Old 3M Computer

From: Rick Bensene <rickb_at_bensene.com>
Date: Thu Jul 17 01:14:01 2003

Hello, all,

Back in high school (1973-1977), the school I attended had an old
computer that was donated to the school.

The machine was quite interesting. It was a transistorized machine,
using a magnetic drum for main memory. It had a bit serial
architecture, with a 24-bit word.

The drum held 8K 24-bit words. Instructions had 5 bits of opcode.
Each instruction contained the address of an operand (track and sector),
and the address of the next instruction (track and sector). The machine
had a 'hardwired' read-in mode that accepted addresses and data as octal
numbers from the console ASR-33 teletype. The input format was address,
followed by a space, followed by data for that address, followed by a
carriage return. All other characters were ignored. The machine could
accept data from the paper tape in read-in mode at about 8 characters
per second (slightly slower than the full rate the TTY was capable of).

The instruction set was fairly basic with two accumulators (A and B),
2's compliment arithmetic, addition & subtraction, various logical ops
(OR, AND, XOR, as I recall), load and store, and various I/O functions.
No index registers, no indirect addressing. These features were
implemented by instruction modification. The instruction set included
suggestions for "optimal programming", which suggested, for example,
plaing the operand at the current address plus 2, and the next
instruction at current address plus 4, to allow for drum rotation during
instruction decode and operation processing. Master timing for the
machine was generated by a fixed timing track on the drum.

The front panel was very simple...there were five lighted (by neon
bulbs) pushbuttons labelled LOAD CLEAR START STOP and STEP. The LOAD
button placed the machine into read-in mode. The CLEAR button cleared
out the major registers. The START button started execution at the
current location, and STOP halted the machine after the current
instruction executed. STEP executed a single instruction. I can't
recall if the machine had any interrupt system. There were five neon
indicators that would display the op code of the current instruction.

The machine was made by 3M. The system consisted of two CPUs, a
"Primary" and "Secondary" , which were identical. These two CPU's fit
into two drawer units in a small rack about 3 feet tall, a standard 19"
relay rack. Each CPU was about 10RU. The circuit boards plugged in
vertically arranged in a "U" shape, around the drum which was in the
center of the chassis. The drum was a fixed head per track drum, with
electronic track switching. Another 6 foot tall relay rack contained I/O
equipment, which included an interface for an ASR-33 teletype, as well
as a large bank of counters which were part of a data acquisition
system. There was also a register that connected the two CPUs together
to allow the CPUs to communicate with each other. Also included was a
real-time clock included in a control panel that used projection-type
incandescent indicators to display the time, and was interfaced through
a 24-bit register that could read the time. The clock was rackmounted
in a standard relay rack (19"). The clock was made by another
manufacturer, apparently OEM'd by 3M. It was about 8RU, and used
transistorized 10-stage ring counters. Also interfaced (but we never
were able to make it work) as a large-carriage IBM output writer (not
Selectric, used indiviual type hammers like a regular electric
typewriter).

The story about the machine was that is was used as a data acquisition
system for monitoring gas line flow/pressure by Northwest Natural Gas in
Portland, Oregon. When the system was decommissioned sometime in the
early'70's, it was donated to the high school.

When I started high school in '73, the drum in the secondary CPU had
suffered a head crash (bearing failure), which made the secondary CPU
useless. The primary CPU ran great. I spent a lot of time tinkering
around with the machine, learning a lot about machine language
programming.
I even wrote a simple BASIC interpreter for the machine..a fun exercise,
but pretty pointless as the machine was so slow that it would take up to
15 seconds to simply syntax a newly inputted statement and store it in
the user workspace.

I went back to the high school in the early 80's, and the machine was
gone. The story was that the primary drum had crashed, and since the
machine was useless, it was given away to a scrap
dealer.


Does this machine ring any bells with anyone?

Rick Bensene
Received on Thu Jul 17 2003 - 01:14:01 BST

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:36:05 BST