Miata Iginition issue - help needed

[MMiskoe]

Looking for some help on diagnosis of an ignition issue that cropped up at the VIR 24 hour.

1994 Miata, stock electronics and engine management system.  Dash board has been removed and gauges rewired.

Car started the race and ran at low RPM for 60+ minutes due to rain and an extended safety car period.  Once we went back green it did 2 laps before it overheated and destroyed one of the coils.  Change coils, send it out, gets about 20 minutes of run time, same issue.  Changed again, same scenario.  At that point we were out of ideas and coils so we packed up.

When they fail, the coils are too hot to touch and some of them physically cracked near the plug when they come off the car.  There is no indication of problem before the car simply goes to two cylinders.

 

We assumed that something was over-driving the coils and they were acting as if they were passing currently 100% of the time which they aren’t really capable of and thus they over heat.  We got our hands on an oscilloscope and measured the signal from the ECU to the coil.  The attached picture shows the racecar and a streetcar (1999).  We also tested a different racecar which looked nearly identical to the streetcar, so we know that the coil signal in the racecar is wrong.

 

So we started swapping parts.  ECU.  Cam angle sensor.  Alternator.  Un-plug the fuel pump while it is running.  Check the signal at the ECU, not at the coil pack.  Unplug all data systems, tachometer, lighting, cool suit.  None of these things had any impact on the signal we were reading on the scope.  All the grounds have been looked at and appear to be sound.  We were fully expecting that the ECU, CAS or alt change would have made an impact.

 

We’re at a loss as to what could be causing this.  The car ran fine last summer and last fall.  This spring it did a test day where we were on track for 10-15 minutes, then off track for twice that time all day.  We don't know it was a problem then because the run time was short, possibly giving the coils time to cool down. 

 

[turbogrill]

So the voltage to the yellow coils are almost always in the middle of on and off? What is the voltage range?

 

Could it be the wiring and the connector being glitchy? If you unplug the connector from the ECU and the connector to the coils and measure, what is there a short?

 

[TimS]

Looks like a short or back feed of sorts.  Try a new temp run of wire from ecu to coil and coil power wiring.  Double check with scope

[Team Infiniti]

I know you mentioned grounds when we spoke but how about addl ground(s) right to the coil/module and ecu to help isolate?

[Team Infiniti]

I had another buddy squint at the waveform his comment was ground related as well.

 

”Ground is ground how can those lines be going below ground”

[LuckyKid]

What's your voltage overall? How did you wire up your excite wire when you removed the dash?

[turbogrill]

Could you clarify the voltage swing on the yellow? Are the downwards spikes below ground or at ground? 

[MMiskoe]

Both yellow & green graphs should be 5 volts at the top of the wave, zero at the bottom.  With the scope we have it is hard to tell the actual scale.  I believe that is how the trigger works, I could be wrong and we're looking at 12V, but I doubt it.  Yes, the noise between coil signal would appear to be below zero, as if there is current backfeeding into the system.

 

I beleive it is a faulty signal from the ECU.  With everything hooked up, we have un-plugged the coil, and also simply cut the wire and watched the display as the motor spins down or runs on 2 cylinders, the wave form changes to being more square (because there is no load on it ) but the noise does not go away.

 

Grounds - I have looked at all grounds, they appear to be intact.  Prior to this issue I added a ground from the intake manifold to the chassis.  During the testing process I have made sure to keep a clip lead between the body of the ECU and ground. 

 

Voltage in the car looks OK.  12.5 with the ignition on, car not running.  14.5 when running, hunts around a little - 14.3 to 14.7.  Perhaps a bit higher than I would like, but not out of the range of what it should be (in my opinion).  On this note, I did a test with the belt off the alternator.  It was no different.

 

 

[TimS]

14v is good. Charges the battery. On the scope can you get the wave on a visible scale?  So you can see where 0vdc is on the wave. My cheap little handheld can’t do it but it will show the min max and average voltage. I thought you tried another ecu already?  You also have a few big spikes between the square waves. Looks like a short 

Edited August 16, 2021 by TimS

[turbogrill]

Do they have condensors? Make something is supposed to absorb that spike? Maybe that is missing or broken.

 

And are you sure there is different scale between blue and yellow? If there is same scale then there is something wrong with the earth on yellow.

 

 

 

 

 

 

[MMiskoe]

Well, turns out it is a ground issue, so anyone who said "ground" wins the chicken dinner.

 

There are 4 wires to the ECU that go to ground.  With it running and scope on, I back probed them and went to ground with a clip lead.  Two of the four had no impact, the other 2 cleaned up the signal so it looked as good as the blue/green image above. 

 

I decided that in the interest of time I was not going to seek out where these went, but simply soldered a piggy-back onto each of the four, and ran them to a ground lug near the ECU.  Double checked the scope, still looked good, turned out the lights in the shop, went in the house and put in an entry for the Thompson race.

 

The joys of a 27 year old wire harness.

 

So to all of you who responded and thought about it, I thank you.  I'd offer to repay the favor, but that means that someone is having a very bad day with their car so I'm not sure I wish that on you.

[TimS]

19 minutes ago, MMiskoe said:

Well, turns out it is a ground issue, so anyone who said "ground" wins the chicken dinner.

 

There are 4 wires to the ECU that go to ground.  With it running and scope on, I back probed them and went to ground with a clip lead.  Two of the four had no impact, the other 2 cleaned up the signal so it looked as good as the blue/green image above. 

 

I decided that in the interest of time I was not going to seek out where these went, but simply soldered a piggy-back onto each of the four, and ran them to a ground lug near the ECU.  Double checked the scope, still looked good, turned out the lights in the shop, went in the house and put in an entry for the Thompson race.

 

The joys of a 27 year old wire harness.

 

So to all of you who responded and thought about it, I thank you.  I'd offer to repay the favor, but that means that someone is having a very bad day with their car so I'm not sure I wish that on you.

Always good to hear about the solutions to crazy problems.  Have fun at Thompson and good luck.