Restoration: how far should it go??

From: Eric Smith <eric_at_brouhaha.com>
Date: Sun Jun 15 01:54:00 2003

"jamesd" <jdickens_at_ameritech.net> wrote:
> I see you have higher standards than IBM, for a while after the AT was
> released they sold ram that was two chips piggy backed so they could get
> more
> ram on the board.

I don't think that's a sign of poor standards at IBM. The IBM parts were
specifically engineered to be stacked. They didn't just take normal DIP
RAMs, stack them up, and have some tech hand solder them together.

The stacked DIP packages were a replacement for the earlier Mostek dual
RAM packages that had two ceramic leadless chip carriers soldered to a
ceramic DIP substrate. The stacked DIPs were less expensive to
manufacture because they could mold both packages and simply use longer
lead frames for the top, but still fundamentally use their normal plastic
DIP production flow. The only additional manufacturing step was to
ultrasonically weld the leads together. (At least, I think that's how
they did it. I'm not 100% certain, and I don't have any close at hand
to examine.)

Note that the two stacked parts were bonded out with a slightly different
pinout, since the RAS and CAS strobes had to be on different pins.
However, since the top DRAMs were being bonded to a special lead frame
anyhow, this was not a big issue.

Both the CLCC on substrate and the factory-stacked DIPs were actually
quite reliable. Until ZIPs, SIPPs, and DIMMs were invented, that was
the highest packaging density in common usage for RAM.
Received on Sun Jun 15 2003 - 01:54:00 BST