The START and STOP signals
So we seem to be making good progress but one thing that we had "left until later" were the START and STOP signals. These are represented by holes in the tape between two existing rows (which effectively makes the tapes six channels, not five). We had tested the amplifier chassis (40 valves in each) and were happy that all the signal looked good. However, the START and STOP signals are now causing worry; they are either not there at all or there all the time (every sprocket hole) which is wrong.
Digging around revealed that some of the power supply rails were now too high by quite a large amount. This was caused by the changing of the 20 output stages which now draw less current. That is the next task, more news will follow ...
It never rains but it pours. The modified power supply worked well with a lower current demand but there was another problem. We found that some of the photo sensitive detectors were virtually dead; they weren't giving any output, or very little. Our spares were down to three and we aren't aware of any other such detectors anywhere in the world. Drastic action was needed!
We calculated that modern silicon photo-diodes could be used, without changing anything in the amplifiers, particularly the 125 volt bias supply. This would mean that we could use the existing devices that still work and replace them with the new version when they fade out.
Charles works next to a company that specialises in glazing products. The boss (Simon) looked a little surprised when Charles explained that he wanted to "cut the tops off twenty-six valves - how could I do it?". Simon rummaged around in a store cupboard and produced a small hand-held circular saw complete with water lubrication bottle. It was tried and was successful. It worked just fine, thanks Simon!
Twenty-seven "failed" cells had their glass bodies carefully cut off. The photo-diodes were soldered into place along with a resistor. Everything was then put in the dishwasher and given a good clean. The glass tops were reattached first with a drop of superglue, then with a bead of two-part adhesive. It was a tricky and laborious job but the end result looks pretty good - the untrained eye wouldn't know what had happened. The photo shows a before and after picture of one of the cells. The altered one is one the right. Testing them is the next job.
A week further down the road and the new photo detectors seem to work very well. Attention is now turned back to the mechanical handling of the two tapes.
Three excellent engineers - Mike Hillyard, Paul Kellar and Steven Brooke - have been doing further work on the tape transport system. Because the tape is physically "dragged" along by the tiny sprocket wheel, it gets damaged after a short time. The metal supports and tape tensioners were produced quite a long time ago and it was decided that these should be improved. Enlarging part of an old (1943) photo gives some idea of how the tape length was accommodated. The two photos below show the original way (left) and the new hardware made by Mike, Paul and Steven (right). The new large aluminium verticals were also made and installed. We can now adjust the length and tension of the two tapes very easily.
The project is now looking rather good with its new tape handling system and new photo-detectors. We shouldn't lose sight of the fact that Robinson never did work very well and that our reconstruction will probably suffer the same problems. We should also remember that when it was designed and built in the early 1940's it was done under enormous pressure and it is a great tribute to the engineers and mathematicians who did it in such a short time scale.