Thank you all for your replies !
The AVO's I keep for this have 2.5 or 3 kV terminals, yours not?
There should be test points for waveforms on the primary side of the HV inverter, and also make sure there is supply to the primary side.
Nope, I have a Metrix 462C and there is a separate socket for each range, the highest one is labeled 1000V. I should have thought about measuring 1-5 kV
voltages before picking up an AVO.
In you second image a few posts back there are a few orange Tant caps...... renown for failures, others might advise replacement but lets get "her" going first.
Those are absolutely all over the place, if I start replacing them, there will be like 150 caps to change. I'm not sure I'm ready to do all this unless I see at least something on the screen.
In this picture, the V-board is also not present as you can see. If this is what you then see, replace the HV multiplier if you can find one, or make your own.
I'm not sure yet, so I don't think I will be touching the HV circuit unless there is strong evidence that it's at fault.
Pull the "Chopped Blanking" cable out for testing. The little toroid transformer is part of a blanking circuit that feeds the z-axis (intensity) section. A fault there can result in the trace being permanently blanked.
You were absolutely right ! With the damaged toroid, chopped blanking was doing something bad to the trigger circuit. When I pulled the chopped blanking cable like you said, the trigger panel suddenly became a lot less dumb. Now the TRIG light blinks periodically and "single sweep" does what it's supposed to. However this wasn't enough to get any visible trace. There must be another problem.
I noticed yesterday, with lights off, that the screen was slightly green, a sort of halo, so I assumed that the bias HV was somehow working (I don't have the equipment to properly measure it) and continued the troubleshooting with that hypothesis.
According to the troubleshooting guide, I have a problem inside the "Sweep Start logic circuit". I will continue working on it tonight. I only hope that the culprit is not U870, that would be a bummer for a scope with so few ICs !
Edit: I have a question: when they say "check QXXX" are you supposed to characterize it with a curve tracer ? I don't have one, so I just checked both junctions with a DMM in diode mode. If both junctions in a BJT are OK, can we say that the transistor is OK in general ? I suppose there may be some weird defects where the junctions are ok and the transistor still not doing its job, but I have no idea how frequent those are and what are the testing practices in the field.