The Emidec 1100 Computer
This page ©2012 Andrew Wylie all rights reserved
The EMIDEC 1100 was a transistor computer dating from 1959. It was made at EMI Laboratories by Professor Sir Godfrey Newbold Hounsfield, recipient of the Nobel prize for the development of computer assisted tomography, amongst many other honours. This is claimed to be the first large commercial transistorised machine in the UK, and Sir Godfrey's autobiography states that 24 of them were sold.
A former Emidec 1100 programmer, Ian Smith, has created a great Web site devoted to the machine, where you can learn about the hardware and software of this important early British transistorised computer.
I possess a number of logic cards from the 1100. The top image shows one of them in side view plus an oblique view of the front panel: click on the thumbnail image of the card to see a larger version (100KBytes). The card measures just over eight inches square, and has a wonderful front plastic panel with handles and thirteen holes which provided test points. Metal tracks on the card match up with these holes, so connectors from an oscilloscope could be pushed through and the behaviour of the circuit studied. This card is labelled "READ C/O GATES". It uses Mullard OC76 transistors (in the aluminium cylindrical cans on the right near the plastic panel) and OA10 diodes (the black painted cylinders between the column of transistors and the column of pale grey boxes). Other cards that I have are variously named "STAND'D COUNTER", "BUFFER LINK", "INPUT GATE", "END OF TAPE DETECT" etc. They use a variety of other transistors: Mullard OC42 and OC83, GEC GET103, STC TK30C and even RCA 2N1170 and Philco T1872. The resistors on the card (the white cylinders with coloured stripes) are beautiful objects: they are 3/4 of an inch long ceramic cylinders filled with a resistive material. These really date from the period of thermionic valves (tubes to the Americans), when resistors needed to dissipate significant power, 1 or 2 watts each. You can't buy resistors like these any longer, modern ones are usually much smaller and only handle 1/8th or 1/4 of a watt.
The grey boxes are one of the most interesting aspects of this machine's technology: they are magnetic logic units. Each box contains a single ferrite ring, with up to fifteen connections to it. (At least, there are fifteen pins on the plastic package - I would have thought that only an even number could be used). The lower image shows one of the packages that has lost its lid, so that the toroid can be seen: it has fourteen wires leading to it, two of which are soldered to the same pin of the case. Each of the boxes on the card bears a different handwritten label (F49 in the image), and each appears to have a different configuration of wires. Unfortunately, the boxes appear to have been filled with a gelatinous substance, which has dried to form an opaque grey layer on the lids, preventing easy inspection. Making these devices, which were hand-wired bespoke units, must have been a tricky task.
In his autobiography Sir Godfrey states that he used the toroids to speed up the operation of the machine, because transistors were relatively slow at that time. By this means the 1100 achieved processing power comparable with a valve/tube computer. The 1100 was not the only machine to use magnetic logic. The later English Electric KDF9 used such devices, albeit simpler ones. There is little information available on how these circuits worked, so if anyone reading this knows about them, please
In the 1960's some all-magnetic machines were built, usually for mission-critical or military applications. Interestingly, a Web search shows that research continues in this field to this today, although now miniature integrated magnetic devices are being studied.
EMI went on to build the EMIDEC 2400, about which I know very little except that it was almost contemporary with the 1100.
If you know about either EMIDEC machine, or have circuit cards similar to mine, please