RCA BA-6A problem - SOLVED!

[noulou]

What about it?  I'm having trouble imagining a transformer problem that wouldn't also present as an obvious response problem.

No idea. It sounded fine when not compressing.  Just impossible to get 2 tubes to match. I had tried around 20 different 6sk7s and their combinations.
I have that interstage transformer here somewhere. I have been planning to test it when I have the time.

 

[scott2000]

Make sure that the voltage regulator tube is operating properly (OD3) It sets the 'threshold' voltage at the GR diodes. I mention it only because it wasn't mentioned earlier, and it's important to the GR process.

This makes sense???

 

[audiopete]

Yes the 6SK7s balance up fine. I could even get them to match almost spot on with DMMs.

Interesting about the interstage. This unit also passes signal fine below threshold. I will try 500Hz and 1KHz when I get to it today. I also have 2 new NOS pairs of 6SK7s to try out.

I tried several OD3s (VR-150s) and no luck.

 

[audiopete]

I'm also going to look into C20.

 

[audiopete]

I wonder if any of the connected heater/hum circuit has any bearing here?

 

[audiopete]

I'm also going to take some DCR measurements of the interstage and compare windings.

 

[audiopete]

On the advice of another tech friend who's helping on this, I added .47uF to C1 and it still clamped down hard, but after that, it was reducing gain a few dB. If I clicked another input level, it dumped again but came back and seemed to provide proper GR. I am thinking about trying the 1M pot on after the 6H6 anodes and see if it can be dialed down.

 

[audiopete]

New 6SK7s, no improvement.

 

[David Kulka]

Sometimes when working on these type problems I'll disable the sidechain section, connect a DC bench supply in its place, and  try using that to control audio level through the vari mu stage. Doing that may tell you whether your issue is in the sidechain or in the vari mu section.

 

[CJ]

send the transformer and we will "figure it out", 

 

[audiopete]

Thanks David. That is a good next step.

Interestingly enough, it is now partially working. It still thumps down hard, but it returns and provides GR. If I click some more input, it shuts down to zero current (according to meter on V1, V2) but comes back to more compression,  and the output signal is staying limited. When I come back down, it un compresses as you would expect until the signal drops below threshold.

2 things changed yesterday - I put in new pair of 6SK7 and calibrated up with a DMM and I rearranged  the wiring a little thinking lead dress oscillation. Something has changed and it might not even be related to those two things, as experience will me. I need to look more into this and go over voltages again and such. Also I have a couple more 6SK7 combos I can try. One of the NOS ones I got was shorted. Back for refund.

I did some measuring of the IT primary and it looks good at about 1.2K a side and 2.4K total.

My father taught me a lot about persistence

Thanks everyone for your time.

 

[emrr]

Schem not in front of me, but recall input Daven between trans and grids, bias dc across it.  Worth disconnecting and measuring at all positions, compare side to side.  Likewise check that input secondary is all there. 

 

[audiopete]

Hi Doug, I did that and all looked balanced and good as far as readings went.

This is where I am:

I will be pulling the 6H6 and connecting a 9V battery and divider pot for a neg voltage source to test the 6sK7 action on signal. Maybe then I will have more info to go on.

My 6J7 stage is running hot and I would like to bring that back into spec.  Tried new tubes - closer but not in spec. Maybe investigate the screen circuit or change the cathode resistor.

Try a pot on the neg control voltage line back to the grids and see if that works.

The owner mentioned that this thing "never worked right" recently to me. I didn't question him more at the time, but it got me thinking. I'm pretty sure this thing is a teeder todder of balance and back when good tubes were solid and plentiful it worked great. Now with sketchy NOS tubes left over, perhaps it needs some mods to calibrate in less than perfect tubes.

Hopefully more troubleshooting will give me these answers.

Cheers everyone.

 

[emrr]

I can only add that I have a pair of RCA/NBC in-house comps that use 6K7's and a 6R7, and I still have yet to find a set of tubes that have the same current draw as what was coming off the line in 1940.    Getting a GR match is not a problem.  Yet the action of the two units is not close enough for most stereo usage (I installed a link circuit) and they've been extensively rebuilt.  They're just such organic beasts in action, there's apparently no taming them entirely, and being from the mono age, there wouldn't have been any thought to exacting consistency.  A real house of cards.  They act a bit differently than they did before, every time they are in use. 

Good luck, I'll be interested to hear what you find. 

 

[audiopete]

Update - Found that C2 (first stage filter cap) was never grounded from day one. I found the original unsoldered end under the wire bundle. It never got soldered to the ground buss bar. Unfortunately that didn't fix the issue. Now I won't get shocked by the top of C2 again.

Pulled the 6H6, hooked up neg voltage supply, works beautifully. Nice smooth cutoff of the 6SK7s.

