I've been blessed!

[old_coot88]

Maybe tack four Si diodes, each with proper series resistance, across the 5U4s(?). Just a thought for a quick&dirty test. That'd also stiffen the B+ against sag.

[EDIT.] Interesting vid here on a CTC-5 resto. https://www.youtube.com/watch?v=3GPrCym5qXw

[Penthode]

Originally Posted by old_tv_nut

Update: No joy yet.
I started chasing the brightness drift this afternoon The suspect resistors measure OK cold, and after the brightness has drifted, freeze spray shows no effect on any components on the video board.

Cathode voltages drift by 10-15 volts, and adjusting the brightness control for the same variation shows the same brightness variation as the drift.

I noted that the far end of the brightness control goes to a negative grid bias voltage developed by the horizontal output tube, but changing the HO tube and/or the H driver made no difference.

I backed up and measured the 385 V supply, and it is low (as low as 330) and varies with beam current. If I understand the HV regulator properly, it should keep a constant load on the HV with brightness changes and therefore the 385 supply should not vary drastically. So, the regulator may be wonky or just acting wonky because the B+ is low.

I think my next move is to replace the 5U4's in the power supply, but I don't have any at the moment.

If the variance of the +385v supply is that great, something is wrong. On my CTC5, upon inspection before I even powered it up, the ground to the transformer center tap was not properly soldered from new! I never had the opportunity to find if it affected the B supply! But a drop to 330 volts can mean a few things eg bad 5U4 tubes, bad fuse holder of a leaky electrolytic. But if it varies with beam current that would suggest the series power supply resistance is too high: bad weak 5U4 tubes seem the most likely candidate.

The voltage sag for a 5U4 at 250mA is about 50volts per plate. The seiers resistor would be 50/0.25 = 200. Two plates in parallel son 1 100ohm 50W resistor per would work.

To test I would simply replace the two 5U4 tubes with a pair of of 2A 1500 PIV diodes.

[old_tv_nut]

Originally Posted by Penthode

If the variance of the +385v supply is that great, something is wrong. On my CTC5, upon inspection before I even powered it up, the ground to the transformer center tap was not properly soldered from new! I never had the opportunity to find if it affected the B supply! But a drop to 330 volts can mean a few things eg bad 5U4 tubes, bad fuse holder of a leaky electrolytic. But if it varies with beam current that would suggest the series power supply resistance is too high: bad weak 5U4 tubes seem the most likely candidate.

That's what I'm thinking - both the time that it takes to drift and the fact that it always comes back to the same starting condition when cold points to tubes.
I ordered a couple, should be here in a few days.

By the way, dragging it from the corner revealed that the convergence plug was not making good contact, and exercising it a bit seems to have restored it.

[old_tv_nut]

I see Tube Depot sells solid state replacements, but warns they will increase B+ by 10 to 20%, so they must not include resistors. I'm doing baby steps by replacing tubes first, then may consider solid state sub.

[old_tv_nut]

I sat down with the service info and reread the grayscale tracking procedure, partly in case it might be needed when the B+ is restored normal, but mainly to try to understand how it works.

The strange thing is that the CRT grid DC bias points in the chroma section are adjusted to correct highlight color, even though the controls are labeled "background" and really do affect the background more strongly than the highlights.

The instructions say to set the "background" controls to counter any tint in the highlights and then adjust the G2 controls to counter the resulting tint in the shadows.

After reading this carefully and looking at the electrical characteristics of the CRT, I realized that the CRT gun is actually a variable-mu (variable gain) tetrode, and when the G2 controls are adjusted after adjusting the grid bias, the CRT is operating at a different mu (gain), effectively changing highlight white balance.

I also note that the CTC-5 called for a 8200 Kelvin color temperature, not as blue as the 10,000 Kelvin that was spec'd for later chassis. In any case, I don't believe accurate color temperature was ever achieved by eyeball, as the eye adapts too readily to color temperature changes. No TV repairman had a colorimeter in that era.

[Penthode]

I try to clolor balance the black and white image against a black and white television of the era. If you had a colorimeter you could do it more accurately but in those days, especially with color drift as the set warms, I find eyeballing it against a P4 phosphor television is the best you can do.

[old_tv_nut]

Originally Posted by Penthode

I try to clolor balance the black and white image against a black and white television of the era. If you had a colorimeter you could do it more accurately but in those days, especially with color drift as the set warms, I find eyeballing it against a P4 phosphor television is the best you can do.

