[GreenKeys] Model 33 bug

[Jones, Douglas W]

The model 33 we've got for our PDP-8 at the U of Iowa failed recently.

The symptom:  Chattering buzz from the selector magnet immediately when plugged in, even when turned off.

Our initial thought was something in the selector driver circuit, but swapping that board with the one from the TTY in my basement made no difference.  A problem with the big transistor mounted on the selector magnet driver frame?  Measuring the voltage across that showed that that transistor was busily switching at 60 cycles.

The selector magnet power supply uses a full wave rectifier, so failures there would tend to give 120 cycle buzz.  The local loop supply, though, has only a half-wave rectifier.  Checking the voltage across its capacitor shows that the capacitor still has some capacity, but not nearly enough.

Our hypothesis is that over the 60 years since the TTY was new, that capacitor has simply dried out.  As it dries, the capacity would go down until it eventually gets too low to sustain a sufficient local loop current.

Scope traces for this adventure are here:
-- https://homepage.cs.uiowa.edu/~dwjones/pdp8/UI-8/log2025b.shtml#2025-09-22

All the power supplies in the Model 33 seem to be floating, and the only way to make useful non-invasive measurements seems to be in differential mode with a 2-channel scope.

              Doug Jones