Hi Bryan,

> We've been selling a product based on a 16C57 for a couple of years.
> Suddenly the same code fails to run properly. But it's more weird than
> that. The old code, when burned into an eraseable device works fine, but
> not in the OTP parts.

Have you tried NEW JW parts with the same date date code as the OTP
parts that won't work ?

> Did I hear than Microchip had changed the die on these parts a few months ago?

Dunno.

> We're using the PIC to output address data to a 27C010 Eprom. The addresses
> get clocked out okay on ports RB and RC, but all output bits on RA seem to
> get set to zero at the very end of the sequence in OTP parts, but NOT in
> eraseable parts. In debugging, I've explicity set RA.2 to 1, but it goes to
> zero anyway as if the code were shutting down before it finished the loop.

Is the watchdog kicking in early, try and disable it completely in the
programmer for a sample chip and see if this copuld be the problem.  The
process variations for the WDT are very loosely characterised and you may
be working at a optimistic point especially since you indicate the fault
occurs at the end of a long function.

Also the chip might have supplemental Brownout stuff that is more
sensitive to power dips.

Cheers
--
Kalle Pihlajasaari   kalle@ip.co.za   http://www.ip.co.za/ip
Interface Products   P O Box 15775, DOORNFONTEIN, 2028, South Africa
+ 27 (11) 402-7750   Fax: 402-7751    http://www.ip.co.za/people/kalle

DonTronics, Silicon Studio and Wirz Electronics uP Product Dealer