[sorenj07]

Finished this last night.



Looks and sounds great except for an ear-splitting oscillation in the right channel, at about 1050 Hz. It plays music on both channels, in other words. I have tried flipping each channel's OPT primaries, and the problem has remained. I moved the little ground wire connecting the RCA inputs to ground down near to where the volume pot is.

I have switched each individual tube between channels, and haven't noticed a change. Other than that, each channel is wired in EXACTLY the same way as far as I can recall.  Anyone have any ideas of what to try? I'll probably take it down to the bench tomorrow and go over each resistor.

[JoshK]

you might consider 100 ohm cc grid stoppers.  i am not that familiar with the 6f6g, is it a video pentode/  if so it might need a grid stopper.

[JoshK]

is the oscillation in the channel where the psu components are next to the output.

[sorenj07]

you might consider 100 ohm cc grid stoppers.  i am not that familiar with the 6f6g, is it a video pentode/  if so it might need a grid stopper.

I'll try them. The 6F6 is an old audio pentode, which was superseded by the more successful 6V6 and 6L6.  Here's some info:
http://www.nj7p.org/Tube4.php?tube=6F6
http://www.mif.pg.gda.pl/homepages/frank/sheets/049/6/6F6.pdf
Mine are Russian copies but the specifications should still apply. 

is the oscillation in the channel where the psu components are next to the output.

Yes. Do you think this might make a difference?

[JoshK]

Yeah...I think you might have too much EMI/RFI emitting from the rectifiers into the first cap that is coupling to the output wiring....it is really easy to do with the proximity that close.  For kicks, put a power resistor after the rectifier before the first cap.  It'll lower B+ but see what it does to the oscillation. 

[sorenj07]

It's a possibility. However, the rectifier diodes are Schottkys, so wouldn't that reduce EMI/RFI emissions? The blue bypass cap is awfully close to the black .22uF signal coupling capacitor, though. Do you think I should try scooting it over to the right?

[JoshK]

It's a possibility. However, the rectifier diodes are Schottkys, so wouldn't that reduce EMI/RFI emissions? The blue bypass cap is awfully close to the black .22uF signal coupling capacitor, though. Do you think I should try scooting it over to the right?

I think Schottkys reduce the effect but don't eliminate it.  If you modeled the PSU in PSUDII you can look at the current waveform in the first cap and it might give you a hint whether you have EMI generation. 

You could try moving the blue bypass over.  Ideally they wouldn't be parrallel, but what can you do?  I am a bit more leery of the stuff up in the right corner of the underside picture. 

[sorenj07]

Here's what the waveform looks like. Pretty bumpy. I'm guessing the sharp knee where the wave spikes up might be where EMI is generated?



What is it about the right corner? Since the oscillation happens in the right channel, isn't the left half of the underside what needs to be worked on?

[JoshK]

Oh, I was thinking you said yes to it being the channel with the PSU components next to the outputs (in the upper right).  Is that just the bias supply?

If the right channel is oscillating and not the left then it might be the coupling caps so close to eachother.

[sorenj07]

Yep, it's the bias supply up there. Not a lot of current.

Two coupling caps of opposite phase can cause oscillation when they're too close to each other? Wow.    I'll try to pry them apart but then there's the blue guy immediately to the right. Serves me right for building it in such a danged small box.

[JoshK]

I am not saying that is a for sure, just a possibility.  I don't have enough experience to do much more then help you think of ideas on what might be. 

[sorenj07]

All right. That's how I build amps anyway, a bunch of rules of thumb you watch out for.  I'll keep you posted on what happens when I do more tinkering tomorrow (today).

[GBB]

Looks and sounds great except for an ear-splitting oscillation in the right channel, at about 1050 Hz.

 Anyone have any ideas of what to try? I'll probably take it down to the bench tomorrow and go over each resistor.

I worry about the way you've set up the bias supply for the 6SN7 tubes that drive the 6F6 output tubes.
By deriving the bias voltage from the -300v supply that you're using for the 6SN7, you are providing a possible feedback path that can lead to oscillations depending upon the parasitic capacitances that exist.  I think you'd do a lot better if you created another -50v supply.  You can still derive that from the -300v supply but you need give it enough capacitance that it really doesn't "see" any of the things happening with the -300v supply.

---Gary

[sorenj07]

Hmm. All right. Do you think I could try upping the value of those little .1uF caps to ground? Maybe add a 100V electrolytic in there? Space is incredibly cramped but I can probably try to re-work things.