10 Apr
2024
10 Apr
'24
6:51 p.m.
I sort of doubt any of these boards were factory supplied this way, but the date codes on
the ram in question are consistent with most other ics . the other banks contain chips
that are months older, or newer.
64-256KB SYSTEM BOARD
18 TI gold capped 4164-20 chips in banks 0 and 1.
Mix of Fujitsu, TI, NEC chips in banks 2 and 3.
There are a half dozen numeric codes present on the board. I don't know what any
signify.
11 Apr
11 Apr
8:55 a.m.
Seems logical enough that someone at the time bought a 128K machine and stuffed in a bunch
of ram chips they may have had laying around at the time. Memory was kind of expensive -
especially new from IBM. That's totally something I would have done at the time (and
probably did, but it was likely an XT for me).
Joe
On Apr 10, 2024, at 9:51 PM, Just Kant via cctalk
<cctalk(a)classiccmp.org> wrote:
I sort of doubt any of these boards were factory supplied this way, but the date codes on
the ram in question are consistent with most other ics . the other banks contain chips
that are months older, or newer.
64-256KB SYSTEM BOARD
18 TI gold capped 4164-20 chips in banks 0 and 1.
Mix of Fujitsu, TI, NEC chips in banks 2 and 3.
There are a half dozen numeric codes present on the board. I don't know what any
signify.
11:01 a.m.
New subject: 5150 mobo?
IIRC, there were two main models of 5150, and a few sub-models.
All 5150 were five slot. (5160 (XT) had 8 slots)
There was the "16-64KB" that had one row of 4116 soldered in, and three
rows of sockets. It could be purchased with those other three rows
populated, at a rather high price for 4116s, or could be purchased with
those three rows unpopulated. Both Apple2 and TRS80 used 4116s, so the
competition had driven the prices way down. Similarly, the Tandon TM100-1
disk drive was available very cheap in the TRS80 after-market. So, the
cheapest way into a 5150 was to buy a minimal system, an FDC board, and a
CGA video board, and provide your own 4116s, TM100 drives, and a composite
monitor.
At Merritt College, we had had a PDP11, with an aftermarket drive, being
used for not much more than teaching FORTRAN and COBOL. The
second time that the machine was down for most of a semester, the college
sold it to the Richmond School District (now "West Contra Costa"), and put
a few doaen 5150s in its place. While far from comparable, there were
never times when there wasn't a computer available for Fortran, COBOL, and
then also BASIC.
When Richmond installed the machine, something "went wrong", quite likely
confusion about delta vs Y three phase power. The official (coverup)
story was a "lightning strike" (at that time of year??!?), and PG&E paid
for a replacement machine.
S, everybody got what they needed.
The 5150s were picky about the RAM. Some types of RAM chips would not
work in it, although would work fine in Apple, or "memory tester"s.
At one point, the college bought some RAM from Fry's, that did not work.
But, at the Fry's store, they retested it and insisted there was nothing
wrong. We escalated. Fry himself came up to Oakland to bring RAM chips
that worked on our 5150s.
Then, there was the "64-256KB" motherboard. It had one row of 4164s
soldered in, and three rows of sockets. Populating those with 4164s gave
you 256K of RAM. BUT, there was an empty socket on the board, that you
could populate; I don't know whether it was a PAL or some 74xx logic, that
then let you use two rows of 4164s (one row of which was soldered in) and
two rows of 41256, giving 640K! 640K was all of the RAM that could be
easily used, other than some upper memory space of the other video or
bits in between other stuff.
We sometimes referred to the two types of motherboards as "16K" and
"256K"
to lessen ambiguity.
--
Grumpy Ol' Fred cisin(a)xenosoft.com
11:21 a.m.
New subject: 5150 mobo?
On 4/11/24 11:01, Fred Cisin via cctalk wrote:
Then, there was the "64-256KB" motherboard.
It had one row of 4164s
soldered in, and three rows of sockets. Populating those with 4164s
gave you 256K of RAM. BUT, there was an empty socket on the board, that
you could populate; I don't know whether it was a PAL or some 74xx
logic, that then let you use two rows of 4164s (one row of which was
soldered in) and two rows of 41256, giving 640K! 640K was all of the
RAM that could be easily used, other than some upper memory space of the
other video or bits in between other stuff.
Bipolar PROM. A few years ago, I published a way for one to use a 22V10
GAL as a substitute to fill in the D000 segment as well as the lower
640K by using 256Kb DRAM. Although the 22V10 is a 24 pin DIP and the
PROM being replaced, 16, things could be arranged to let the "tail" end
of the 22V10 hang out of the socket and have no changes made to the
planar traces themselves. So completely reversible.
--Chuck
39
days inactive
39
days old
3 comments
4 participants
participants (4)
-
Chuck Guzis -
Fred Cisin -
Joe George -
Just Kant