CT100 owner request

[dtvmcdonald]

I have a lingering problem with my CT-100. I'm not sure whether this
is just my set, or a design problem.

The problem is that if the far left edge of the image, immediately
adjacent to the end of the horizontal blanking, is a near saturated
color, not green or magenta, it leaks into the tail end of
the "selected" color burst and causes the color sync to go wrong.
This appears as horizontal bars of wrong color. The size of
the problem varies with the position of the horizontal hold. At the
full CW setting it is weak or absent and at the full CCW position
where sync is maintained it is worst.

So I ask ... can anybody with a CT-100 check for this. You need
a test signal with a full intensity red, orange, yellow, or blue-green
color. It could be a full field or horizontal color bars. Such test
signals exist on most standard calibration DVDs. Just run the
pattern and see if it is the right color and if it changes as the
horizontal hold is run over the full in-sync range.

I can see the impingement on a scope looking at the burst
at the AFC diodes.

Thanks in advance for helping.

[miniman82]

Have you ran through the horizontal oscillator and color AFC adjustments outlined in the service instructions? That clears up 99% of the problems, but I suppose the CW transformer being out of whack could cause it as well if it was way off. How do the rest of the bars look?

A pic of what's happening usually narrows the possible faults to only a few, but I would do the horizontal and AFC stuff first. And make sure you don't have any peaking coils going bad! Can't tell you how long I chased my tail because of bad peaking coils in that chassis.

[dtvmcdonald]

The horizontal and color burst adjustments were done exactly as specified
and the test results were perfect. There is a test specified for this exact thing
in the RCA service man, i.e. Fig. 34 and note in left column of p. 23. Its not
supposed to change as you change the horizontal hold. And it doesn't ... if you use
standard VERTICALLY positioned color bars, i.e. a bar running from top to bottom of the screen.
That's because the standard ones don't have a 100% color intensity along
a lot of the left edge of the picture. True, the real SMPTE ones do have
a bit of blue there, but its so few lines that
it doesn't really show up well looking on your scope for fig. 34. I can barely
see it, expecting it.

I don't have an image, its not needed. Look at a pure 100% intensity
100% saturated 100 screen coverage red field from any test disc. it should be red!
And the top is. But after 1/2 to one inch from the top it turns magenta.
The distance depend on the position of the horizontal hold.

[miniman82]

So, you're saying the color smears from red to something else from top to bottom depending on the setting of horizontal hold?

Are you running the signal through the antenna or composite input? I suppose it's possible for the tuner or IF to be slightly out of whack one direction and the color circuits to be skewed in the opposite direction causing it to get confused (some sort of odd phase shift or somesuch), but it should be noticeable in normal viewing if it were that far off.

Try this: whenever I troubleshoot a color problem, I always inject a signal at the first video tube's grid (NTSC signal). That eliminates the tuner/IF as a possible source of confusion. If it's still doing the same thing, you know the issue lies in the chroma/demodulation section and you should focus your efforts there. My guess will be that the CW osc transformer (Sams A36 and A37 respectively) is tilted too far one way, and because the feedback signal for the AFC is taken from one side of that transformer, it's easy to see how even with correct colors on the screen a slight imbalance will throw the whole thing off. In other words, I've found that it's possible to have that particular transformer set with correct colors on screen, but have an overall phase shift with respect to burst that it tends to get the AFC circuit all bent out of shape.

It's the color equivalent of the horizontal circuit being on the verge of falling out of sync, so you're seeing the color 'pulling' just as the whole screen would tend to tear when on the edge of the horizontal sync pulse.

I think this is because the color burst is gated by the horizontal flyback pulse, and if those 2 transformers are set incorrectly it will try like hell to lock in but it's too far off for an automatic lock. What happens to a set of color bars when you ground point 'M' in the Sams? If you get a rolling color rainbow, your AFC is out of line with respect to burst. If that's fine, my money is on A36/37 being tuned incorrectly.

Best way I've found to get those 2 close to good is have color bars on screen, and tweak them both till you zero in on a normal looking pattern. You can look at the grid signal from the CRT with a scope to try and get the ladder to look right, but I always just tune it by eye.

Hope this helps, but I really didn't think it would be possible to have vertical color tearing. Sounds like a strange problem indeed.

[dtvmcdonald]

Its NOT out of sync. When its wrong its in sync, quite tightly.


