On 3/22/22 21:32, dwight via cctalk wrote:
Years ago when I was at Intel, they would take EPROMs
that had only a few defects and attempt to program them with specific applications. They
would then usually paint the window black and put on that particular product number. When
we needed EPROMs in the lab, we would take a pile of these that were no longer used and
scrape the paint off the window. Some would take the desired program and some
wouldn't.
I suspect that they can fuse identify specific bad EPROMs to use for such harvesting. The
fused value may be laser blown as well. My guess is that that is the reason for the
different ID number. If it will take pre-programming with a particular data, it is worth
harvesting.
This way they can find value in an otherwise unsellable product. I doubt the harvesting
methods have changed much over the years.
That's not surprising. Probably cheaper even for IBM than
mask-programmed PROMs.
--Chuck