I finally managed to loosen the screws holding PC148. The actual issue was that the screw heads had started cold welding to the chassis. Thus, the probability to be able to remove the screws only by using a screwdriver was exactly zero. The trick was to get a 6BA open ended spanner and give the four hexagonal mounting posts a 1/4 clockwise turn. Then the screws came off easily. The only mounting post that also could have been turned with a pair of pliers is the bottom right one. The other three, however, ... no way.
After that, I could replace the 10M resistors on the brightness chain with VR37 ones and the scope worked again. It still has small occasional sudden brightness changes, but nothing critical. Maybe this will be fixed, once I manage to find a way to also replace the resistors on the focus chain.
With the D75 working again, I continued to debug the D755. Unfortunately the issue seems to be kind of a heisenbug (it disappears if you try to observe it). It can run for hours, without the issue showing up. However, it seems that the issue is not in the bright-up circuit on PC128. Its output signal is exactly like expected, a square wave with the lower level at 6 V and the upper level depending on the set brightness. Also the bright-up output signal does not change when the issue occurs.
Next, I examined the low voltage side of PC148. The amplitude and the DC offset of the oscillator signals at test points 1 and 2 seems to be right. The frequency, however, is not at 28 kHz but at 40 kHz. In addition to that, the sawtooth at test point 3 has no flat top and its amplitude is too large. Unfortunately, I could not observe yet what happens to these three signals when the issue occurs. I have to find a way to trigger the issue. I does not seem to be directly heat related, since heating PC148 with a hair dryer could not trigger it.
At the moment, I am kind of expecting something on the high voltage side on PC148 to draw too much power, driving the EHT supply out of regulation. All in all the issues look a lot like the issues @Anode_to_Joy described in
this thread. Unfortunately his solution was close to replacing everything on PC148.
One time, I could observe continuous arcing at the output of EHT transformer connecting to D303. It occurred on the back of the soldering post, inside the transparent, wax-like coating near the paxolin board. While it was there, the brightness of the trace was fluctuating.