[Resending with proper tagging that I had forgotten the first time]

Last week a prototype of a 16F877-based instrument started acting crazy
(naturally, this was while
my client was demonstrating it at a trade show).  It was as if the code
was running properly most
of the time, but jumping to the wrong places occasionally (for example,
menu items were getting
written to the wrong line of the LCD).  When I got it back, I verified the
chip and discovered that
it verified at 5V but failed the low-voltage verify.  It was both
programmed initially and verified
on a Needhams EMP-11.  Erasing it and reprogramming it cured the problem.
The EMP-11 uses the
"production" algorithm, so the chip had to have passed a low-voltage
verify when it was initially
programmed.

Has anyone seen anything like this before?  The chip has been programmed a
few tens of times,
nowhere near the write limit.

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads