Pulse Generator Control
The pulse generator starts when V5 is triggered by a pulse which originates from the carry check line which is connected to all terminals A1 in the carry plug-in units, and to socket 4 where the pulses can be inspected. The pulses are only available when all the carry circuits have passed on carry pulses to the next decimal place. The over-fill check and circulation check circuits effectively short-circuit the carry check line if the arithmetic circuits have discovered that a problem exceeds the store capacity. The cancel check key isolates these circuits for test purposes. If the pulses are available at socket 4, they should, in the normal operating condition, appear at socket 6 and if relay GST is pressed manually V5 should strike to start the arithmetic operation.
The main problems associated with this circuit have been in maintaining the correct voltages across V2 and V3, and originally rectifiers W1, W2 and W3 were of the copper oxide type, but were replaced to prevent errors due to leakage. The bias on V3 is rather delicately held, relying on the stability of the -180V supply with pulses and the -200V line, and the total current of the right-hand side depends on the bias of V2, although the left-hand side of V2 is normally cut off until the arrival of a stop pulse, and if one of these valves fails it is best to replace them both to maintain some equivalence in cathode emission.
A few years ago this section of the control circuit was checked very carefully and all resistors and (capacitors measured as it was discovered that because of their age some values had changed significantly, and at the same time we replaced the 6 J6 by an industrial type with similar, but much more reliable, characteristics.
At one time I considered modifying the cathode circuit of Y3, rewiring RV1 as a potentiometer to give more reliable control, but it would have created further problems of heat dissipation. One modification which helped considerably was the addition of a zener diode between the -200V and -180V lines in the power supply unit, as previously the 20V difference varied depending on the load. Perhaps a few more zener diodes (which had not been invented when the computer was designed) would make the circuit much more reliable.
Cecil Ramsbottom 13/6/1974