My 91 accord has been giving me grief 4 months. Sure would appreciate any help - TX GREASEMONKEY, u have been on the mark before, want to take a shot... Its my only transportation and has been down for a month. Months ago it would not start after driving until hot. It would start after a couple of hours with no problems. I tested it when it wouldn't start: fire to plugs, gas, air. A couple of months later, it started dying when driving. Sometimes it would start, idle rough, then die. On second go around it refused 2 start under any conditions. Same situation: fire to plugs, fuel, and air. Since then I have tested the main relay, fuses, ignition switch, wiring harnesses, coil, and icm. The icm tested as bad according 2 chilton (not sure now because their wiring description wasn't consistent which led to a guessing game). I found an aftermarket distributor cheaper than icm, so I replaced the distributor. It started up no problem and ran like a top 4 about 5 minutes, sputtered and died. Same situation as b4: fuel, fire, air, yet no start. After reading that aftermarket distributors may behave this way (TX GREASEMONKEY), I swapped out the icm and new coil and put the old distributor back in. Still nothing. I used chiltons ambiguous tests again and both the coil and icm checked out. This has long since gotten old. Any ideas?

If the CEL consistently comes on for 2 seconds and then goes out, when the ignition switch is turned to ON (Position II), then focus on the distributor. I would install a new name brand aftermarket ICM, coil, rotor, and distributor cap on a new genuine Honda distributor housing. This should solve the problem and keep the cost down. You require the reliability and safety of knowing your engine is not going to die on you in traffic--your life is on the line. Aftermarket distributors for Hondas are not reliable--they have a hard time getting the internal sensors to work properly. Distributors are the weak link on older Hondas. Ensure the aftermarket parts are compatible with the new Honda distributor housing.

P.S. I would not consider Chilton's tests for ICMs and coils to be definitive, by any stretch of the imagination. Nor would I rely on the Well's machine at AutoZone for testing ICMs, because I've seen many that tested okay that were actually bad. Instead, I recommend replacing ICMs and coils every 125,000 miles or 10 years, whichever comes first.

Thanks for the quick response. Really appreciate your help.
Haven't been able 2 work on since reading your post. Worse yet I'm broke until the 4th. Even then, I won't b able 2 swing the 5 billsthey want for an approved new distributor with all the sensors.
Again, I really appreciate your help, and am not taking your advice lightly - but I don't have the cashola. Im not on a low buget, imm on a NO budget.
Is there anything I can do in the mean time?
B4 I read your post I did the following:
- retested all the fuses and relays
- tested the input voltage to the ecu with ignition at II (ffull voltage)
- tested the continuity from icm to tach, ecu, ignition (all good)
- tested the ground to the ecu (good)
- found that the grnd at the thermo housing was loose > cleaned, sanded, replaced, tightened well
- tried 2 start > cel came on and stayed on for the 1st time since it has been broke down, but no start.
- rechecked both grnds, all basic wiring.
- attempted to start / check codes > back to no cel at all, no start.
- attempted to test ecu by testing the voltage to the fuel injectors > full voltage when ignition at II, but only 2.73 volts when cranking. Read somewhere that if voltage < 5 volts = bad ecu. What do u think?
Unfortunately, there is an added variable - my battery won't hold a charge so I had it on a 6 amp charger for all the above. After cranking the charge would drop to 10.26 volts, take ~ 2 minutes to recharge to 11.76 (supposed to be 12.5?).
Any more counsel till I can buy a battery? All I can afford till next month.
Thanks 4 all your help. God bless.

If the CEL comes on and stays on, the ECM is likely bad--especially if the K-Test test voltage is way low.