I located the tubes of AMD PALCE16V8H-25/4 parts. These were in three separate tubes stored in the same box as the prior mentioned Cypress parts. These have date codes that appear to be a mix from 1994 and 1996. (I say "appear" because AMD had used some gazootnik type date code system ... but I recall the timeframes of the various prototype runs these parts had been purchased for). This collection had 19 parts. Three parts had green stuff growing on the lead frames right next to where the leads exited the black body plastic. I threw those away!) The remaining 16 parts all passed an inital blank check in the Bytek programmer. I performed an erase cycle on a few parts and re-verified the blank check and this also passed. So at least for this small sample it looks like AMD GALs maintain their erased state in long term storage.

An interesting side note. The erased AMD part shows all 1's in its erased fuse map edit buffer. This is just the opposite of the Cypress part.

Michael Karas

PS. I have AMD PALCE22V10-25 parts to check but that is for another day.