Did some bassmans have POSITIVE feedback?

[ELS]

I was researching the bassman when I thought I didn't really have it sounding how I thought it needed to sound.
and realized, in some amps the plate wires on the output tubes are connected backwards, but the output connections stay the same.
particularly in Johan Segeborn's video:


here's some screenshots:


You can see that the brown wire connects to the top tube (from this perspective) and blue wire to the bottom tube.
and the grid wires dont seem to be flipped because you can see that the red wire goes to the bottom tube, and the 'not red' wire goes to the top.

here's the layout:

you see that the only thing that seems to be backwards is the plate wires, which would make the amp have positive feedback instead of negative.

and the output connections arent flipped either, you can see here:


they are correct.
other than that the amp is exactly to spec (to the reissue spec, not the original schematic which isnt correct for most bassmans)

Johans video that is here:


at 1:30 you can see the output speaker connections and right after the output connections.


now, yes the amp is much more unstable with positive feedback, when I connected the amp to a speaker it oscillated badly, only could play it with prescence on max (which would reduce positive feedback in higher frequencies), and even then it oscillated at low frequencies sometimes.
But, connecting the amp to a dummy load, it worked fine.
I started wondering maybe my speaker is mostly to blame, since 1. it's 12 ohms, not 8 (dummy load is 8 as it should be), and 2. it's not the same speaker as the bassman.

so I used amplitube's speaker/cabinet sims after the dummy load and recorded some samples (excuse my guitar playing )


With positive feedback it definetly has more gain, as you'd expect. but idk if it sounded better, in certain ways I though it did, but I felt it lacked the bassman's "thonkyness", didn't have the bite. when I played the riff for the negative feedback configuration I couldn't stop playing even (so many hours of riffin' I cut out)
but with positive feedback I didn't feel the same addicting force.
positive feedback did feel 'lighter' on the playing tho, while negative felt tougher to play with.

idk, what do you think about this?


I also researched specs on the output transformer, but the best I could find is DC resistance readings,
49 ohms to blue, 42 to red (or brown most likely)


if the amp was wired according to the layout, the higher resistance winding would be connected to the 100k plate resistor side of the phase inverter
and other to the input side (82k plate resistor)

but ofc if the plate wires are reversed, it would be the other way around.
this in theory should form mismatched frequency response from output tube to output tube, especially after you consider that the 100k side of the PI has a hotter output signal...

but I don't haven't tested what impact does this have, but wanted to point that out anyway.



my setup was:
Les Paul with EMG F-H2A bridge pickup, straight into bassman-esc amplifier, with the preamp more like a blackface fender.
into an 8 ohm dummy load, Bass on 2, Middle on 8, Treble and Volume all max (10)

 

[StingRay85]

It's all about the phase of the signal.

I don't think any amp can sound good with positive feedback, unless you like a squealing noise.

 

[Pete Farrington]

To be fair, positive feedback per se is used in several common guitar amp stages, eg cathode follower, long tail pair, cathodyne. And by being returned to the ‘tail’ as well as the ‘other’ grid, almost every global loop around the power amp includes an element of it.
Positive feedback doesn’t necessarily equate to oscillation or even instability.
But yeah, flipping OT polarity in an amp with global NFB is going to result in a weird sound, probable instability, if not outright free running oscillation.
My guess is that the OT in that amp may be an oddball, eg a manufacturing error in either the primary or secondary wiring colour codes. Which Fender corrected for when they built the amp.
A speaker has massive resonance impedance peaks / phase shifts, so compared to a resistive load, they’re bound to have a greater tendency to induce instability.

 

[XTRXTR]

If its not what Pete suggested I think you have the potential for high freq oscillation out of hearing range with POS feedback. Put a scope on the output to see. If the tubes are cool biased they may not red plate in that condition but likely lots of odd order harmonics. As you said is some cases of dialing in some tones you might find something you can't get otherwise but it could be prone to harsh and fatigue the ears quickly. Your low volume cleans may be edged with super highs but as you said not giving that feel you can play hours with.

Just intuitizing

 

[ELS]

It's all about the phase of the signal.

I don't think any amp can sound good with positive feedback, unless you like a squealing noise.

I said this already... when connected to a constant 8 ohm load (yeah it will have some UHF affecting inductance and capacitance) but it doesn't oscillate, it works fine.
yes real work speakers aren't a constant 8 ohm load, far from it. but it's not far fetched to consider that maybe some speakers will cope with a positive feedback amp.
even a random 12 ohm speaker I had connected was pretty usable, albeit with presence on max.

To be fair, positive feedback per se is used in several common guitar amp stages, eg cathode follower, long tail pair, cathodyne. And by being returned to the ‘tail’ as well as the ‘other’ grid, almost every global loop around the power amp includes an element of it.
Positive feedback doesn’t necessarily equate to oscillation or even instability.
But yeah, flipping OT polarity in an amp with global NFB is going to result in a weird sound, probable instability, if not outright free running oscillation.
My guess is that the OT in that amp may be an oddball, eg a manufacturing error in either the primary or secondary wiring colour codes. Which Fender corrected for when they built the amp.
A speaker has massive resonance impedance peaks / phase shifts, so compared to a resistive load, they’re bound to have a greater tendency to induce instability.

it sounds like your describing bootstrapping, getting voltage gain same as MU of tube.
but yeah positive feedback is also used in some hifi amps that I've seen, when done right it is stable.

yeah could be just a manufacturing error, but I doubt a transformer manufacturer would make such a trivial mistake.
one likely posibility is that the grid wires going from the PI under the board were connected backwards, and afterwards the factory reversed the plate wires to account for this.

