Paper tape and photo cells
Robinson paper tape and photo cells
Robinson’s task was to read two synchronised paper tapes at the same time and to make decisions depending on what was on these tapes.
In the 1940’s paper tape was used by teleprinter machines. Teleprinters were something like typewriters, but they had the ability to communicate with other teleprinters anywhere in the World. If this sounds familiar then this was the 1940’s version of the internet!
The paper tape was quite narrow: about 2cm wide punched with holes 2.5mm apart. Teleprinters read the tapes by poking five mechanical feelers at the tape, if some went through then this showed that there was a hole in those positions. Alphabetic letters and numbers were encoded as different combinations of the five holes. (An interesting exercise for the reader is to find out how the 26 letters of the alphabet, the digits 0 to 9, and lots of other things like space, colon, comma etc were all coded using just five holes in the tape.)
Reading at about ten characters per second was normal in those days. The mathematicians who specified Robinson wanted to read the tapes at 2000 characters per second, which meant a very different method was needed.
A motor was used to drive the tapes (they were in a loop with the end of each tape was glued to its beginning). The really difficult thing was to sense the holes in the tapes: mechanical feelers were not a feasible option at these speeds. Arnold Lynch designed a photo-electric system using a lamp, lenses, masks and photo-electric cells. The documentation suggests that Lynch didn’t know what the final use was; secrecy was vital.
The photo-cells look like glass valves. The 20p piece gives a sense of their size.
Robinson read two consecutive characters from the paper tape; each character had five data holes and a sprocket hole. To read two characters (2.5 mm apart) needed two lots of five plus one sprocket hole (there was no need to read two sprockets) that is eleven photo-cells! Oh, and there were START and a STOP holes as well. That's thirteen photo-cells to read a very small piece of tape.
So how was it done?
The photo (left) shows one of the two chassis we have made; they are pretty much the same as the original.
Can you spot a slight error? If you count the photocells you will only count twelve, the thirteenth one has yet to be mounted. The spacing up-down and left-right is very critical; we have spent a long time working out exactly where the cells must be. Don't forget, we are magnifying two small groups of holes many times and any errors will also get magnified. That's all the more credit to the designers and engineers in the 1940s and the very short time-scales they worked to.
The chassis sits behind the tape drive and lamp, as in the photo (right).
The optics consist of a low voltage lamp bulb which is mounted close to the paper tape. Behind the tape is a lens (from a 35mm slide projector) and this focuses and projects an image of two characters onto the bank of photocells, as shown in the photo.
The two paper tapes run over pulleys. The massive steel frame, known as the Bedstead, supported the two tape systems. The original Robinson used wooden pulleys, but we had other sets of the aluminium pulleys which were made for the Colossus Rebuild. (Robinson pre-dated Colossus of course.)
We had little to go on as to what sort of motor would have been used in the original, so we scratched our heads and guessed that, because the Post Office was designing the system, a teleprinter motor would probably have been used.
In order to get variable speed a controller was built based on resistors and relays. It works well. See photo (right).