Just one additional thought:
The 12LP4 datasheet for your CRT shows that the Heater Negative with respect to cathode shouldn't exceed 140V.

Without the sams, its not immediately obvious what the plate supply voltage is on the video amp. But, if you implement this, make sure the average heater to cathode voltage doesn't exceed the limit.

dobulee, for the love of god don't take this conversation offline please! I'm working my way through a couple of TV engineering books right now and this discussion was super, super helpful to me. Thanks!

Most of the DC restore circuits are more complicated than this, and a lot of them use the same triode for synch separation and DC restore. It makes some sense, since the triode grid gives diode action when driven positive.

Im interested if the above circuit functions decently or not. Im sure the original set designers (and others here!) are aware of more practical limitations and tradeoffs than I am.

screen shots and/or scope captures would make for further discussion! that little simulation doesnt model the video pentode output impedance or anything...so...mileage may vary!

Heater negative is just ground, I believe. So basically cathode to ground, then? Voltage chart says video amp plate voltage is 120vdc, measured to pin 1 of V11, I'm guessing a positive... rail... or whatever. Main voltage supply type thingy. The crt cathode says 115vdc, presumably to ground.

I have ordered some diodes. Will keep you posted.

Alright, so I assembled the circuit as described, and it causes some pretty wild image instability :/

Some tearing near the top always, with image/scene changes the horizontal and vertical stabilities change, sometimes the image jumps halfway down the screen, other times its starts to roll. Contrast control used to affect image stability, but now contrast, brightness, and focus have drastic effects on it. And at all times, no matter what, the image is never perfect. At best it's skewed and jittery, and ripple-y.

That all having been said, I *think* it helps the DC restoration, but it's kind of hard to tell when the image is getting as screwed up as it does.

So I'm thinking. This shouldn't affect anything but the brightness and contrast, yeah? Is there some way of isolating it (more) from the rest of the workings? Or wait. The path to the cathode used to be a capacitor and resistor, so that would mean it's capacitively coupled, but with a limited straight connection. Now it's just capacitive only. Would a resistor over the first cap help?

EDIT: I also tried a 1k resistor for R5, it barely made a difference.

if you put a resistor over the cap, you wont have DC restoration anymore :D the resistive path to common is now through the 2 series 470k resistors.

unfortunately I do not have a cathode-video set to experiment with, however, i suspect the tearing problems are due to the circuit interfering with the synch separation, which also comes off the video amp in your set.

anyhow, thats the basic idea:
a diode to clamp the black level
capacitively isolate the CRT grid from the DC level of the video amp
set the DC cathode level with a resistor divider

the details are the complicated part!

go even to 10k at r5. without a similar set, I cant do much other than consult :/ do you by chance have a scope to post some waveforms? as a last resort, i can try to simulate your entire video amp circuit, but i wouldnt have a lot of faith in the tube models up in the Mhz range, if the models even exist.

for what its worth, the 48 RCA 6t45 chassis has dc restoration, and even it doesnt work thr greatest. I took some captures of thr cathode waveforms with all black and all white screens -- i will post them over in the 6t75 thread. But the front porch "black" level on the CRT grid drifts by about 10 volts from "all black" to "all white", causing all black to look gray. That set uses a positively biased triode grid to get diode action from grid current - and it probably doesnt make for the "tightest" clamping.

if anybody else has ideas, id love to discuss. seems like an interesting little project.

Alright I experimented a bit today. First thing I tried was to change R5 to 10k. It made a big improvement in stability. yay! In fact, about 99% of the problems could be adjusted out, and did not come back with changing images. Now, I made a set of test images and use them in a slideshow for testing. I have a regular test pattern, all-white, all-black, then 4 sets of half black on top, bottom, right and left. They don't really approximate rapid scene and lighting changes, but they work well enough for basics.

First thing I noticed was that on the all-black and top-half-is-black* images, the black level starts out at about dark gray, then fades to medium gray over the course of 1-2 seconds. However this happens in reverse when going from all-black to side-half-is-black, where the black area starts (darker than it is with all-black) dark gray and darkens down to a good black. Which, I'm guessing, is what you intended to happen. Why it happens in reverse on all black is beyond me.

*top-half-is-black improved with later modifications.

Then I tried lowering C3. This brought back some of the stability problems and generally made it worse. I then raised it to .2uf and it made things better. Better stability and better evenness of black areas, but not really better black levels. All-black was still the same medium gray. It also seemed to help a bit with retrace lines. I added .1uf more to it, which improved it only a tiny bit more, it seems it's reached the point of diminishing returns at .3uf total.

