My plan was to take the best parts and build one working system, but it has gone awry. I've ended up with four RF source/frequency control units and one display/processor unit.
During my repair adventure I rebuilt two of the optical rotary pulse generators (we'd call them encoders now, I think) with LEDs instead of filament lamps and 74LS instead of 74 series. This took the power requirements for each encoder from 5V at 170mA to 5V at 17mA.
I also discovered what must be a design fault in the sweep generator, which destroys two 1.1kΩ resistors responsible for passing current through a stack of Zener diodes that form a -13V reference. Two machines had lost the reference completely with both resistors open-circuit and a further machine was heading that way, with the resistors a long way from 1.1kΩ
The strangest fault was in the 8505A that I already had, a LM339 quad comparator on the front panel had corroded, perhaps because a nearby wet tantalum had leaked on it. This caused the "overload" reference voltage to be absent, which made the warning annunciators for all three inputs show overload.
One unit had suffered a failure in its negative voltage regulator board, blowing a few fuses and another had lost the 20dB section of its attenuator. I consolidated these broken bits into the fifth unit, which also donated a few bits of case hardware.
Three of the four RF systems are "fully loaded" with the label generator card and the phase lock option but I only have two display/processors with the label generator option and one of those has a non-working HV supply.
Here's the stack of working bits (I don't know if the 8501A and 8503A work fully yet as I don't have interconnect cables). Yes that is the wrong bezel on the top one, I'm going to make a crude laser-printed bezel for it for now.
![Image](https://live.staticflickr.com/65535/53877885816_5d0a7fefdc_b.jpg)