Some news, but not what you may expect.
As there were some doubts (and I agree, for good reasons), I opened the display area again; soldered two wires to shutter connections - totally random, I have no idea for which colors and I believe it's not relevant. Shutter connector in normal place, put everything back and started the scope. Here is the moment to say, for Siggi, that drivers for the shutter are indeed powerful, initially designed to drive (servo) motors at 20kHz, max 0.5A DC.
OK, see pictures 5 to 8 in the album.
Synchro signal on Ch1, frame, as I said before, around 17 ms (by signal I mean waveform).
Picture 5: on Ch2 (should be green but, at low light, camera doesn't make much of a difference between colors) is signal from output of driver for one of the shutter's colors.
Picture 6: same signals, different timing.
Picture 7: same as 6 but signal is displayed on Ch4, not Ch2. Some vertical vectors are obviously of different color (naked eye sees clear red). If Ch4 light is OFF (pressing Ch1, for example) signal is all blue (no red vectors).
Picture 8: Same as Picture 6, different shutter channel ( different driver output/color).
1. Shutter works fine; period of signal, roughly 370uS or so (just to have an idea how often the shutter refreshes the colors - definitely not line frequency). As I don't have the pair signal for the same color I cannot say if they are really opposite; it may be that some time they are not (and that color is not ON), depends how the control signals come from the memory.
2. Some waveform information for Ch4 is incorrectly displayed when a different color (driver output for red) is ON (and on condition that light of Ch4 is also ON). This back the same conclusion it was suggested before: Between the point(s) of split from VGA the signal path toward Video for CRT is corrupted. Still to be discovered where and how...