The problem clearly is that it is in sync to the AVERAGE of the
color burst and the part of color in the left hand edge of
the image. You can see this easily on the scope when
doing the Fig. 34 test. Different color give different effects.
If the field color is in phase with the burst, it gets longer.
IF out of phase, it gets shorter. In these cases the color on the screen stays OK.
If its 90 degrees out, it gets longer and you can see it change
phase during itself, on the scope.

In other words, the gating signal is too long or too late.

The effect is RARE on ordinary program material. It is
sometimes seen on football games where there is lots of green
at the left edge, and always on commercials with solid saturated color backgrounds.

I suppose I should check the tuning of 2L141 and 2T122 again.
What I'm really looking for is confirmation that my set is off,
that its not a design problem. It could be a problem with a
capacitor or resistor in that area, making the gate be too long
or too late.

[Pete Deksnis]

Originally Posted by dtvmcdonald

What I'm really looking for is confirmation that my set is off,
that its not a design problem.

Last May at the ETF convention I gave a PowerPoint presentation about the final tweaking that took place before the CTC2 was released for initial production. I had a sidebar that had a screen shot of an anomaly that is similar to that in your post, and I asked the same question you have, but at that time left it hanging without an answer.

Since then I'd troubleshot the problem, and to get right to the point, yes, I think there was a design problem. Highly saturated chroma that appears a few microseconds in from the beginning of horizontal sweep does not cause the anomaly, because it occurs far out of range of the keying pulse trailing edge.

Highly saturated chroma, which appears right at the beginning of the horizontal sweep, occurs close enough to the keying pulse trailing edge to literally climb up the edge of the pulse, slide onto the top of the keying pulse, and infect the color burst with chroma. This anomaly can occur because rotating the horizontal control varies the phase of the horizontal oscillator enough to shift the timing of the keying pulse -- and drive it into the garbage chroma signal.

Why garbage? Because the color burst is tapped from the plate of the 6CL6 first video amplifier by a tiny (3 pF) cap that can't help conducting some of the chroma signal as well. (Of course, if there is no chroma there is no chroma signal, which is why, if there is white or black at the beginning of the horizontal sweep, there is no chroma to contaminate the color burst at the top of the keying pulse, and the anomaly under discussion does not and can not take place.) Only highly saturated chroma signals cause the anomaly because only large signals will be great enough to get through the 3-pF coupling cap in sufficient amplitude to cause color burst contamination.

Another quick point: I believe Philco, when designing its first color TV, the Model 22D5102 with the TV 123 chassis, ran into the same problem as RCA with its CT-100. Philco's fix was to add a pot to the horizontal oscillator circuit that was adjusted to limit the range of the horizontal hold control.

I have been fleshing out that ETF PPT video; when finished I will try to make it available if there's interest.

Of course, there's still the chance that it's an alignment problem... but I'm beginning to believe not.

Pete

[miniman82]

I've never had this problem, so I'm struggling to understand what's happening. Perhaps some careful tweaking with the 3.58mhz trap would help? All you really need to do is sweep it, and make sure it's peaked correctly.

[Pete Deksnis]

Originally Posted by miniman82

I've never had this problem, so I'm struggling to understand what's happening. Perhaps some careful tweaking with the 3.58mhz trap would help? All you really need to do is sweep it, and make sure it's peaked correctly.

It's hard to see; a 100-percent saturated solid color (I use red) test signal is needed that begins at the far left of the screen. Similar hue distortions have been witnessed occasionally with 'regular' video programming, where rotating the horizontal hold control adds and removes hue distortion. For example, some years ago when the Sarnoff Library was still in existence and had a working CT-100, I and others observed the same effect. Can't yet be sure if the same process was at work there as is here.

I assume you are referring to the 3.58-MHz trap in the plate of the first video amp. The burst is taken from the plate (via the 3-pF cap) before that trap, and so the trap, if misadjusted, I don't believe would reduce the amount of chroma contamination in the burst processing circuitry. It might leak some chroma into the Y channel of the matrix.

Pete

[dtvmcdonald]

I'm now convinced its a design problem. I believe that a possible solution is
to reduce the values of the series resistor (22k) and capacitor (150 pF) coupling
the horizontal output transformer to the grid of the burst keyer tube,
perhaps to 4.7k and 50 or 75 pF. A simulator says this should reduce the
problem, and the 21CT55 does this, though only to 10k and 100pF.

I'm not going to change my set as of yet.

[Sandy G]

Call up RCA & DEMAND they come & fix it !