While repairing the set I found some errors in the schematic I had drawn. So here's the new version, also showing replaced components:
Now I did not have enough gain in the video output. This was put right by building a simple solid state video amp, powered from the 15VAC filament voltage of the PL83:
This solved the drive problem easily, as is shown in the previous page. Of course this is a huge cheat in such an old set but serves nicely as proof of concept until I decide whether to fix the RF/IF strip and build a modulator for 819 lines or not. Of course, this will also mean building an 819 lines (system E) video generator. If I leave it like this I have to fix the horizontal oscillator to compensate for the much longer sync pulse of a system B/G video signal (625 lines).
Some more gratuitous close-up porn:
The stupid bulb sockets are gone now and I added a new capacitor which solved the hum issues. Note in the last image the corona in the mica line linearity trimmer.