65 Bassman oscillation problem

[mresistor]

Guys, I've got a 65 Bassman AB165 on the bench.  It had the 84 Twin OT in it which is now replaced with a CT 40-18001 OT. The local feedback around the power tubes is removed and the bias has been converted to AB763 style. The Volume and Treble pots have been replaced.  When I first hooked up the OT secondary leads I got that nasty sound, so I reversed them and then it sounds pretty good However a little bug has crept in. This concerns the Normal channel.  I've been monitoring the 5881 cathode current while operating the amp and when I turn up the treble I get a nasty oscillation sound, very high pitch and when it happens the cathode current doubles at least, it jumps easily from 33ma to 66ma and even higher. It has gone as high as 100ma. It is worse with the bright switch on. I am assuming this is an oscillation, and at times it has had a sort very faint tremolo effect to it.
The bass channel does not exhibit this behavior at all.
The volume and treble pots have been replaced because the old ones were not good. I've subbed in another bright cap to no avail.
I am going to check the tone stack caps here shortly.
Does this sound like something a tone stack capacitor could do?
Since the bass channel is basically stock and has no problems, one would think that the power section is ok, but the bass channel doesn't pass high frequecies like the normal channel does so I hesitate to assume that the power section is OK.
Oh  and this amp came to me with a 27K feedback resistor in it. 
Any ideas what could cause this oscillation?

[mresistor]

https://el34world.com/charts/Schematics/files/Fender/Fender_bassman_ab165_schem.pdf

[mresistor]

Also I have substituted every tube in the amp with no change.  The .1uf and .047uf capacitors don't show DC leakage and check within tolerance, they are original blue molded caps.

[pdf64]

So the local and global nfb on the power amp has been removed? That’s increased the gain and lost the margin of stability.
Does reinstating the nfb restore stability?

[mresistor]

The global feedback is there with a reduced value resistor, 27K instead of 47K.  The third stage 12AX7 feedback is present.  The local feedback around the power tubes is removed.   I was thinking I'd put it back in and see if it corrects the problem.
Thanks.

[pdf64]

27k is more feedback, ie a greater degree of nfb with respect to the regular signal input.
Again, that could reduce the margin of stability too.
So maybe put that back to stock.

[DummyLoad]

there's a much better way to insert NFB - ditch the .1uF and stuff a 47R under the tail - use the AB763 super reverb NFB values. leave out the 220K local NFB for the 5881s 6L6GCs. install a 47-100pF cap across the LTPI plates - install that cap at the tube socket or in the board across the plate Rs.   


--pete

[mresistor]

Thanks Pete.  I'll give this a go.

[pdf64]

Changing from an inverting (Bassman AB165) to non inverting (eg SR AB763) power amp arrangement also requires, as the description indicates, the polarity of the balanced signal to be reversed.

[mresistor]

there's a much better way to insert NFB - ditch the .1uF and stuff a 47R under the tail - use the AB763 super reverb NFB values. leave out the 220K local NFB for the 5881s 6L6GCs. install a 47-100pF cap across the LTPI plates - install that cap at the tube socket or in the board across the plate Rs. 
--pete

Well, inserting a 47R in the tail is going to be quite difficult physically, as there is an under board jumper from the 22k/.01uf junction to ground. Have you done this before and if so how did you accomplish it?
 Pdf64 - I'm not sure he's talking about changing the whole PI input topology, what I was understanding he said is eliminate the .01uf cap in the feedback loop and change the feedback R to 820R. Now I'm not sure what exactly Pete was saying there.

[DummyLoad]

Changing from an inverting (Bassman AB165) to non inverting (eg SR AB763) power amp arrangement also requires, as the description indicates, the polarity of the balanced signal to be reversed.

how is that changing the polarity? - grid input and cathode of the LTPI are in same phase relationship? looking at the layout for both confirms that it would work with some hacking - see below. 

Well, inserting a 47R in the tail is going to be quite difficult physically, as there is an under board jumper from the 22k/.01uf junction to ground. Have you done this before and if so how did you accomplish it?

i have not done this mode before, i have only read about it. sad to say, but zero AB165 have crossed my bench. looking at the layouts, i see what you mean. so back to the drawing board.    you could try to separate the LTPI/power stage from the preamps at that same NFB insertion point by lifting the .1uF back to the preamps then test for oscillation without the LTPI/power stage in the mix. could help to determine if the instability is originating in the preamp section or is originating from the LTPI onward. 

--pete

[sluckey]

I would do this. Yes, the underboard jumper must be removed. Just requires some patience. Unscrew the board. Remove all solder from the two eyelets involved. Use curved needle-nose, forceps, spring hook, or whatever you have to get hold of the jumper wire and pull it out.

Will this solve the oscillation problem? I have no idea. If you need to reverse the mod, put the jumper on top of the board.

[mresistor]

Thank you.   Great suggestion..

