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