Got a weird problem, hoping someone out there will come up with something I've missed here. I have a '94 F-car LT1 with an MSD Optispark replacement (which fortunately has not fallen apart on me in about 10,000 miles so far, unlike some of the ones I've heard about on here, none of that high RPM miss stuff).

What happens is: when the engine's warming up from a long cold soak near or below freezing and the ambient temperature is around there as well, the engine speed is between 1500 and 1800 rpm, and the throttle is above about 20%, I get a misfire, which is happening on all cylinders; the engine cuts out completely for periods of under a second. The computer sets the dreaded Code 42 (EST Grounded) and no others. When I first looked at the codes, there was a Code 36 (high resolution pulse fault), but since I cleared it the problem has occurred several times and it has not come back.

This same problem happened last winter, and I never found a solution other than avoiding that RPM range. None of the wires involved appear to have anything wrong with them (switched power to the PCM and coil, white wire from the PCM to the ICM, grounds at PCM and ICM), and I seem to remember trying my spare PCM with no difference. I had just bought a new ICM before last winter, and it seemed unlikely that one would fail to work only when cold and only at a certain engine speed, so I didn't try replacing it. It went away after changing the plugs (one of which was fouled with ash from what I suspect is a slow coolant leak into one of the combustion chambers), but that also coincided with the last real cold we had here, so I can't say for sure that the problem was fixed. I was able to capture a data log of the problem happening a couple days ago using FreeScan, which I can send to anyone who's interested. Mostly you just see the effects of the car repeatedly cutting out, so that my foot jerks up and down on the gas pedal; however, I noticed that at one point during the misfire event, the indicated RPMs jump from 1225 to 2125 in 1/8 second (which seems to be the time resolution for engine speed) and then drop to 1575 in another 1/8 second--and this happened with the clutch engaged, unless my foot hit the clutch at some point. That's way too much to be explained by driveline slop, which would suggest something strange is happening with the timing reference.

Any ideas?