You will see from the connectors to the Punch that this photo was taken prior to conversion to 64-column. Several of my unofficial modifications are visible:-

1. Four more matching knobs on the Punch specify a 4-digit Case number/Extension of Job number on card columns 1-4.

2. A switch above the Minor Cycle Slip button splits the vertical scan of both monitor tubes into four groups of 8 minor cycles, easing identification. It operates via a relay clamped to the handle of Unit Y (or was it X?), which switches resistors in series with the adjustment pots of the vertical ramp renerator.

3. A knob under the left hand monitor tube splits the horizontal scan of both monitor tubes into Instruction groups NIS, Source, Destination, Characteristic, Wait, Joe, Timing ; or into Data groups of 5; or into Data groups of 8. This is achieved by a small chassis carrying many diodes forming OR gates from the appropriate Q pulses, and a few valves to buffer and select, built into the gap between Units D3 and D4 (where the signal tray stops short). Control is all neatly fed via spare cores in the 24-way muticores from the Painton plugs of the Junction Panel, while the output pulsed line is via a new wire fed into the signal tray to Unit X (or was it Y?), where just a few components have been added to the tag strips. Each pulse so received then pumps a step into the horizontal ramp.

4. To the right of the I.D. keys a push button has been added. When inserting a program manually, for each instruction we press various P keys (lighting the lamps above), culminating in P32 (the GO digit). Then we give a Single Shot. Then we press the Clear I.D. key in preparation for the next instruction. This sequence gets repeated very many times. It occurred to me that having to raise the hand to the Single Shot key in the vicinity of the telephone dial was inefficient. It was ergonomic to have a further Single Shot commander immediately between the P32 and the Clear I.D. keys, and I chose a push button for this to give speed of operation (and ease of installation!)

5. The changed colours of the Instruction Stat lamps are hidden in this photo, but the Power indicators above the right hand monitor tube show where the replacements were obtained!

6. On the left upper fascia panel are the magnetic track neon indicator repeaters. I can just make out that the Read heads are in Position 4 and the write heads are in Position 1, and the last transfer was a Read by Head 7 (i.e. from Track 71).

7. Knowing what to look for, I can make out that a faint smudge in the card stacker of the Hollerith reader is a matrix of holes that I drilled, covering them with grey cloth. I mounted my loudspeaker behind. A horn loudspeaker system! Well, where else could I put it tidily? Again, spare cores were used in the multicore cables, conflicting with Kidsgrove drawings.

8. Finally, on the main fascia I put a clear-faced clock. This was deprecated by Derek Royle, as compromising the Junction Panel behind. But operationally, for timing job throughput and filling in the log book, that's the best place. The clock shows that the photograph was taken in the evening; this was because special lighting was bought in, which would have impeded the hurly-burly of daytime operations. The workload was heavy and urgent. Like everyone else, Barbara quite often stayed late despite having two buses to catch for her 10-mile journey home.