[mresistor]

  you could try to separate the LTPI/power stage from the preamps at that same NFB insertion point by lifting the .1uF back to the preamps then test for oscillation without the LTPI/power stage in the mix. could help to determine if the instability is originating in the preamp section or is originating from the LTPI onward. 
--pete

Thanks Pete   good idea,  I do need to isolate where this is coming from.  The treble pot is acting strangely ..   It seems to only start working halfway up.

[mresistor]

Many thanks to DummyLoad, Sluckey and PDF64 for their assisstance in fixing this beast.  It is working well now with the little mod in it. I did have to swap the OT plate leads so insertion of this mod did  put the feedback out of phase.  ( note: I hate that noise and so does my speaker LOL )

[pdf64]

Changing from an inverting (Bassman AB165) to non inverting (eg SR AB763) power amp arrangement also requires, as the description indicates, the polarity of the balanced signal to be reversed.

how is that changing the polarity? - grid input and cathode of the LTPI are in same phase relationship?...

Indeed but that’s just looking at its input stage, whereas when considering feedback implementation, we must consider the power amp system as a whole.
Note that for the Bassman AB165, the voltage sample of the output signal is fed back to the same power amp input as that used for the signal. Hence for the feedback to be negative, the power amp must be configured for that to be its inverting input.
Whereas for the SR AB763, the voltage sample of the output signal is fed back to the ‘other’ power amp input. Therefore the ‘other’ power amp input must be its inverting one; if not, the feedback would be positive not negative.
So from the above, it is apparent that the Bassman AB165 is configured as an inverting amp, whereas the SR AB763 is configured as a non inverting amp. Hence when changing from one to the other, one way or another, the polarity of the power amp will need to be flipped.
Hope that made sense 

[2deaf]

The LTPI is a differential amplifier that amplifies the difference between the signal at its two inputs.  With the AB165 LTPI, the input of the non-inverting stage is referenced to ground so that the difference between the two inputs is the same as the input of the inverting stage.  With the SR AB763 LTPI, the NFB signal is presented to the input of the non-inverting stage.  The NFB signal needs to be in phase with the signal at the input of the inverting stage in order to reduce the difference between the two inputs.  The reduction of the difference between the two inputs will result in the reduction of the outputs from the LTPI, which is what NFB is all about.   
     

[pdf64]

The LTPI is a differential amplifier that amplifies the difference between the signal at its two inputs...With the SR AB763 LTPI, the NFB signal is presented to the input of the non-inverting stage.  The NFB signal needs to be in phase with the signal at the input of the inverting stage in order to reduce the difference between the two inputs...

The above seems topsy turvey.
The NFB signal is a voltage sample of the output signal. If it was applied to the non-inverting input, the feedback would be positive and the amplifier would be prone to instability.
Hence the output and input signals must have the same polarity; the power amplifier block may be represented -

With Rf being 820ohms, R1 being 47ohms.

...With the AB165 LTPI, the input of the non-inverting stage is referenced to ground so that the difference between the two inputs is the same as the input of the inverting stage...

Indeed, the power amplifier block may be represented -

Rf is 47k and R1 is formed by the output impedance of the preceding stage.

With both the AB165 Bassman and the SR AB763, the input signal is applied to the same LTP input. Hence when comparing the 2 power amps, somewhere in the power amp block, the signal polarity (and so inverting / non-inverting nature of the LTP inputs) must get flipped.

[2deaf]

The LTPI is a differential amplifier that amplifies the difference between the signal at its two inputs...With the SR AB763 LTPI, the NFB signal is presented to the input of the non-inverting stage.  The NFB signal needs to be in phase with the signal at the input of the inverting stage in order to reduce the difference between the two inputs...

The above seems topsy turvey.
The NFB signal is a voltage sample of the output signal. If it was applied to the non-inverting input, the feedback would be positive and the amplifier would be prone to instability.
Hence the output and input signals must have the same polarity; the power amplifier block may be represented -

With Rf being 820ohms, R1 being 47ohms.

Topsy Turvey?  I looked that up and it seems to be characteristic of those that have difficulty in distinguishing the difference between their anterior and posterior.  That could well be, but what I said is dead-on correct.

The NFB signal is a voltage sample of the output signal.

Yeah.

If it was applied to the non-inverting input, the feedback would be positive and the amplifier would be prone to instability.

Not necessarily.  It would depend on the polarity and you didn't specify that.

Hence the output and input signals must have the same polarity

That is exactly what I said despite my alleged condition, so I really don't see any argument here.

the power amplifier block may be represented -

That seems a little topsy turvey to me.  I prefer my page.  It doesn't specify polarity, but the equations make it abundantly clear that V_in would have to be in phase with V_t in order to have negative feedback.


If you are going to quote me, then your comments have to refer to what I said in that quote.  You keep going back to the entire voltage amplifier block with its inverting and non-inverting nature.  I was specifically referring to the LTPI alone and its inverting and non-inverting stages.  What you said about the entire voltage amplifier block and what I said about the LTPI alone are not mutually exclusive.

