21-CT-55 power supply failure
 

The other day I installed the last of the new caps into the '55, so it's a good reliable runner now... or so I thought.

A couple seconds after it came on, the picture got all fuzzy with hum bars slowly drifting up the image. Voltage checks revealed B+ was 360 when it should be 400, and there's 36VAC of hum now. I suspected one of the new filters I installed might have bit the dust since they are only 80uf when they should be 200 (doubler circuit), but replacing them didn't fix it. I suspect the seleniums might have gone bad, so I'm going to replace them with new diodes. Trouble is new silicon outputs more voltage, so I need to know what value resistor to use. I could always just run the variac, but I don't want lower filament supplies. Can anyone with a CT55 tell me what they have done with the power supply?

Nick,

I used a 7.5 ohm 25W tubular ceramic with the adjustable slider on my set. I was able to tweek the B+ right on the money. I used 1A 1000PIV diodes. I hid the resistor inside the rectifier cage mounted with a screw and fiber insulating washers through the center of the resistor bolted to the wall of the rectifier cage.

7.5 ohm? That's it? I would have thought more, since the CTC-4 needed 75...

IIRC, Seleniums have a forward voltage drop of about 5 to 9 volts, and a silicon rectifier has a forward voltage drop of .7 to .9 volts.

Georg Simon Ohm's little contribution says that E=IxR, and we need to drop approximately 7 volts (somewhere in the middle of the selenium voltage drop range).

I'm assuming a 21-CT-55 would draw about 450 watts from the power line at 115/117V, so the current draw would be about 3 Amperes. Not all of that is B+ current, so let's say (worst case) the B+ current is 1 Ampere. It is more likely in the 550-800mA range, I suppose.

Now that we know the current and voltage, getting to the resistance needed to drop that voltage is simple Algebra 0.5:

R=E/I, so 7/1 gives you about 7 ohms, well within the adjustable range of Bob's 7.5 ohm resistor.

My calculations/explanations are full of assumptions, and are VERY open to correction.....I just did the same math for an old 6L6-based amp, and it worked well. Your mileage may vary. Fuses are a must, I might add.

Cheers,

Nick,

The ctc4 has a 10 ohm 25W resistor in the ac side between the common lead of the diodes and 4.5A power supply fuse. The other power resistor you saw on my CTC4 chassis is original to the chassis and is somewhere else in the power supply ckt.

Not a good idea to calculate the resistance needed based on the DC (average) current. It might be right but probably won't be. The voltage drop actually depends on the instantaneous charging current, and that depends on the size of the filter caps and how much ripple is on those caps.

Why not just try different values until the B+ is what you want?

Phil Nelson

I believe that is what Wayne was implying. The overall forward resistance of the silicon diode plus added resistance must match the forward resistance of the original selenium rectifier. This should maintain the instantaneous charging current.

The net result of the properly chosen extra resistor will ensure the same ripple. Hence the resistor is chosen to match the original B+ voltage.

Terry

those seleniums seem never to last long. I just used newer diodes in mine too and kept sels. for looks

OK, so I did some poking around tonight and it's definitely not the power supply like I thought. I see no ripple when the deflection circuits are disabled, but the metal ballast is still getting pretty warm. It seems something from somewhere else in the set is causing the issue, but I don't know where to start.

Does this set have a boost cap? If so, I suspect it might be leaky. There's other caps in the HV garage that still have not been replaced, so I'll start there. Any other suggestions?

I have the orginal seleniums in my Model 5, CT-100 and Westinghouse 15 inch. They've held up well.

Now it's obvious that some circuit is taxing the power supply big time, even with the sweep tubes out the ballast is getting really hot and B+ is still 15 volts low. The drain appears to be on the 285 volt rail, so I guess I'm just gonna have to start disconnecting things till the load goes away. I think maybe a cap somewhere has (partially?) shorted or something.

Also tried replacing the seleniums with diodes, didn't change a thing.

Nick, you could try measuring the resistance of the 285V circuit (with the power off and any voltage bled off). It may measure low since you have a high-current problem. The next trick is to test the "other side" of each part that branches off from that line. One of them would likely measure lower resistance than the main circuit itself, and that would be where your shorted part is.

Something just occured to me. My CT-55 chassis is sitting in the dining room so it's easy to access. If you need a resistance measurment, let me know what you'd like to compare; that may be helpful.

Regards,
Mark

Sure, what's the resistance reading you get on the 285v line with the ballast pulled? More importantly, how much current does the 285 rail pull when working correctly?

Please give me a few days; due to the upcoming holiday weekend I have commitments that will take me away from home. I can easily measure the resistance; however, because the chassis is not in the cabinet, I am not presently able to fully power-up it up with a deflection load to measure total current. I may have that information from the manufacturer though. Alternatively, we should be able to calculate the nominal current based upon voltage drops (if indicated on the schematic) across the ballast. I will study the diagram if other readers cannot answer this empirically. I prefer to make these measurements with my Simpson 260 rather than a digital meter. I will provide measurements with reversed polarity as well, since in theory, they could differ.

