96 miata stops running on 2&3 after warm

[garybled]

I have a 96 Miata that stops running on 2 & 3 after warm. After 2 trips to shop in Austin same problem. The first time it was 1&4. Then coil was replaced and then it stopped running on 2&3 after 15 minutes. Coil was changed again, and it also stopped on 2&3 after warm . That is when they told me PCM was out. New PCM and another new coil and it still happens. Catalytic converter is out. I unplugged second oxygen sensor and as soon as engine warmed up and engine light came on, 2 & 3 went dead again. Unplugged first Oxygen sensor to see if it would run in open loop mode and it still does not fire on 2 & 3. Allow to cool off, and it runs perfect again until warm (about 10-15 minutes). Any ideas? I am now going to replace the bad gutted cat and see if that helps. Dealer doesn't have a clue. They just keep saying coil is out and PCM is bad. This is 3rd coil and second PCM. Nothing affected the problem. I am losing my hair. My tech manual is just a Haynes and does not have codes for OBDII. I do have access to an OBDII reader at a local Honda shop. PCM runs just fine in another Miata, and coil runs just fine until warm. ???????? thanks for your ideas in advance ???????

[MadMerv]

I know this is stupid, but did you change the plug wires first??????

[Mark52]

Some kind of thermal short in the wiring between the ECU and the coils?

[channelmaniac]

The coils share a common ground... G117 @ Right rear of engine... a common power source... the Blue wire off of the ignition switch... and separate connections to the ECU.

I don't have a factory service manual but from looking at the wiring diagram it's something that could be narrowed down this way:

1. Replace plug wires.
2. Clean & retighten the ground connection.
3. Clean & reassemble the wiring connector for the coils.
4. Clean & reassemble the ignition switch.
5. Replace coil
6. Replace ECU

Keep in mind that if the connections are bad then you could theoretically blow coil packs.

EDIT: Oh yeah, check your wiring for shorts/opens too... You didn't do any underhood work recently did you? Don't laugh... other people have crimped wires under the valve cover

RJ

[channelmaniac]

I'm also wondering if you have a problem with the cam angle sensor (CAS) as that tells the computer when to fire the plugs...

RJ

[Kags1969]

I would tend to think in terms of a short in the wiring harness and the #2/3 coil as well. If the wire is shorting, then it could be taking a while for it to heat up and remelt the plastic around the wire, causing a miss. You can easily physically swap the #4/1 coil and 2/3 coil and see if the problem follows the harness. Or just swap the wires and plugs on the back of the coil.


Chris

[garybled]

Still no luck. I checked the plug wires, and also replaced them. I also checked the grounding. I thought about the cam angle sensor, but it doesn't make much sense that it would be affected by heat at just the point that the ECM goes from open loop mode to closed loop mode. Before I unplugged the oxygen sensors one by one the same thing would happen after it got warm but the check engine light wouldn't come on. After the rear oxygen sensor was unplugged, the car started and ran perfectly and at the moment the check engine light came on is when it began to cut out on 2&3. This would seem to indicate that it has somehing to do with the point that it switches from open loop to closed loop mode. Prior to that point, the injectors run on a preprogrammed setting. As soon as it goes to closed loop mode, the two oxygen sensors are monitored to look for a difference that corresponds to the catalytic converter efficiency, and the system then tries to optimize the fuel to produce the leanest burn consistent with a good fuel mixture. However, the oxygen sensors read correctly. I will replace the gutted catalytic converter first and see if that solves the problem. Incidently, it doesn't make any differene which ECM I use, the old one or the new one, and the same for the coil packs. In fact the 2nd coil pack that the dealer said was burned out, reads perfect on the resistance check and also works perfectly until it hits open loop mode. The dealer said the codes showed defective coil and computer, but something else is wrong, and they couldn't find it either (Mazda dealer in Austin). I trailered the car home and am now thinking about it. If anyone has any ideas about who is real good with this kind of problem here in the DFW area, please let me know. Thanks again for all of your help. Gary B

[Majik]

Just a note, on my 99 heat was apparently causing my cam position sensor to go haywire. It wasnt a specific temp, it was just when the car would get hot from running with the a/c on and little to no airflow into the rad.

[garybled]

OK, the latest thing. I replaced plug wires again, checked to cam angle sensor, cleaned the ignition switch (and gained 0.3 volts increase), checked to see that the voltage regulator was not over volting, checked the harness, puilled each pin every plug between the coils and the ECU and cleaned the contacts and rebent the middle clamping part of the connector for good pressure, replaced the ECU, another coil pack was installed, cleaned and sanded all grounding connections, measured voltage at the coil pack of 12.38 vs. 12.42 at the battery terminals, measured voltage with engine revving at 3000 at 14.48 v at battery and 14.41 at coil pack, and drove it. This time instead of going for 15 miles it ran perfectlyfor 3 days before it caused one of the coils in the pack to burn up. I am at my witts end!!! Any idea who is really good at electrical problems on these things in the DFW area?
Thanks!

[Titus]

Yikes... doesn't sound like there is much left to replace other than the entire wiring harness.

[garybled]

Hey, anybody have an electrical schematic from the factory that covers the 96 Miata? The Haynes schematic is crap. I have finally isolated the problem to what lies between the brown/yellow wire form the PCM igniter for coil 2-3 and the actual coil itself. I ran a bypasss wire from the PCM direct to the #2-3 coil and it didn't fire at all. when I hooked the original wire back tothe bypass wire at the pcm the coil fired perfectly. However, the problem obviously lies somewhere don this wire. It must be "teeing-off" and going to somehing else that returns info back to the PCM. Without a factory schematic, I can't find it. ANy ideas?
Thanks, garybled

[POS Racing]

Here is one : http://www.madracki.com/miata/wiring.html

There is another source on the web, but I got it bookmarked on the work computer, not home.....

[POS Racing]

There is another source on the web, but I got it bookmarked on the work computer, not home.....

Well I was wrong, I thought I had the other source bookmarked on the work computer but I don't.