 

[PRR]

>> If it was applied to the non-inverting input, the feedback would be positive
> Not necessarily.  It would depend on the polarity and you didn't specify that.

I think the 'polarity' is implied by 'non-inverting', which is always relative. To what? I would assume the output which is sampled.

The naked long-tail-pair has "both polarities" of input and output available so that's not carved in stone.

[pdf64]

...

the power amplifier block may be represented -

That seems a little topsy turvey to me.  I prefer my page...

I can't perceive how one might consider one topsy turvey, yet the other not; please explain? I thought I had explained why I thought the text of yours that I quoted was topsy turvey, ie because applying NFB to an amplifier's non-inverting input seems inherently self contradictory.

Your page only appears to deal with the SR AB763 non-inverting amplifier arrangement; how about for the Bassman AB165 inverting amplifier, is the block diagram I linked to in accordance with your preference?

...If you are going to quote me, then your comments have to refer to what I said in that quote...

Apologies but I can't see where I haven't done that?

...You keep going back to the entire voltage amplifier block with its inverting and non-inverting nature.  I was specifically referring to the LTPI alone and its inverting and non-inverting stages...

I don't understanding how, when referring to the LTP alone, its inputs / stages might be considered inherently inverting or non-inverting; the LTP, as you note, is a differential amplifier, so to do that seems conceptually invalid? As each input and output is inverting with respect to the other. Hence the inverting / non-inverting characteristic / terminology seems only to be meaningful with respect to the power amplifier's (ie the entire amplifier block / feedback loop) output.

The differing use of the inverting / non-inverting terminology and the mis-communication resulting from that, might be at the crux of at least one of the points being debated.
It seems to me that your referring to the circuit node that forms the inverting input of the power amp block as the non-inverting input (of the LTP) makes a fine example of how such mis-communication would be a near inevitability 

[2deaf]

Reply #16 merely offers one explanation for the necessity of reversing the phase of the NFB signal when moving from an AB165 to a SR AB763 topology.  Reply #16 contains no quote, so it is not addressing anything in particular that anyone said.  Reply #16 is not a rebuttal to Reply #15 since Reply #15 and Reply #16 are talking about two different things.  Reply #16 is related to Reply #10 with respect to a need for NFB phase reversal, but again, Reply #10 quotes Reply #8 which is about something different than Reply #16. 

Reply #16 is not talking about a generic LTP differential amplifier.  I'm talking about a LTPI, which is a phase inverter.  And not just any phase inverter, but the ones found in an AB165 and a SR AB763.  It's right there in black and white with no edits.  Since I am talking about these specific phase inverters, the terminology used refers to these specific phase inverters.  A widely used convention is to refer to the stage of these LTPI's that is connected to the preamp signal as the "inverting stage".  The other stage would then be referred to as the "non-inverting" stage.  Although the terms "inverting" and "non-inverting" occur across a wide range of electronics, it is difficult to construe my use of these terms to designate anything other than the two stages of the particular phase inverters under discussion in Reply #16.       

[2deaf]

...

the power amplifier block may be represented -

That seems a little topsy turvey to me.  I prefer my page...

I can't perceive how one might consider one topsy turvey, yet the other not; please explain? I thought I had explained why I thought the text of yours that I quoted was topsy turvey, ie because applying NFB to an amplifier's non-inverting input seems inherently self contradictory.

Your page only appears to deal with the SR AB763 non-inverting amplifier arrangement; how about for the Bassman AB165 inverting amplifier, is the block diagram I linked to in accordance with your preference?

...If you are going to quote me, then your comments have to refer to what I said in that quote...

Apologies but I can't see where I haven't done that?

...You keep going back to the entire voltage amplifier block with its inverting and non-inverting nature.  I was specifically referring to the LTPI alone and its inverting and non-inverting stages...

I don't understanding how, when referring to the LTP alone, its inputs / stages might be considered inherently inverting or non-inverting; the LTP, as you note, is a differential amplifier, so to do that seems conceptually invalid? As each input and output is inverting with respect to the other. Hence the inverting / non-inverting characteristic / terminology seems only to be meaningful with respect to the power amplifier's (ie the entire amplifier block / feedback loop) output.

The differing use of the inverting / non-inverting terminology and the mis-communication resulting from that, might be at the crux of at least one of the points being debated.
It seems to me that your referring to the circuit node that forms the inverting input of the power amp block as the non-inverting input (of the LTP) makes a fine example of how such mis-communication would be a near inevitability 

Ruffled your feathers, eh?

[fredouille]

Hi all, I jump into this topic, found in google

I have such issues with a 6G6B, waterboating + treeble oscilliation when volume over 25%.. the osciloscope becomes crazy
but I have this only with a signal generator a 0,1 V with a guitar, even with a humbucker the amp works perfectly fine even at 50 100% of volume, treeble at 100% (with a power attenuator of course).
I heard that it could an issue with the impedance of the generator ? strange is that I don't remember on a blackface to have the same neither to have this before rebuiling the 6G6B