brettp1

I finally think I have found the issue to my stumble issue at around ~2500 RPM. I used Dataminer to capture several ECM dumps and found that the pulse injector period and duty cycle is set to 0 when I have problems. All sensor data at that time looks good. I am suspecting the ECM is the problem. Any suggestions on a good ECM source? Any other things that I should consider?

SunCr

I'd check MAP and TPS. Looks like the ECM thinks you've taken your foot off the pedal. MAP is somewhat problematic in that an intermittent can be so quick that a lot of scanning software won't pick it up. Graphing helps. Check the TPS with an old analog meter. Digitals just give you an average. With the analog, a fluctuating needle as you open the throttle is a good indication that it's the cause. If it's nice and steady, it's probably ok.

I buy reman ECM's from the Dealer (for no particular reason that I can think of).

vader86

Better double check the other sensors before spending money on an ECM, its very rarely the culprit.

brettp1

I guess I would be suprised to get a scenerio were no fuel is requested. Is 0 pulse width and duty cycle a valid response?

0ski_dwn_it

shoot me the logs, I would be happy to take a look at them.



Describe what your seeing when you see the pulsewidths go to zero. Its not decel fuel cuttoff is it, when you leave off the gas peddle and the car is "coasting" is it?

SunCr

Per above - decel (above 1500 rpms) cuts off the fuel. Monetarily lose the TPS or MAP signal and guess what happens? Those are the signals the ECM uses to enter decel. You should be able to duplicate at other rpms by closing the throttle. Graphing (at least for me) works really well for this type of problem. Set it up so it draws TPS, MAP and Pulse Width and/or RPMS.

brettp1

Thanks for the help. I was graphing the response when running the car from 1000 RPM to 3000 RPM in first or second gear. Typically as I see the RPM rise, the TPS voltage looks good, and the MAP output starts to fall with increased RPM. During this acceleration profile, the injector PW is set to 0 for one sample. This does not always happen. Half of the runs with the same profile show a minor change in the injector PW at the problem RPM. It seems to be sensitive to temperature, which leads me to the guess of ECM. However, the sampling time could be masking some weird MAP or TPS input. I guess I never expected to see 0 PW on the injectors while accelerating. I did put a fuel pressure gauge on the fuel rail during the runs to make sure I didn't see a drop in fuel pressure.

TA

If it ends up being the ECM, I got one for my 91 at Napa for like $90.

SunCr

"and the MAP output starts to fall with increased RPM"

Voltage should be linear and rising. I'd go back and use an analog on the TPS signal and you can try it on the middle/signal wire of the MAP (though that's a little more difficult because you need to have it running - though I've often thought about trying it with a vacuum pump and the key on). You're looking for a flat spot/drop, needle wiggle from either. Intermittents - unless heat related (aim a hair dryer at the ECM to find a cracked solder joint) are usually the sensor or wiring - I created one from hell by crimping (not soldering) in a new MAP harness - finally picked it up with a Tech 2 and some Mitchell software (crap, I should have just wriggled the wires!).

brettp1

Thanks again. I will eventually track down the issue and put the results back on the forum.

0ski_dwn_it

This is an odd one. I took a look at a log that brett sent me.

The TPS is good, as far as reporting to the ECM. Not that.

The event occurs at ~2200 RPM give or take a few RPM nearly every time. But not necessarily every time the RPMs pass over 2200 rpms. And even at the same lv8 values, which can eliminate several controlled parameters like timing tables, and some fuel events.

also some of the evens are followed by knock, which does mean that the fuel shortage at the moment could be causing real knock. Not bad knock, but its causing some.

My thoughts are its possible something in the ECM is tripping up or in the memcal. Its odd that it does only happen at 2200rpms, which might be related to a vibration or something causing the problem to spike up.

These are all theories. I would like to look at some more logs Brett with some more driving in different conditions to perhaps give me a clue if other events are happening.

I just spent about 40 minutes running the events against other parameters trying to see it there are some correlations and nothing is jumping out. About the only correlation I could see, which is not real convincing is a voltage change from 14.1 to 14.0 sometimes when the event occured. This possibly could be a short happening, but I have to say there are other times when it also fluctuates and it doesn't occure. ??

Get me some more logs at say WOT, and some other driving conditions and maybe I can put 2 and 2 together...

Some other suggestions are if you have a buddy that you can borrow an ECM from to give a try to see if the problem goes away with the swap out, you can nail it down to the ECM itself.

Lastly, this is just a quick thought, but check your battery connections. I have seen MORE issues related to a loose battery in EFI cars and non-EFI cars than you care to know about. LOTS of VERY small annoying problems can be tracked back to this problem. I once chased a problem for over a month of lost ET, due to a battery problem. Could also be the battery itself shorting out. Since the battery voltage is reported though the alternator, you may not see the change in voltage much.

Just read a good racing/battery article that stated from a very reputable company/race group that on a hard launch of a racecar your typical battery can see several shorting outs through just the 60' mark! Perhaps that is why I am going through batteries like water. They said not to expect a battery due to this abuse to last over 1 season! Thats about what I get, if that.

Just some food for thought. Get me some more logs, you got my interest up now. lol

brettp1

The voltage change is interesting. I will make a few more runs and forward the data. Thanks again for your help.

The car sat for a while I replaced the clutch and I had to replace several batteries that would not hold a charge for more then a week. I had to jump start the car several times and I am sure this was not good for the ECM. I will check to make sure the battery connections are good and see if I can procure an ECM.

What symptoms did you see with the loose battery connections? I'm surprised that the ECM did not reset in this condition.

0ski_dwn_it

A loss of about .2sec through the 1/4 mile. No missing or anything, just down on power. Drove me NUTS! Thought something was seriously wrong with the motor and spent about 20hrs trying to find the problems and SEVERAL trips to the track, traveling 2.5+ hrs each way - which just ended with frustration. Finally at the track the one time, the car would not start- which about sent me over the edge. I ran to get the jumper pack, and upon putting the jumper pack on, I saw the nut move on the terminal! I ran and got the wrench, tightened up the fitting and the next run was right back on target! It was the positive terminal, which I seldom disconnect when working on the car, so all that time when working on it, I always undid the negative.

lesson learned, start with the easy stuff. People have a tendency of thinking the worse things first, and going for big ticket items. All my problems like above have all turned out to be stupid little things.

Get me the logs and i will be happy to look them over. check your battery ground, as that could also be a source of the problem.