Advanced Features Microcomputer

After graduation in 1976 I worked seven years at Danfysik A/S, Jyllinge, Denmark. Around 1979 we developed a microcomputer with advanced features that were ahead of their time. The picture is from the little short-form sales brochure that someone made for us. As you can see, there were three things that were absolutely outstanding: Menu-driven programming, touchpad and in-circuit emulation.

In the late seventies microprocessors were usually programmed using a text-editor on a larger mini computer, and afterwards the text was translated/assembled into machine code on the minicomputer. When this task was finished, the text was printed on paper and at the same time the machine code was stored on paper punch-tape (I still have the old Olivetti that we used, and I am trying to restore it).

The punch-tape would then be sent to a company specialized in producing PROMs (read-only-memory chips) that you could plug into your microcomputer. If the code did not work, you had to repeat the whole procedure! Later we used EPROMs that we could burn and erase in-house, but that actually only saved the trip to the PROM-manufacturer, and the EPROMs were not very reliable in areas with radiation.

With our own new machine, the often cumbersome assembler-code programming was done by selecting commands from a menu, using the touchpad. The menu was context-driven in a way that only the correct sets of commands could be selected. The display would show the complete program, and you could test it right away, by using the in-circuit emulator. Notice that there was also both a built-in impact printer and a small cassette tape drive for storing the program.

It was the ingenious R&D manager Hans Erik Jørgensen who got most of the ideas. I worked as an assistant to him for extended periods of time, and often we would work late into the night with our crazy ideas. One of my own contribution was a smart dual access memory, that enabled code from one processor to be used simultaneously by another running CPU. This technique was used both for emulation with an external CPU, and internally for fast communication between the device’s two processors. I do not remember who came up with the idea of the touchpad, probably Hans Erik. The first versions we made, instead had a light-sensitive pen to point at the menu: A hollowed-out 0,5mm pencil with a photo-transistor inside.

It is part of history that our micro-computer was a spin-off of the microprocessor-controlled high-voltage supplies that we developed and sold to laboratories and accelerator centres. We simply did not have the budget to buy the “right” programming and development equipment, so we had to make our own. Microprocessors were introduced to avoid large amounts of traditional electronics, switches, potentiometers, etc. on the front panels of the instruments we made. Everything was based on the Signetics 2650 processor, which had a nice logical and efficient instruction set, context switching and good interface options without too many external circuits. 8 bit architecture of course, but unfortunately with too little addressing area (15 bits) compared to Intel and Zilog. Lately I found out that the 2650 was almost a copy of the IBM 1130 computer.

How did it all end? As with many other pre-PC initiatives of that era, it died. Since it was not exactly core business, Danfysik sold the design to Signetics, which in the same period had been purchased by Philips. Afterwards it was completely forgotten. I just found some of the parts in my drawers: the tape deck, the printer, some 2650’s and some compatible 2651 UART chips. Maybe some day I will build “something” with it (if it still works!):