With regards, to your recent acquisition, congratulations. I seem to recall this prototype was well described in the Second Color Television Issue of the Proceedings of the IRE, 1953. If you have not already studied this report, let me know and I will dig it out for you.

Regards,
Mark

Anything for the proto set should go in its respective thread: http://videokarma.org/showthread.php?t=251369

I too have commitments this weekend, I'll get back to the CT-55 next week.

K, so I finally had more time to look at this.

I have now ruled out the power supply as a failure source. I have disconnected the following items from the power supply, to try and isolate the current drain:

Everything on the 285v 'IF' rail. Nothing in the audio or IF sections is getting power anymore, I accomplished this by disconnecting pin 2 on the ballast.

I disconnected the tuner from the other 'normal' 285 rail so the only things getting power are the sweep circuits, video (Y), chroma paths and whatever convergence circuits may still be hooked up.

I am putting a composite signal from my gen into the grid of the 6CL6 to get video and the set again has a nice color picture, but read on.

Still getting the hot ballast, so I started looking hard at the horizontal section. Oscillator still had some wax caps in it, but replacement did nothing. Adjusted the wave shaping coil for equal peaks (it wasn't far off to begin with), no change. HOT is drawing 225ma (cathode), too much for my comfort. Tube data sheet for the 6CB5 says 220 is the safe limit as a deflection amplifier, so I have to get that down somehow. Messing with the drive adjustment won't get it low enough, and the lin coil is set correctly.

Grid bias on the HOT is high by 16 VDC: should be -56, but I get only -40 on the meter. Also, with the HOT pulled, I get a perfect waveform on the scope. Putting the HOT back in causes the waveform to become sort of flat-topped, which I don't think is good. I could be wrong, but I believe this means it's being driven as hard as possible and may be drawing excessive grid current. This and the high grid bias are likely related, but I can't figure out how.

At this point, I have subbed out the HOT, damper, shunt reg, horizontal osc, and vert output. None had any impact at all. The fly windings are in the ballpark with the resistance values called out in Sams, so I have a hard time believing it's got a shorted turn.

Only tubes I can't sub are the 3 rectifiers, since I don't have any 3A2's. Only thing I have left is maybe the shunt reg is set incorrectly, sucking away too much power from the HV. Wayne says he will stop by with his HV probe tomorrow evening, so we can check. Beyond that I'm not sure where to go next, short of cap shotgunning and a test of all the tubes left in the chassis.

I apologize for not making the resistance measurements in a more expedient manner. Weather and family issues are, or were, weighing on me heavily.

I will look at my RCA service literature again for this receiver to see if I can offer any useful guidance. It is doubtful, in my opinion, that there exists a shorted winding on the horizontal output transformer. In my experience, even one shorted turn results in catastrophic effects; nonlinearity, foldover, or complete failure of the transformer after a few minutes of operation. A “ring-down” test is simple to execute and provides a pretty comprehensive assessment of the transformers health. Certainly you are familiar with this, but for readers who may not be: http://www.repairfaq.org/sam/flytest.htm

While a transformer failure isn’t out of the question, those employed in the CTC-2B chassis were far more durable than say ones in later models like the 16’s or 17’s.

The drive waveform asymmetry would be a place I’d start investigating.

Regards.

Agreed. So I understand that the waveform applied to the tube is what develops the bias, but what could cause it to be too high? It's possible the high voltage doorknobs are getting leaky, causing HV to sag. I already repaired one that had fallen apart, but I'm not sure it's possible to get an exact replacement anymore.

Any chance the HOT grid coupling capacitor might be leaky, raising the grid voltage?

That was the first thing I replaced.


Wayne stopped by tonight, so we had some fun. First we used the HV probe, which revealed a perfectly healthy 24.5kv. The set makes a nice picture when driving a composite signal directly into the video tube, but going through the tuner/IF produces snow. There's still some kind of load on one of the B++ lines, it's still measuring 40 odd volts low. We tested all the tubes I haven't already subbed, no surprise we didn't find a single one bad.

This thing is kicking everyone's ass all over the place, we just can't figure out why it's doing this. I have my suspicions about a leaky cap still, but the only thing I have left as far as troubleshooting is a shotgun recap of the leftover micas.

CT-100: is there any way you can give me a current reading from both legs coming out of the ballast tube? Just insert the meter between pins 2 and 3 on the ballast and let me know what you find. At this point I have nothing else to go on.

Refering back to your first post: When you replaced the two voltage double caps in the LV PS (C103 & 104), did you replace with correct 200 uf, or did you put new 80 uf back in?

If your HOT grid bias is too positive, that could be caused by not enough grid drive. Do you have the correct 200 and 270v on the plates of the Horz OSC tube? Does the drive signal amplitude change much with the HOT pulled? How's the voltage on the screen grid of the HOT?

