21-CT-55 power supply failure

[Penthode]

Originally Posted by miniman82

Since it's not a vertical blanking issue, how exactly does this chassis blank horizontally? I don't see a blanking tube.

There is not blanking circuit as such. Blanking is done by the video waveform. Hence if the if there is a bad VIF response or misadjusted fine tuning, high video frequencies will overshoot producing a white line.

Try adjusting the fine tuning and reduce the RGB CRT screen adjustments.

[miniman82]

Here's a scope trace, maybe you can get something from it.



Bottom is the composite output from the generator, top is from the grid of the second video amp tube. It does look like there's some fuzz on the blanking pulse, like the whole signal slid over a little. If so, are we talking sync issues or IF issues?

[Penthode]

Originally Posted by miniman82

Here's a scope trace, maybe you can get something from it.



Bottom is the composite output from the generator, top is from the grid of the second video amp tube. It does look like there's some fuzz on the blanking pulse, like the whole signal slid over a little. If so, are we talking sync issues or IF issues?

Maybe IF and maybe sync. Maybe both. What I am trying to say is that if the video frequency response is not flat or the phase response linear, you can get an overshoot at the sync pulse rise and fall. Getting a good wide bandwith with the tube technology of the past was a bit of a challenge. The problem is exaserbated with the use of vestigial sideband trasmission: broadcast transmitter design went to great lengths to try to mitigate the problem.

Also the broadcast sync waveform had a carefully controlled rise and fall time to help reduce this effect. Your test generator might have too fast a rise and fall time which may cause problems. Your modulator, like most consumer modulators, is double sideband so that should not cause a problem.

If your IF response is not flat or if the video carrier is too low on the Nyquist slope of the VIF response, this could lead to the white line appearing. I have attached a jpg photo I found on the web to illustrate what I had expected the sync to look like as applied to the CRT. Your waveform seems to have very fast rise/fall time sync pulses. Try expanding the waveform around the sync fall as an overshoot there may be seen in the video as the white vertical line.

Are you injecting at baseband or RF input? If at RF, if you adjust the fine tuning control to raise the video carrier to the top of the Nyquist VIF slope (that is adjust until you begin to lose picture detail and color and picture starts to smear) see if the line diminishes.

Lastly, consumer DTV boxes have proper broadcast sync rise and fall times so see if another video source causes problems.

[miniman82]

It's getting worse, now an RF signal with a staircase won't even get through. All I get on screen is a jumbled mess, horizontal tearing, ect. I think it's high time to pull this chassis for a good while and put the B&K 415 on it, real detective work is in order.

[CT-100]

I'm on mobile and cannot easily follow all responses. With regard to your hum, did you bypass your Si diodes with .001's? That looks like switching noise to me.

[miniman82]

I decided to bypass the IF for the time being, injecting video into the 1st vid amp and disconnecting power to the IF strip. Found something odd: the bandpass amp grid is registering 0 volts, when it should be -15 or so. I can't figure out why it's not biased right, all the resistors in the circuit measure correctly and the caps are new. With the grid at 0, that tube would draw more current that it's supposed to which might explain the power supply drain. It might also explain the dot crawl I'm getting now.

[miniman82]

For continuity, this problem was finally fixed here: http://videokarma.org/showthread.php?t=252849


The voltages on the bandpass tube ended up being a Sams typo, the CT-100 chassis had the same readings. After fixing the problems detailed in the thread above, dot crawl is much better as well. I can probably make it disappear by tweaking the 3.58 trap in the luma line. So glad this is finally over.