1989 2.5L stops 2-3 seconds after starting... for the experts:

1989 2.5L stops 2-3 seconds after starting... for the experts: My local dealer has admitted great difficulty in diagnosing this recurrent problem. They have the car at this time. History: A few years ago, the 1989 Dynasty had a similar problem due to a vacuum sensor failure, which was replaced by the local dealer. Subsequently, all of the vacuum lines were replaced after one cracked. The new "T's" melted down, and were replaced. Last fall, the car would start and run smoothly for 2 or 3 seconds, and then shut down. This time, the dealer indicated that it wasn't the sensor, vacuum, or continuity problems, but appeared to be fuel starvation (the pressure was low). The filter and fuel pump were replaced which had no effect. Eventually, they replaced the computer, and the car was running smoothly again, except for a few stalling episodes while coasting to a stop which didn't recur. Other components were checked, and appeared to be OK. A month ago, the symptoms reappeared. The mechanic felt that it was not an electronic failure. The engine stopped running after a few seconds, and fuel pressure would drop. The injector was firing. Clamping the fuel return line permitted the engine to run. Swapping the injector with a used known "good" injector and seal permitted the engine to run with normal fuel pressure (+14 psi). The old injector seal was crushed on one edge. Replacing it seemed to solve the problem. Presure remained high and stable. Now... the same scenario. The engine would start and run for two or three seconds; when it stopped, I could detect the odor of gasoline, which suggested flooding (that no longer occurs). The technician at the dealer indicated that the "clamp" trick didn't work in the parking lot, and that I might need a new injector. After working on the car, they determined that the injector was firing, and the ignition system appears to be OK. Holding the throttle open while spraying starter fluid permits the engine to run. The current suggestion is to replace the computer, and assume that last month's the leaking o-ring simply was coincidental with an intermittent computer problem. (I do not know if they checked fuel pressure, the o-ring or swapped the injector for a good one this time around, but I assume they did.) One additional bit of information... After the car shuts down, it does not restart unless the key returns to the "off" position. Otherwise, it just cranks. The implication is that the computer is being "reset". That sequence also occurred last month. A year ago, they indicated that they were baffled by the problem, and a straightforward solution still is not apparent. Any clues? I'll pass them along today. Thank you for reading the post. Mike in RI