If it can be of help I have a heathkit eye type cap checker that I recaped reciently which I'd be willing to lend to you. It seems to work well... new caps test good and old ones in various TVs test anywhere from good to completely dead.

Tom C.

I suppose I should have my own by this point, but I've always taken the position that if in doubt replace it.

Another day, another 12 caps replaced, still no change. There's definitely hum bars rolling vertically up through the picture when using an RF signal through the tuner, perfect picture when injecting composite into the video amp. I'm at a loss.

Have you considered a possible heater to cathode short in a stage prior to your baseband video injection point? That is the Tuner through the Video IF. A heater cathode short could inject the hum and could create a higher current draw if a stage is dependent upon a cathode resistor bias arrangement. And a partial short is sometimes not obvious in a tester.

Terry

I thought the tester would catch something like that, but I suppose anything is possible. I don't have any 6DC6 tubes, so I can't sub anything in the IF strip. I can probably borrow a complete set from Bob G though, I'll ask next time I see him.

A 6CB6 or 6BZ6 is a reasonable substitute for for the 6DC6 for testing purposes. The tranconductance is a little higher for these tubes (8000umho as opposed to 5500umho for the 6DC6).

Of course you would not want to mix up the existing tubes or necessarily permanently exchange the IF tubes as realignment would be advised. But I would try a little step by step substitution to see if the hum clears up.

One other thing: when you inject the baseband composit signal to the video amplifier, do you deactivate the RF/IF section? Or does the hum only appear when the signal is applied to the antenna?

Also, is it 60Hz hum or 120Hz hum?

Bought some IF tubes today, didn't make much difference. The noise got a little lighter, but it's still there. I'm not sure how to tell what frequency this is, I only know that there are bars floating up the screen. I'd shoot a video, but I'm not sure the camera would pick up on it.

The noise is only there when passing a signal through the tuner/IF, when I put composite on it the tuner/IF were not powered and there was no noise.

I also have an unrestored heathkit cap checker I'd be willing to sell that is burried in storage, but can probably be located next time I go there, if you want to own one.

As long as you take that Silvertone in trade, I'm game. lol

120Hz is ripple from the B+ supply. With 120Hz you will see two distinct bars floating up the screen.

60Hz will originate from the mains directly and you will see a single bar.

How many bars are there? Can you post a photo of the screen with the bars present?

And have you substituted the tuner tubes?

If you are refering to the console that I did not take then the only way that will happen is if I sell some TVs or toss some out. My folks have ordered me to stop collecting anything bigger than large table radios for an indefinite period of time.

Tom C.

Yes I can make the measurement, but this again, is going to take a few days. My primary delay is that I have not powered my chassis up since I am working on the CT-100. I am aware of your request and will do what I can as quickly as I can. If you lived in my neighborhood, I'd give you my chassis and you could take it to your lab for precise comparisons.

I have a lot of background with this particular set as my dad brought 5 or 6 of them home as they became obsolete; one even had an IF-RF problem like you have described. We're talking 1968, so I have to review the schematic and try to reverse engineer the solution he discussed with me. (I couldn't work on sets then, he did and I only watched the process)

If the grid bias is too high, then the grid current must be too low [OR] the tube is being over-driven. I need to get my diagram; is there a cathode resistor in the deflection output; if so, is it by-passed? I will look this weekend for sure.

I'm max'ed at work and have not fairly studied the responses ahead of this, so forgive me if I am repeating suggestions. I admit, I have not done my homework yet so please take these comments with that in mind.

Regards.

To me it looks like a pair of bars on screen at one time travelling up through the pisture, but the power supply has no ripple so I'm not sure what to think. I've already subbed every tube in the tuner/IF and even the first video (6CL6), nothing made an improvement. If I had to pick a word, I'm flummoxed. :sigh:


I think the tube is being overdriven, because it's DC grid bias is not what it should be. As I said before, I'm going on the assumtipn that the cause of the flat topped waveform at the grid is being caused because the tube cannot be driven any harder than it already is. Result: power comes from the only place it can, in the form of grid current.

It seems you explored the most obvious plus some of the most obscure prblems. maybe signal tracing with a scope is the best way to track this down.

Can to apply to the antenna a test signal of a flat field of grey. Ensure you can see the bar clearly on the CRT. Then 'scope back from the CRT stage-by-stage. I'd expect you should see the hum when scoped at the CRT and I would trust at some stage the hum will disappear from the scoped waveform.

At this stage I am as flummoxed as you...

Terry

I'd suggest trying to duplicate the SAMS resistance readings at the tube pins in the IF and tuner - might lead you to an anomaly.

You know what's interesting? My SAMS shows 60 Hz Hum even though this is, in fact, a full wave PS configuration.

My RCA data is quite explicit, the drive waveform geometry is critical to the proper operation of this circuit.......