London Power/Kevin O’Connor Master Volume Circuit

[syscokid]

Continuing with my previous post, yesterday I realized that the dual 1 MEG Alpha pot that came with the kit does not have enough bushing length to clear the front panel enough to add a washer and nut. I’ll be using a dual 500k PEC pot instead, which I have in my stash.

[Helmholtz]

Bootstrapping between Vout and common will increase the effective value of R2.
eg if the bootstrapping doubles the R2 in circuit effective resistance, then
Vout = 0.2Vin.

The bootstrapping makes the whole pot (between its outer ends) look like 2M to the input signal before the pot.
Why should it only alter the lower resistance?

[pdf64]

Bootstrapping between Vout and common will increase the effective value of R2.
eg if the bootstrapping doubles the R2 in circuit effective resistance, then
Vout = 0.2Vin.

The bootstrapping makes the whole pot (between its outer ends) look like 2M to the input signal before the pot.
Why should it only alter the lower resistance?

As I see it, the bootstrapping can only affect the resistance between grid and cathode's common point, ie R2?
Anything preceding that seems to be outside the (local LTP) feedback loop.

[Helmholtz]

As I see it, the bootstrapping can only affect the resistance between grid and cathode's common point, ie R2?
Anything preceding that seems to be outside the (local LTP) feedback loop.

The bootstrapping happens between the 2 voltage sources: input signal and tail signal.
Grid signal is not input signal.
When the pot is rolled back, both grid signal and tail signal reduce, thus decreasing the bootstrapping and input impedance.
But pot taper should stay the same.

[martin manning]

The thing I keep wondering about is the pot on the grounded grid side is dialing down the feedback signal reaching that grid. Does that mean that the feedback signal relative to the input signal is being doubly reduced?

[Helmholtz]

The thing I keep wondering about is the pot on the grounded grid side is dialing down the feedback signal reaching that grid.

Not sure what you mean. The NFB signal is not meant reach the input grid anyway.

[pdf64]

Grid signal is not input signal.

Yikes, that's left me bewildered!
Can you explain / provide a link?

[martin manning]

The thing I keep wondering about is the pot on the grounded grid side is dialing down the feedback signal reaching that grid.

Not sure what you mean. The NFB signal is not meant reach the input grid anyway.

LTP has three inputs, two grids and the tail. The FB signal is applied at the tail and to the second grid through a cap. When this MV is turned down, one of the pots directs the FB going to the second grid to the tail, which is a common mode input.

[Helmholtz]

LTP has three inputs, two grids and the tail. The FB signal is applied at the tail and to the second grid through a cap. When this MV is turned down, one of the pots directs the FB going to the second grid to the tail, which is a common mode input.

Ok.
(Referring to the LP-MV implementation in the JCM 800 circuit of post #1.)
When the "A" pot is in the upper position, the input signal to the first triode is zero.
Without an input signal, also the NFB signal must be zero.

Generally only a few percent of the NFB signal injected to the bottom of the tail reach the cathodes, because of the low cathodes' input impedance causing heavy voltage dividing.
I think the NFB path via the tail is rather insignificant.
It does somewhat bootstrap the tail above the injection point and so slightly improves PI balance, though.

[LOUDthud]

Feedback to be bottom of the tail causes a nearly constant current to flow through the tail resistor in the unmodified circuit. This improves balance at the outputs of the PI except for the imbalanced plate resistors. As the pot on the feedback side reduces the feedback, the signal at the bottom of the tail doesn't change in relation to the output. This is going to screw with balance.

Too many things to try to figure out what it might sound like. This is a job for a modeling or simulation program.

[Helmholtz]

This is going to screw with balance.

The imbalance of a typical LTPI with equal plate resistors and balanced triodes is around 8 % without NFB.
So the difference can't be big.

https://www.ampbooks.com/mobile/amplifi ... alculator/

[Helmholtz]

Yikes, that's left me bewildered!
Can you explain / provide a link?

Somehow my earlier response to this post got lost.
Lemme try again.
There are actually 2 interactive (?) bootstrap paths.
One is between the outer input source (that's what I call the input signal) and the tail signal.
The other is between the grid signal and the tail signal.
Bootstrapping depends on the gain between the 2 points involved.
The gain between grid and tail is fixed. Gain between input and tail varies with MV setting.
Maybe that's all and means you're right with pot taper changing.
Originally I had only looked at the outer bootstrap path, where the resistance between grid and tail would be increased proportionally the total pot resistance.
I'm still not quite sure and might calculate some examples.

[LOUDthud]

https://www.ampbooks.com/mobile/amplifi ... alculator/

That calculator doesn't take into account the effect of feedback to the bottom of the tail resistor.

[Helmholtz]

That calculator doesn't take into account the effect of feedback to the bottom of the tail resistor.

Yes, but it shows the gain difference of only 8% with equal plate resistors and 6% with the typical 82k/100k pair without NFB.
So there's not really much to improve with a 6% or 0.5dB imbalance.
After all it's a guitar amp.

BTW, NFB improves balance even without feeding to the bottom of the tail.

[Helmholtz]

@pdf64:
I came to the conclusion that a 10% log pot should act like a 20% log pot at low MV settings, gradually morphing into a 10% taper with increasing the MV.
So you were right and I was wrong.
My considerations don't take into account additional effects caused by simultaneously varying the NFB level.