Negative Feedback Causing Oscillation?

[AHeck]

Good afternoon, hope you are all very well and thanks for considering this post.  This is an amp I put together and I'm having some stability issues at very high volumes only when I engage the NFB options. The more feedback, the more howling.  What have I not considered? 
Also wondering about the P to S voltage on V1?  is this extreme?
The amp sounds good, but it's not as stable as one I built earlier this year. 
Thanks for the help.

[AHeck]

This is the speaker out and NFB circuit.

[AHeck]

Layout Picture

[AHeck]

Another Layout Picture

[nandrewjackson]

470R and 1K should be huge amounts of NFB.


Is the black through-panel thing the impedence switch?
It's my understanding that the NFB resistors are to be tied to only one of the OT taps. If that black receptacle is the impedence switch, you're also switching your NFB back and forth between OT taps.


I don't know how hot the signal is coming out of your CF, but having a 0.1uF there, and having 0.1uF on both sides of the phase inverter output seems like too high capacitance.


I don't know anything about your input tube stage, but I would lower the CF output coupling cap to 0.02uF.

[AHeck]

It's my understanding that the NFB resistors are to be tied to only one of the OT taps. If that black receptacle is the impedence switch, you're also switching your NFB back and forth between OT taps.

Of course, you are quite right.  I’ll go and correct that immediately. Good looking out. Thanks.

[sluckey]

Your schematic is not believable. Output tube grid circuit is really messed up.

[PRR]

470R and 1K should be huge amounts of NFB....

Or Positive Feedback. Can't tell from here.

[AHeck]

Your schematic is not believable. Output tube grid circuit is really messed up.

I was trying to write out the dual 250K pot master volume.  I didn’t get it right, did I?

[pdf64]

470R and 1K should be huge amounts of NFB....

Or Positive Feedback. Can't tell from here.

The key question is, compared to open loop, does the feedback cause gain to be increased or decreased?
Positive feedback doesn’t necessarily result in free running oscillation.

[sluckey]

Your schematic is not believable. Output tube grid circuit is really messed up.

I was trying to write out the dual 250K pot master volume.  I didn’t get it right, did I?

Ah ha! Now it makes perfect sense.

You used a triangle for your ground symbols and you used the same triangle for the pot wipers. But the pot wipers did not actually touch the pot. My mind just processed the pot wiper triangles as two more ground symbols. Sorry for the confusion.   

[AHeck]

470R and 1K should be huge amounts of NFB....

Or Positive Feedback. Can't tell from here.

The key question is, compared to open loop, does the feedback cause gain to be increased or decreased?
Positive feedback doesn’t necessarily result in free running oscillation.

The gain increases when I switch the NFB on.  Did I miscalculate the polarity through the preamp stage? 

[AHeck]

Your schematic is not believable. Output tube grid circuit is really messed up.

I was trying to write out the dual 250K pot master volume.  I didn’t get it right, did I?

Ah ha! Now it makes perfect sense.

You used a triangle for your ground symbols and you used the same triangle for the pot wipers. But the pot wipers did not actually touch the pot. My mind just processed the pot wiper triangles as two more ground symbols. Sorry for the confusion.   

Yessir.  Sorry about that.  I’m still trying to find and install a good potentiometer model on LTSpice. 

[pdf64]

…
The gain increases when I switch the NFB on.  Did I miscalculate the polarity through the preamp stage?

Feedback that increases gain is positive 
The balanced area of the circuit is usually where mistakes are made.
If you’ve got a scope and sig gen, it’s simple to verify the signal polarity.

[tubeswell]

Aside from having positive feedback, the voltage divider options in the NFB loop on your schematic will provide quite a lot of feedback (as someone else already mentioned). With LTP inverters, ‘less is more’. Look at tried and true examples: 820R and 47R (BFDR) or 56k and 1k5 (6G16), 27k and 5k (5F6A). Even too much NFB can result in positive feedback at HF (particularly with output transformers that are wound with several interleavings, where interwinding capacitance produces more phase shift at HF, which can add up to positive FB).


But try swapping the OT primary ends to the opposite output tube plates first.

[nandrewjackson]

Other examples of that type of MV show 2M2 resistors to bias tap. Yours shows 1K8.
That's quite a difference.

[AHeck]

But try swapping the OT primary ends to the opposite output tube plates first.

Thank you, no more oscillation.  I can max out the Vol, Master and guitar and it's quite quiet and stable.

[AHeck]

Other examples of that type of MV show 2M2 resistors to bias tap. Yours shows 1K8.
That's quite a difference.

Yes it is.  There are 1M8 resistors at the wipers of that pot.  Thanks for catching that.

[Williamblake]

So you heater centertap is connected to 0VDC via a 50uF capacitor? Why?

[AHeck]

So you heater centertap is connected to 0VDC via a 50uF capacitor? Why?

There’s a voltage divider off of B+ at the 47uF Cap. The resulting voltage is a little over 50VDC and tying the heater CT in there elevates heater ground so the Heater to Cathode voltage comes in under the 200V maximum at V2b.  The capacitor to ground helps with heater hum.