Hi,
I recently had an experience with a 27c256 chip. I did some tuning
and corrected the checksum, and everything seemed fine. Dynoed the
car, and ran it around for a while, no problems. The car went to
the track, and after 5 minutes shut off. Until it was towed back to the
pit, it started and ran fine. Could not see anything wrong. Went back out
and did the same thing, 5 minutes or so, almost like it was on a timer.
Took it back to the shop, and checked wiring, power and grounds, charging,
and anything that could be thought of in two days. Also it never even fluttered
even moving connectors etc in all this time. Went back to the track and did the same
thing. Driver input proved to be little help. After the third time I was able to get to
the car in time, and it had the symptoms of a bad chip. (relays clicking, did not start, etc)
Took it back, and made a new chip, same program, car worked perfectly. I have read
that this can be caused by not erasing long enough or a bad chip itself. Anyone have any feeling
on this subject. The ECU is a Motronic 2.1. Thanks, Patrick

I don't think that the long-time-no-erasing could be the problem.

More like a faulty contact somewhere in the relay box, a bad solder in the ECU (on this chip?) or something similar (only god knows what). :?

Anyway, if it ain't broken, it can't be repaired, so leave it until it works! :thumbup:

I finally got a chance to look at the trouble maker chip. By the way the car has not missed a beat since replacing the trouble
maker chip. I first read the chip and there were no errors in the program. Thought maybe I did something wrong. Then I
heated up the chip with a hot air gun to about 110 degrees F. Reread the chip and there were many bytes that had changed,
probably at least 100 bytes changed. Then I let it cool down, and it read the program perfectly again. This also went right along with
the sympton the car had on the track.
So I erased the chip and it would not program again. My conclusion is that when the
chip heated up (15-20 minutes) it electrically failed, and corrupted the code. Never seen this before, but maybe it will help
someone else. Patrick