- Fireworks and Slot Machines
- Fractals – Mathematical Art
- Pythia – Natural Language Processing
- UltraKiss – The Kisekae Set System
- Waves – Navier–Stokes and Graphics
Throughout my career I have worked on many computer information projects, large and small. My latest personal work is listed above, but if you read on you can learn about the time before personal computers when I actually built my own fully capable CPU.
In the late 1960’s while a teenager I worked with vacuum tubes, transistors, integrated circuits and digital logic. Before graduating high school in 1970, one of my long-time school friends and I undertook to build our own CPU in my basement. In time we each moved on to other things, but at the end in 1978 this machine was fully operational on my workbench. A simple program could be entered through the front panel switches. We had wire-wrapped together many 7400-series TTL logic gates, flip-flops, LED’s, and other components on two vertically layered perforated boards that were each about 16 inches square. We had developed and burned in ROM a full set of micro-programmed instructions for arithmetic commands, jumping, testing, machine state changes, and so on.
The computer had 16 general registers segregated into 8 user registers and 8 system registers, where each register was 18 bits in size with two parity bits. There were 2048 words of 48-bit microcode ROM that implemented over 100 different machine instructions modeled after the PDP-11 architecture. The instruction set for dual register commands had a 4-bit opcode and used two 3-bit fields for source and destination registers, leaving a 6-bit address offset. Single register commands allowed for a 7-bit opcode and commands that did not require a general register specification had a 10-bit opcode. The power supply delivered about 20 amperes of current at 5 volts.
After I graduated from University I had opportunity to acquire a surplus Univac DCT-2000 data entry terminal from my employer. This was to be my computer desk. The DCT-2000 came with a punched card reader and full sized line printer. My friend had earlier acquired a 5-bit baudot Model 33 teletype and paper tape reader to use as the system console.
I really need to find an old picture of this! But, there were no CMOS digital cameras at this time? I may have a picture on film or 35mm slides?
Regretfully, I never completed building this personal computer. Main memory was far too expensive for my limited budget at the time. By 1978 I was moving on to other things. Unfortunately, I didn’t know about business, venture capital, and all the other things that were necessary to bring ideas like this to fruition. I can only imagine what might have happened had someone understood what had been accomplished? We were a decade ahead of Bill Gates in this regard.
I ended up destroying my work by 1980. I had moved, married, and the DCT-2000 was in storage and costing money. I eventually scrapped it. I kept only the two 1024-bit ferrite core memory cards that were used as buffers for data. I mounted one card on a plaque and kept it on my desk until it was stolen in 1985. Someone in the Univac office had an eye for computer history. The other I still have.
In hindsight I now wished I had kept more of my old equipment. It is part of my history and the history of computing.
And, as we are in to nostalgia, I should show you the microprocessor chips I have for the Burrough’s A-series computer. This is called the SCAMP chip (Single Chip A-Series Mainframe Processor). The SCAMP chip had 10 million transistors and was one of the first VLSI desktop mainframe computers. This chip essentially implemented the dream I had 20 years earlier. I acquired these chips for teaching purposes in 1989 while working for Unisys. The photo shows a populated chip and a chip blank.