If its not what Pete suggested I think you have the potential for high freq oscillation out of hearing range with POS feedback. Put a scope on the output to see. If the tubes are cool biased they may not red plate in that condition but likely lots of odd order harmonics. As you said is some cases of dialing in some tones you might find something you can't get otherwise but it could be prone to harsh and fatigue the ears quickly. Your low volume cleans may be edged with super highs but as you said not giving that feel you can play hours with.

Just intuitizing

the amp isnt oscillating.

I think that 'feel' was just from a nicer response in the output section, was tighter than with positive feedback.

 

[Pete Farrington]

it sounds like your describing bootstrapping

Yes, bootstrapping, which is an application of positive feedback.
Also when used in a typical global NFB loop, consider that the LTP grids carry signals of the same polarity. But as the NFB signal is also applied to the tail, that must also be a positive feedback loop. As with bootstrapped grid leaks, it acts to make the effective value of the tail resistor larger, thereby improving balance.

getting voltage gain same as MU of tube.

Dunno about that, could you explain your thinking?

 

[ELS]

Yes, bootstrapping, which is an application of positive feedback.
Also when used in a typical global NFB loop, consider that the LTP grids carry signals of the same polarity. But as the NFB signal is also applied to the tail, that must also be a positive feedback loop. As with bootstrapped grid leaks, it acts to make the effective value of the tail resistor larger, thereby improving balance.

Dunno about that, could you explain your thinking?

no bootstrapping isn't really adding positive feedback, bootsrapping can be achieved with a constant current source in place of the plate load resistor.
it isnt positive feedback, it's unity feedback. which then will give you a voltage gain the same as the mu of the tube, for 12AX7 that would be 100.
the thing your describing for LTP is bootstrapping too, the tube acts like a buffer, which then is fed back into the grid voltage supply, tube will pull it's own shoe laces, but same as what would happen when you try to pull your own shoe laces in the real world: you wont start to fly.
to 'fly' you would have to add positive feedback, more than unity feedback, with that you can get a lot of gain, but that can turn unstable real quick.
although when you mention the NFBL affecting the input impeadance of a LTP PI... yeah I guess it would, have to thing about that..

bootstrapping, it's really useful for a lot of things, I used bootstrapping in a tube microphone and it has more than enough gain. you can also boostrap by adding AC feedback from a cathode follower back into the plate supply of the previous stage driving it, which is what I did in the mic.

it works like this:
the tube wants to pull the signal low trough the plate load resistor, but then the plate supply voltage gets pulled down too because of the cathode follower being AC coupled to it.
so it's like having near infinite AC impeadance for the plate load, thus the voltage gain is nearly same as MU of the tube (you ofc have losses that reduce the gain a bit, so it wont be exactly the MU)

 

[william vogel]

I think the point here is the plate wires being reversed from what’s seen in the typical layout. I’d suggest reversing the wires to confirm that the wiring although looking backwards is correct. If the amp goes crazy with the symptom of reversed primary leads then we know that the transformer was built with incorrect wiring or the phase inverter plate wires are reversed. I’ve never built an amp that would have stability with NFB and reversed wiring but I’ve only mistakenly done it a few times.

 

[Browneyesound]

I made a single ended EL34 home brew amp that initially had the feedback tap as positive feedback. It didn’t have wildly unstable oscillations, but the thing I noticed was of was the bass was very boomy. Not a pleasant bass boost, but quite an annoying boomyness.

 

[ELS]

I made a single ended EL34 home brew amp that initially had the feedback tap as positive feedback. It didn’t have wildly unstable oscillations, but the thing I noticed was of was the bass was very boomy. Not a pleasant bass boost, but quite an annoying boomyness.

well that could possibly be the amp oscillating, oscillations can often be ultra-sonic, meaning that they're too high in frequency to be heard by humans. but those oscillations can cause weird frequency modulations, as well as increase filter hum because of the high current draw.
boomy bass is a big red flag in tube amp design since it's one of the most frequent symptoms of ultra-sonic oscillations.
best way to know that it is happening without a scope would be with a light bulb in series with the amplifier, like a 250w bulb for a 50w amplifier should be good, at idle the bulb should be completely dark, when cranked it should shine slightly, but not super bright, if it's shining pretty bright all the time, and usually dims down when putting a signal trough the amp, that's a sure sign of oscillations.

I haven't messed with many single ended amps, I know that SE amps draw less current the more signal you put trough them, so it probably isn't a good way to test that for an SE amp.

 

[Pete Farrington]

SE amps draw less current the more signal you put trough them

You may be getting anode current mixed up with anode dissipation?
Anode current in class A should remain fairly constant from idle to full unclipped output (though g2 current will increase).
But anode dissipation reduces as power output increases, due to power that would be getting dissipated by the anode being delivered to the load.

As noted previously, positive feedback can exist without oscillation (free running or parasitic) occurring.