The following scope pictures illustrate the behavior of the DC restorer in the CT-100, and also show why my picture turns green if the contrast is too high.
These are all shots of the red CRT grid signal. The brightness control and DC restorer (a separate diode for each CRT grid) set the level of the bottoms of the blanking pulses on the video.
With the brightness all the way down, the bottoms of the blanking pulses are at zero (bottom of scope grid):
The scope vertical scale is at 20 V / div. Increasing the brightness a little just moves the whole waveform up. In this case, so the bottoms are at 40 V:
Now the brightness is set to a good point (bottoms of blanking pulses at 60 V), but the contrast is still fairly low. Note that the red cathode is at 124 V, near the top of the screen, so things are still pretty dark overall:
Now I increase the contrast control, and you can see that the bottoms of the blanking pulses are still at 60 V (held there by the DC restorer) and the contrast grows the waveform upward for a brighter picture:
Here is the point of maximum usable contrast, and the picture is quite good, with good greyscale color tracking. The tops of the waveform are now at or slightly above the cathode voltage, so this is the threshold for CRT grid current starting to flow:
If I turn up the contrast more, grid current in the CRT pushes the whole waveform down. Since there is a 100 K ohm resistor going to the DC restorer cathode, the action of the DC restorer trying to maintain the bottoms of the blanking pulses at 60 V is now overcome by the lower impedance "grid leak" biasing of the waveform down by the CRT grid current:
What this means is that a lot of the red content in the picture is now being pushed toward black, even though a few of the brightest red areas are at maximum brightness. The effect of this is to cause the whole picture to shift toward green and goof up the color balance. Since the green phosphor is capable of higher brightness at lower current, the green is no where near this kind of grid current situation, so the green keeps getting brighter while we start losing red. The contrast control can be advanced further still, driving the bottoms of the blanking pulses all the way down to zero on the red (and still topping out at around 130 -140 V), giving an extremely green picture.
I suppose I really should have captured some screen shots along with the last two scope shots, but I trust you can imagine the problem.
So with these modifications, I now get a well behaved picture if I turn up to the contrast just to the point that things start turning green, and back off a little. Good greyscale tracking, and colors stay pretty true from dark to bright. The picture is not all that bright, and is best viewed in a reasonably dark room. Not very viewable in a sun-lit room by any means.
So I'm not sure why my CRT seems to need these circuit mods to work. There may yet be something wrong in the circuitry, but it looks from various voltage measurements and scope traces that the circuitry is doing what it should.
By the way, the tip on the white coils was probably a good one. I do still have a few of those white coils in place, but I checked them again and they're still OK for the moment. In the video output stages, there are no white coils left. L51 and its brethren were bad when I got the set, so those have been replaced. L54 and its brethren are not white coils, but red ceramic coated ones that seem to be fine (maybe were previously replaced?). A few white coils still left in some other circuits.
I also swapped the 6BC7 DC restoration tube for a brand new one, but that didn't have any positive effect.
For completeness, here are some voltage measurements with the set running well at its optimal contrast point:
Screen voltages:
R: 448 V
G: 275 V
B: 306 V
Cathodes: all at 124 V
Grid voltages (varying with scene, of course, so don't pay too much attention other than the general range of these):
R: 87 V
G: 88 V
B: 95 V