I then tried doubling C5, which didn't do much besides raise the brightness momentarily while it charged up.

I played with the value of R6, but that didn't seem to do much besides alter the overall brightness.

Also, I just noticed when I turn the set off a spot remains on the screen, which it never did before. But only for maybe 5 seconds, and it's about the size of a quarter, so I don't think it'll hurt anything.

So then I put it back to factory spec, and examined the black levels. Interestingly with the test images I have, they all exhibit perfect black levels until you get to all-black, which is so light gray that it's nearly white. So at least the all-black has some improvement with the new circuit. However when watching actual widescreen video with black bars, it is noticeably worse than it is with the new circuit, that is to say - the changes in black level are more prominent. And also, because of that fade in time with the new circuit, it sort of takes the edge off of rapid light-to-dark and dark-to-light scene changes, by allowing it to slowly fade down or up, so it's not as jarring. So that's good. Which capacitor manages that, though? I changed the values of both of them and it didn't seem to make much of a difference in the fade in time.

Also I measured the crt cathode voltage to ground:
Sam's says: 115
Factory spec: 150
Current circuit: 145
idk, make of that what you will. However I do not have a VTVM, I'm just using a cheap digital multimeter.

Below is attached a schematic of how I have it currently. I understand that it'll never be perfect, but I feel like it could still be a bit better. If you have any more suggestions, I'm listening. And thanks again for all the help!

Oh also, if I should hook up a scope, would I connect to the cathode? And what would I be looking for?

After reading the Rider's TV Manufacturers' cure for retrace suppression, I was kind of disappointed because the resistor numbers do not match the numbers on the Sams schematic. Frustrated, I had put the issue aside. Recently I decided to give it a go again, and through very careful reading and educated guesses, figured out where the connection points were.
For the 1st, the resistor and capacitor numbers didn't even matter, as pin 2 of the crt only goes to one junction, so that was it. For the 2nd, the major giveaway was the 3.9k ohm resistor. There are only two in the entire set, and one is like in the tuner or somewhere completely unrelated. For the 3rd, it calls for the junction of C408 and the already-identified R447 and it's a resistor to ground, so it could only be the opposite side.

I'm pleased to say this retrace cure works like a charm. It's perfect, there are NO retrace lines, period. Whereas before it was installed, glaringly obvious retrace lines were onscreen about 90% of the time. I used a .047uf capacitor. I had a .056 ready to try, and even if I put them both in parallel it made no difference. So it's not very picky about that capacitance. Also the old 3.9k resistor tested at 4.5k... makes me wonder how many other resistors have drifted. >_>

I then finalized the DC restoration. I made a tidy little prototyping pcb. Sacrilege, I know. I think it's a much neater solution than perhaps sticking a lug strip somewhere. Plus the fact that it's obviously modern will be less likely to confuse anyone servicing this unit in the future. The schematic posted in the previous post is the current iteration - big thanks to Dobulee for that!

As stated before, it's not perfect dc restoration, but it is an improvement. And it does this thing where a dark scene will start dark and fade to brighter (in about 1 sec). But at least that takes the shock away from the transition. Besides, even without that, it does actually make an improvement in keeping dark levels. However slight an improvement, it was pretty bad without it, so I'll take it.

Next step is putting the chassis back in the case. I've already fixed a couple broken bosses in the case - for the screws that hold the back cover on. And retapped the screw holes on the chassis.

Oh yeah, how would I direct inject video into this? I tried putting raw video right into test point B (I think) which is right before the video amp. It came through but was very faded and gray. Maybe it needs a transistor to amplify the signal? The test point being right before the video amp makes me think that's what it's for.

Test points were usually more for measurement than injection back then and even most service/ manufacturing test injection points were not designed with monitor use in mind (composite video as a standard only existed in broadcast studios which didn't buy consumer model monitors)...You want to inject as close to the video detector output as you can while still getting the correct video polarity.

I had this set out on Christmas day running a fireplace and playing xmas music. It performed well, only had to adjust it once a couple hours in, but I was running the contrast right near the point where it's about to make the image unstable, anyhow. I think it's just fine. And the sound is really crisp and clear, even with a couple of weak tubes in the audio section. Something I noticed though, is that it no longer has a spot when you turn it off - the screen just goes black instantly. Which is what it does from the factory, but after adding Doublee's DC restoration mod, it would leave a spot the size of a quarter when turned off, for about a minute.

It just now dawned on me that the retrace suppression circuit was the only change made between then and now. So I guess it kills the spot too. Just thought I'd throw that out there.