 0-15 or 0-20 will work and are readily available.

Commercial smoke test machines include a sensitive flow meter. Which is handy as a clue that there is a leak even if you haven’t yet seen smoke.

#21

Jim,

I know the intake is connected via the breather tube, but my understanding was that it did not open into the intake tracts, only into the head, to permit pressure equalization. Which is why i’m wondering how so much smoke is getting in there. Unless it’s leaking at the center intake manifold gasket, of course. Are there any other paths from the intake to the crankcase?

Even if I had a sensitive flow meter or a 0-15/20 gauge, how would this help for a rather coarse system? I haven’t wrapped all the joints in my exhaust with tape, but I’m sure they’re permitting loss of air pressure because they’re slip fits with no clamps. Clearly, I have a gross leak at the vent tube and I’m wondering 1) whether that’s indicative of a problem and 2) if not, how to close it so that a sensitive flow meter and a pressure gauge would help. Unfortunately, removal requires pulling the intake and it’s also very difficult to reach. Even with a new set of seals, apparently it doesn’t seal well at all.

Matt

#22

The top of the head is open to the crank case via the oil return passages and oil return tube. The valve cover breather hose goes directly to the intake manifold (no pcv valve). So if you pump smoke into the intake it has no choice but to also flow into the block via the cylinder head.

The point of the pressure regulator & gauge is to make sure that the smoke pressure within the intake/engine is high enough to find the small leaks. But you don’t want too much smoke pressure as that can create a leak. The gauge lets you know when the desired pressure has been reached and then you make sure that pressure is held for the duration of the test. The flow meter will show evidence of flow going into the engine/intake before the smoke may have worked it’s way to a leak. The flow meter is nice to have, but not a necessity.

On a stock car the exhaust is essentially gas tight unless it has been damaged. On a Spec E30 that won’t be the case and the slip joints will have to be taped up.

As you may have figured out by not, while not rocket science, a properly run smoke test is not just “blowing smoke”.

The oil return tube, aka “bitch tube”, is no problem to seal. The intake does have to come off, but new oil return tube o-rings and intake manifold gaskets will get the job done. Unless you get real lucky, the bore in the block the tube fits into will need to be cleaned up with a brake cylinder hone and the ends of the tube run against a scratch wheel to remove corrosion. The intake manifold side is usually okay. For the spring and o-rings to work, the tube must slide easily in the block and intake.