This would suggest the problem is in the neg control circuit. I'm still seeing wicked neg spikes of voltage when it crosses the threshold and would love to know what's causing that. I wonder if it's output tranny related. I've swapped the caps off the OT and the feedback caps back to the 6J7s no change.

Also interesting is when I bring the level close to threshold, a tap or bump up around the 6SK7s triggers the dive.

Any thoughts would be most welcome. I need to think about this here.

 

[Monte McGuire]

Progress! Given one soldering mistake, maybe there are others? The sensitivity to shock seems to suggest some sort of mechanical defect, but it could simply be a microphonic tube. As long as you can safely shotgun the GR section solder joints, it might be worth it. Another thought is the GR hold capacitor has faulty internal terminations, and 'comes to life' only with voltage across it.

Good luck - sure is an annoying problem!

 

[emrr]

Leakage/arcing at the 6H6 socket?  Like the filament hitting the diode? 

I'd think any output transformer problem would be seen in the audio output with the 6H6 pulled.

Any chance the Bal-B pot is going open at the wiper intermittently? 

 

[audiopete]

Monte:

It's is very strange. I will confirm today, but pretty sure when I had the external neg supply hooked up and the 6SK7s were working perfectly, it didn't react to mechanical shock. On the other hand, it must also be electrical since it will cycle and do it all day if I set the input level high to trigger it.

Doug:

Since that socket is original and all the others have been replaced, that did cross my mind. I have a socket, but it's one of those can't know till I do it things. There is a rise of a volt and a half in the diode bias when this event triggers. More on one diode than the other.  I could see a volt or two of AC here potentially causing this. It seems like it has to overcome a momentary higher threshold, and when it does the recitified negative DC is too high, but maybe the leaking volts are contributing and the result is too high rectified DC.

haven't observed any Bal B issues. Screen voltages are solid there and do not move, other than rise a little when tubes cutoff, as expected. Both tube screens and cathodes matched by DMM.

 

[audiopete]

Update - still looking at this and I have it down to what I believe is an oscillation building up on the bias line to the 6H6 between R44/45 and between R53/54. I can clearly see on the scope the 100Hz signal start to wobble, distort, then finally a spike, and the 6SK7s shut down.  Then it comes back up clean and starts again as it hits threshold. All other places like 6V6 grids, plates, 6J7 is clean sine until it collapses. I believe the resonant peak takes off and rectifies into the massive negative spike. If I clip my scope probe to one side of the bias resistors, it stops pumping, but an input adjustment higher will trigger it. All this reinforces my oscillation theory.

I noticed on the clone schem that not only is the bias line moved onto the OD3 regulator, but there is a 22uF decoupling (?) cap there that doesn't exist in the original.  Could this be the key to my issue? I'm curious about the reasoning behind these changes.

Cheers and thanks for your time.

 

[jim-analog]

Greetings,

I had a unit in my shop about 10 years ago with what seems to be the exact same problem. Worked on it for well over a month and it drove me nuts. I also replaced or swapped everything at least twice, all tubes, all caps, many resistors, went through each of the multi deck switches several times re-soldering every connection in the entire unit and verifying every wire, took apart and cleaned the Davon attenuators,  etc.; literally everything I could think of.

Well, it DID turn out to be the IT!  What was interesting is that under DC bench test conditions, the IT tested fine. Trying to test it in circuit, the Fluke 8060a meter I was using put too much load on that high Z portion  of the circuit and was completely confusing things (and me).

  If you have a VTVM available, you may have better luck and may be able to see the overshoot occur. What was happening was an internal breakdown of the IT windings. There would be a dramatic overshoot, and then it would gradually recover till the next over threshold signal started.

  I proved the cause by quickly jumping in a tube P-P output tx in place of the IT. I can't recall exactly what it was that I used and it wasn't close to the same spec, likely something from a 2 x 6L6 or 6V6 gtr amp. It worked well enough to absolutely prove the IT was the problem. If it would help, I can look back in my files and get you all the details.

  I was able to order a very good replacement from Sowter.  Getting the old core out of the can was another difficulty. A hot air gun,  several hours and MUCH smoke till enough of the black tar goop melted to pull it out. I used bees wax to "pot" the new one in and also had to repaint the can.

Of interest is I had two units from the same client. He had purchased them "as is" untested and without tubes, so there was no prior history to go on. They had not been stored in a very good environment.

The first one I rebuilt came out great with no difficulty. The other one I refer to above.  When they were both complete, testing showed no discernible difference between the unit with the factory IT and the Sowter replacement. The client was unable to hear or notice any significant differences between them and was very happy to get them back.

Good luck, I'm glad to go look up my notes if it would help you out. I know how frustrating that one was for me!

Regards, Jim