Those screen caps were not very bright when I took those a couple of years ago. Always have needed to run this set quite dark to get enough red.
Today, I took a quick shot during the Super Bowl. This is the west coast, so there was still significant light coming into the room from the window at 5 PM or when I took the shot. As you can see, red is decent. I'll take some nicer screen shots after I've had a chance to fine tune things better.
Noting the concerns about positive grid voltage and grid current mentioned above, I made an update to the circuit by adding two 1N4148 diodes in series between the red grid and cathode. I used 1N4148 to have fairly high speed diodes. Two in series since they are only good for 75 V reverse voltage each. This limits the grid to only +1.2 V on positive peaks of the video.
Here's the circuit:
With this circuit, the DC restorer sets the black level ahead of the cathode follower, and the diodes clip the peaks off the video drive to the CRT grid. I'm pleasantly surprised that clipping the positive peaks still lets plenty of red drive through (so I don't really need the strongly positive peaks on the grid signals).
I compared this to the original to reconfirm that this was all doing some good. The difference is very compelling. Without the cathode follower, the DC restorer isn't able to keep the black level fixed on the red gun any time the red signal becomes large. As reported above and on the last page, grid current competes with the action of the DC restorer, and the DC restorer loses. Now with the cathode follower and diode clipper, the DC restorer keeps the black level fixed as it should, and I can get decent red.
Not like a new tube, but pretty watchable. With the clipper in there, hopefully I am not damaging the red gun.