I bought my '98 GTP in 2011, then went back overseas until late 2012, leaving the car at home. Several people drove it during that time, and due to a coolant elbow leak, it was severely overheated at least twice. Also during that time, my dad visited my house and replaced the snout of the supercharger with a remanufactured part.
When I came home in late 2012, the car had a check engine light (MIL), and at cruising speeds on the freeway it had a flashing MIL. (No feel of a misfire - the car still ran smoothly and accelerated like it was supposed to, but that light wouldn't go away.)
I hooked up my scan tool to it and found a P0300 code. Because I didn't know much of the recent history, I started by cleaning the injectors, replacing the fuel filter, replacing the plugs and wires, and replacing the long, spidery plastic vacuum hose that goes to the fuel pressure regulator and EGR (broken during supercharger repair). No luck... still had the exact same problem.
I purchased AutoEnginuity so I could graph the data from the PCM and found that only cylinders 3 and 6 were misfiring. Since they share the same coil, I swapped coils, hoping for an easy fix. No luck - the problem stayed with 3 and 6, even with the 1-4 coil firing those cylinders.
At this point, I was starting to think that perhaps the severe overheating had damaged a head gasket or cracked a head, but I had no mixing of coolant and oil, and I wasn't losing either fluid. I gave up on it for a while and just drove it with the flashing MIL, because it didn't run at all like it was actually misfiring.
Recently I decided to tackle it again before I went ahead and tore down the engine, so I grabbed an Ignition Control Module (ICM) and coil packs from the junkyard for $35. Installed the new ICM and coils - no difference... still misfiring on 3 and 6.
More reading online led me to an article on Edmonds with some suggested fixes. (http://answers.edmunds.com/question-...lp-167605.aspx) Based on that article, I borrowed a Tech 2 and performed a Crankshaft Position System Variation learn. This procedure is only supposed to be needed if the crank sensor is removed, replaced, or disturbed, or if the PCM is replaced. As far as I know, none of those things happened, but the CKP System Variation Learn solved my problem.
I brought up the misfire graph on the Tech 2 and drove for over 30 miles without a misfire. I then went home and reinstalled the original ICM to verify that it wasn't part of the problem. Still no misfires. It's now been almost a week and no more MIL. As far as I can tell, the problem is solved, so I'm posting my experience here to hopefully help someone else out.
