This will change your mind.

A few years ago I attempted to upgrade my pc532 from 1MB simms to 4MB.
I applied the 74S10 patch and it did not work. I actually could not
find an S10 so I used an F10. Naturally, I blamed this chip.

I recently had a great urge to complete the upgrade, and finally
managed to source an 'S10. I also scavenged at the local tip and
pulled a dozen 4MB simms from discarded 486 machines.

But no beef. No minitor prompt. Nothing.

I removed the mobo and checked all the new wiring. All looks
OK. Well, the machine is working with the extra circuit for a
few years now so it probably is correct.

I installed the 1x8MB simms back and was screwing the mobo
to the baseplate when my eyes gazed at the chip which had
three consecutive pins left unplugged. It was done on purpose,
as this is the old /WE gate. I noticed that the chip I am
looking at is U18. But I could swear that it was U30 which
had the disabled gate... a quick check confirmed this to
be true. I reinstalled U18 fully and pulled pins 11-13 of
U30 outside the socket and tested the machine one more time
and surely it came up with the 4MB simms installed. It was
fully dead before, and was dancing now.

I booted NetBSD and saw a huge, 32MB memory reported. I am
now very happy.

Ho, about Dog. Looking at the pc532 logic I can see that the
pins I incorrectly pulled out are in charge of A10, a memory
line that is only used with 4MB simms. And not disabling U30
probably left a wired or that did not cause any problem.

Surely this cannot be an accident, there must be an intelligent
being, a very intelligent one, who planned this whole trick
on me. If the mistake was to make the machine unbootable then
I would probably have checked more closely back then and fixed
it. This way, I suffered for a few more years with only 8MB
memory. This being is both intelligent and cruel.

It must be Dog.

--
Eyal Lebedinsky (eyal@eyal.emu.id.au) <http://samba.org/eyal/>