RCA 21CT55 #8802897 is Dead

[Kirk]

Hey Tomcomm,

What did you ever do about the Flyback in the 21CT55?

Thanks,
Kirk

[Electronic M]

Originally Posted by Kirk

Hey Tomcomm,

What did you ever do about the Flyback in the 21CT55?

Thanks,
Kirk

IIRC he cobbled a CTC-20 flyback in and was running that.

His hotrodding antics have slightly rubbed off on me...A while ago I acquired the quadrature transformer from a CT-100 that I want to try swapping into my 21CT55....the CT-100 was an I,Q chroma demodulator set and the CT55 was a R-Y,Q demodulator set... Supposedly the only difference was the quadrature transformer and one resistor in the matrixing... I'd kinda like to see IQ on a 21" screen and it'll be a low risk, reversible sleeper mod.

[dtvmcdonald]

Tom: The "one resistor" means one less. A couple of values in the matrix are different. I doubt you will be able to see a difference. Only the relative noise in the colors and the (small) spurious color line caused by the different demod axes
will be different.

The difference you will see will be determined almost
entirely by differences between the types of peaking coils used
to replace the nasty white ones that self-corrode. That makes huge differences.
I spent hours and hours getting the right coils to get the sharpest and least ringy color bars. I tried a couple of the ones specified as replacements in the manual, and they were disastrously wrong.

[Penthode]

Originally Posted by dtvmcdonald

Tom: The "one resistor" means one less. A couple of values in the matrix are different. I doubt you will be able to see a difference. Only the relative noise in the colors and the (small) spurious color line caused by the different demod axes
will be different.

The difference you will see will be determined almost
entirely by differences between the types of peaking coils used
to replace the nasty white ones that self-corrode. That makes huge differences.
I spent hours and hours getting the right coils to get the sharpest and least ringy color bars. I tried a couple of the ones specified as replacements in the manual, and they were disastrously wrong.

In my CT100, I rewound almost all of the peaking coils. It was surprisingly easy. And the frequency responses the exercise yielded resulted in an excellent ringing free picture.

[Electronic M]

Originally Posted by dtvmcdonald

Tom: The "one resistor" means one less. A couple of values in the matrix are different. I doubt you will be able to see a difference. Only the relative noise in the colors and the (small) spurious color line caused by the different demod axes
will be different.

The difference you will see will be determined almost
entirely by differences between the types of peaking coils used
to replace the nasty white ones that self-corrode. That makes huge differences.
I spent hours and hours getting the right coils to get the sharpest and least ringy color bars. I tried a couple of the ones specified as replacements in the manual, and they were disastrously wrong.

I just use the peaking coils off digikey that Nick Miniman82 recommended many years ago. Such as this (this is one of several from that list): https://www.digikey.com/en/products/...-223-rc/969835

I've heard those coils criticized by more than one person, but the people doing the criticism don't site modern off the shelf parts as being part of their better solution...I don't have time to hunt out of production borderline unobtainium, I also don't have the equipment or skills to rewind much less wind new multi-layer coils or properly characterize them or verify performance outside of eyeballing the screen (I also don't have money or time to pick that up)...So until someone comes up with a better list of current production off the shelf parts I can buy off mouser or digikey I'm going to stick with what I know works decently in both of my sets and very soon a set I'm restoring for a customer.

[dtvmcdonald]

Originally Posted by Electronic M

I just use the peaking coils off digikey that Nick Miniman82 recommended many years ago. Such as this (this is one of several from that list): https://www.digikey.com/en/products/...-223-rc/969835
.

Those or the API Delavan (Spelling?) look-alikes were what I ended up with
in all cases. Only one, the 6.8 mH one in the Q, needed a value much different than the spec. I also found them to be the best ones in my HF radio 7 to 11 pole bandpass filters.

[Electronic M]

Originally Posted by dtvmcdonald

Those or the API Delavan (Spelling?) look-alikes were what I ended up with
in all cases. Only one, the 6.8 mH one in the Q, needed a value much different than the spec. I also found them to be the best ones in my HF radio 7 to 11 pole bandpass filters.

Interesting. Is there any significant difference between the API made ones that makes them better in some circuits or is it the same either way? Do you remember what value or part you used for the 6.8mH Q coil?

[Penthode]

Originally Posted by Electronic M

I just use the peaking coils off digikey that Nick Miniman82 recommended many years ago. Such as this (this is one of several from that list): https://www.digikey.com/en/products/...-223-rc/969835

I've heard those coils criticized by more than one person, but the people doing the criticism don't site modern off the shelf parts as being part of their better solution...I don't have time to hunt out of production borderline unobtainium, I also don't have the equipment or skills to rewind much less wind new multi-layer coils or properly characterize them or verify performance outside of eyeballing the screen (I also don't have money or time to pick that up)...So until someone comes up with a better list of current production off the shelf parts I can buy off mouser or digikey I'm going to stick with what I know works decently in both of my sets and very soon a set I'm restoring for a customer.

Just to let you know how I wound my peaking coils, I used a 1megohm 1watt resistor to wind the coil on. Cut out with scissors two cardboard disc from old business cards and slipped them over the resistor to hold the windings in place. Roughly calculated the number of turns to achieve needed inductance from an on line calculator (I Googled). Bought off Amazon appropriate gauge wire small spool. Made a spool holder from a coat hanger wire C clamped to a table. Fit the end of the resistor I was using to wind the coil in my variable speed power drill. Slowly (5 turns a second) ran the drill. (There was typically around 600 turns for a 1 mH peaking coils so winding that did not take too long).

