Gain control in local NFB instead of shunt to ground?

[thelonious]

I have been learning as much as I can about pedals/solid state stuff recently, since I knew so little about it before. I've noticed that with overdrive pedals that use op amps set up as inverting amplifiers, the gain pot is often in the feedback loop of the op amp instead of between stages.

Now... I know that op amps and tubes are very different animals. But are there tube amps that put a gain control in a local NFB loop like that? And when we design amps, why do we usually make the gain pot a voltage divider between stages instead of putting it in a local NFB loop?

I'm sure there are good reasons; just trying to learn!

[tubeswell]

Merlin's triode-pentode morph control is a type of local NFB gain control for pentodes (where the screen is variably AC-tied to the plate).

Don't know how many example of straight NFB gain controls there are in triode stages but lots of amps use NFB for tone controls (resonance, presence etc)

Here's a link to a Baxandall article on various NFB tone controls: http://thermionic.info/baxandall/Baxandall_NegativeFeedbackTone.pdf


This quirky circuit utilises a Global NFB 'tone-compensation' network

[HotBluePlates]

And when we design amps, why do we usually make the gain pot a voltage divider between stages instead of putting it in a local NFB loop?

Because most preamp tubes have a gain from 10-120, unlike transistors and opamps having gain in the thousands-to-millions.

But guitar amps do have something similar in the way of presence controls (or a NFB control if you swap a resistor for the cap in the presence circuit).

You can do something which uses a gain pot inside of a NFB loop to control a stage's gain, but the low end of the gain may be too low, and the high end not high enough. An audio pot has a range of maybe 100:1 from the volume max'd down to a setting of 1. You can't match that with a triode, but you might be able to have something similar with a pentode and a NFB loop of some kind.

I'm guessing that if you want the effect to sound like "more gain" instead of "more volume" that you'd need some kind of volume control after the loop turning the output down at the same time the NFB pot is turning the gain up. Or maybe that just defeats the point... I don't know.

[PRR]

> put a gain control in a local NFB loop

Your drawing does not show NFB gain control.

> I'm sure there are good reasons

The best: "we've always done it this way!" (Variable-loss volume control pot.)

The volume pot works good.

A possible disadvantage of variable NFB (in a technology which, as HBP says, never has "large" NFB) is that the distortion level and flavor will change when you only want to change gain.

[thelonious]

Thanks, all. That is great info. Definitely going to study that Baxandall PDF; I've been wanting to play around with NFB tone control ideas for a while but haven't done enough research yet. And PRR - thanks for catching that error. I should have drawn the pot from plate to grid, not cathode to grid. That's what I get for sketching something quickly at work when I should be working instead.

[PRR]

> I should have drawn the pot from plate to grid, not cathode to grid.

Is _that_ what you meant?

In that cases, it is incomplete without an impedance (often resistance) in -series- with the input.

Not making fun of you... but confusions like this may be another reason the Old Men didn't do it that way. That they tried, without mature understanding, got it somehow wrong, it didn't work as expected, they went back to the Old Ways.