I do have a colorimeter, but I usually balance by eye against natural light coming in the window. I haven't tried measuring the CTC-5 for a long time, but using my DSLR as a rough colorimeter, it measures somewhere around 7700 to 8000 Kelvin as set by eye. I think many/most P4 tubes from the early/mid 50s onward were far too blue as a color set reference. I seem to recall GE boasting how blue their P4 tubes were.

[old_tv_nut]

By the way, since I got some adjustable color temperature LED light panels, there's much less post processing needed to match the screen and room light in photos.

[Penthode]

Originally Posted by old_tv_nut

I do have a colorimeter, but I usually balance by eye against natural light coming in the window. I haven't tried measuring the CTC-5 for a long time, but using my DSLR as a rough colorimeter, it measures somewhere around 7700 to 8000 Kelvin as set by eye. I think many/most P4 tubes from the early/mid 50s onward were far too blue as a color set reference. I seem to recall GE boasting how blue their P4 tubes were.

I was using a late 1940's 10BP4 as reference. It is has a lower color temperature as it is much less blue.I agree as I have seen how mid to late 50s monochrome CRTs do have a higher color temperature.

[old_tv_nut]

Originally Posted by Penthode

I was using a late 1940's 10BP4 as reference. It is has a lower color temperature as it is much less blue.I agree as I have seen how mid to late 50s monochrome CRTs do have a higher color temperature.

That's great. I have definitely seen the difference in the older sets at the Early Television museum.

I have an I1 Pro3 spectrophotomter I use for calibrating my computer monitors, as the less expensive colorimeters I had in the past could not do a good job of matching different models of monitor.

This discussion prompted me to get out the Minolta spot colorimeter I got when Zenith closed down their lab in 2000. Its last calibration was in 1998. It is way off when measuring the two LCD computer monitors (Viewsonic with sRGB gamut and HP with wide gamut). These monitors match visually when calibrated to 5000 Kelvin with the I1, but measure 4000 to 4400 with the Minolta. So, I guess the Minolta suffers the same problems as other colorimeters even though it was an expensive top-of-the-line unit at the time.

I know the home-built colorimeters that Zenith used on the assembly line had to be recalibrated any time the phosphor formulation was changed. This was a long chain of calibrating a spectrophotometer (which used a photomultiplier tube at the time) using an NBS traceable incandescent lamp and then measuring a CRT with the new phosphors and calibrating the production line colorimeters to match those readings. I never talked to whoever was using the Minolta, so don't know if it required the same calibration procedure, but I suspect it did.

[old_tv_nut]

The standard incandescent bulbs I mentioned were calibrated with respect to standard bulbs maintained by the NBS, and originally traced back to the definitive standard consisting of a black-body cavity in a piece of platinum at a specified temperature

Today, of course, instruments use stable solid state silicon photodetectors. I believe (but don't quote me, have to research it) the base standard now consists of lasers whose output can be measured accurately thermally and then transferred to a standard photodiode and from there to secondary standards.

[etype2]

Here is an excellent PowerPoint Level 2 presentation by ISF. (Imagining Science Foundation) A certified calibrator will have completed Level 3 training.

https://visions4netjournal.com/wp-co...-2017-33.3.pdf

[old_tv_nut]

I got the new 5U4GBs from Tube Depot (new from JJ Elctronics, made in the Slovak Republic). They may have increased th B+ a few volts (hard to tell because of the loading effects of beam current) but have not fixed the brightness drift. Cathode voltage still drifts up about 10 volts during this long warmup.

Brightness still starts high and drifts lower slowly, and returns to too high after the set is off a while. Still wondering what can have this characteristic. Still a possibility of a resistor under the chassis where I can't get to it with the freeze spray?

Don't know - more head scratching is in order.

I tried running with the line voltage 5 volts higher to bring up the B+ some, but this doesn't seem to make a big difference in the drift.

I'm not running totally out of brightness control range, but have to set it fairly high after the set is warm. Running through a slide show, I just see the usual effects of less than 100% DC coupling.

[bgadow]

Had you swapped the HV regulator? (I can't remember, was it a 6BK4 that early?) Maybe a bit gassy? Stabs in the dark.

[old_tv_nut]

Originally Posted by bgadow

Had you swapped the HV regulator? (I can't remember, was it a 6BK4 that early?) Maybe a bit gassy? Stabs in the dark.

I'll try that tomorrow, but as you say it feels like a stab in the dark. I have already tried the horizontal oscillator and output, which seemed more likely because one end of the brightness control gets a negative voltage from the grid of the horizontal oscillator.

Unfortunately the horizontal circuit board is hidden behind the high voltage cage and practically impossible to get at, but I am going to measure voltages at the brightness pot tomorrow - slapping my forehead for not thinking of it before.