When wound, I would apply the nail lacquer I bought at Walmart to hold the windings in place. I would then take the coil and measure the inductance using a series resistor, signal generator and oscilloscope to determine the 0.707 point when the inductive reactance equals the resistance. I would tend to be a bit generous with overwinding so that I could remove turns to hit my target inductance.

I did this for almost all of the 8 or so open peaking coils in my CT100. The coils in the I and Q chroma path were fairly critical in order to get a nice flat 1MHz response for the I channel and 500kHz for the Q channel. I frequency swept the luminance and chrominance channels and they matched closely the RCA published curves. And the net result was a an extremely nice sharp picture, free from ringing.

Looking at my bunch wound coils, it would be difficult to distinguish from a new commercial replacement. And my coils looked a heck of a lot better than the original ugly white blobs.

[Kirk]

I have used the Digi Key replacements for the failed white coils- the only difference is that I put them in Series to add up to the desired value using a L-C-R meter. I believe it is the difference in DC resistance that causes the ringing- not the coil itself. By increasing the DC resistance it minimized the undesirable ringing (you can see it on the scope) to be indistinguishable.

Great job winding your own!

Kirk

[Electronic M]

Originally Posted by Kirk

I have used the Digi Key replacements for the failed white coils- the only difference is that I put them in Series to add up to the desired value using a L-C-R meter. I believe it is the difference in DC resistance that causes the ringing- not the coil itself. By increasing the DC resistance it minimized the undesirable ringing (you can see it on the scope) to be indistinguishable.

Great job winding your own!

Kirk

Did you add series resistors or get a large assortment of coils and test them in series with each other until DC resistance and inductance matched target values?

[dtvmcdonald]

I think I used 3.9 mH for 6.8. I did buy an assortment ane tested each one for inductance using one of those yellow meters anf for self-resonance using a frequency synthesizer, a 1 megohm scope probe, feeding the signal in through a 1 pF capacitor. I used my 5-spice simulator in the actual circuit and tried the best two
in the actual set, sweeping them (with the horizontal system fully disabled as the yoke was not connected.) I used the best looking sweeps. All looked fine when tried on actual pictures except the Q one. That one I tried by mounting the chassis behind the set and trying coils and different values oif the nominally 27 pF cap from the Q demod to ground. It a compormise between the best looking Q channel response and the problem that the best response looks terrible because its much too early ... it would need less delay in the Y delay line and the I filter, which has intentional delay in it . I have actual photos of the restult on +-I and +-Q bars posted on an earlier thread here. I will repost here. Note that the time error is obvious here ... but not on program material. Note that there is an orange bar from crosstalk on green to violet but the corresponding bar on green to violet looks more like luminance. Except for the black and white bars, Y is constant. Note that th I channel looks really good.

[Kirk]

The Digi Key coils tended to have a lower DC resistance than the originals- Putting them in series until I had the correct inductance value worked for me.

[old_tv_nut]

Originally Posted by dtvmcdonald

...It a compormise between the best looking Q channel response and the problem that the best response looks terrible because its much too early ... it would need less delay in the Y delay line and the I filter, which has intentional delay in it . I have actual photos of the restult on +-I and +-Q bars posted on an earlier thread here. I will repost here. Note that the time error is obvious here ... but not on program material. Note that there is an orange bar from crosstalk on green to violet but the corresponding bar on green to violet looks more like luminance. Except for the black and white bars, Y is constant. Note that th I channel looks really good.

If by Q "best response" you mean sharpest transition, then that may correspond to wider than normal Q bandwidth and the early Q timing you have. This will result in some quadrature crosstalk of I into Q, which I think is part of what you see in the yellower left edge of the green in the violet to green transition, and the bluer left edge of the violet in the green to violet transition.

Of course, this could be confused by the presence of a slight ringing in the Q circuit response - if you see a ringing transition, it's hard to separate the Q baseband response from quadrature distortion. To separate things, you need to input a step waveform at the Q demodulator output and see that it's not ringing at the CRT.

Also, there's a question whether the chroma in the test source has wideband I and narrowband Q or is equiband.

[old_tv_nut]

I want to modify what I posted above.

To look for quadrature distortion, you can look at the I demodulator output during the Q (green/violet and violet/green) transitions, and vice versa.

If the source has wideband Q, it will cause quadrature distortion in the I demodulator, even in a perfect receiver. This could be a lot of what you are seeing in the transitions between green and violet (but there could also be ringing in the Q signal itself).

NTSC screwed up the Q channel color encoder specs by not including a 920 kHz trap in Q baseband (or a lower sideband trap following the Q modulator), so even with a perfect receiver, it is possible to get quadrature crosstalk from Q into I if the lower Q sideband is not strongly suppressed at the encoder. (The upper Q [and I] sideband is always strongly suppressed by sound traps.)

[Penthode]

I reckoned the effort RCA initially went to in the CT100 was to minimize quadrature crosstalk. With the inductances I mad as described, I was able to achieve these response curves which yielded a picture free from ringing yet obtained opptimal chrman resolution.