Local NFB in preamp

[Dagan8]

Hello

I was looking at schematic that was posted in another forum. It looks to have some sort of negative feedback loop. I think. I have attached a picture and outlined the area in red. Is this some sort of NFB? Is the amount of FB dictated partially by the series 1M resistor as well as the split cathode resistance ratio?

-Thanks

[soundguruman]

That's positive feedback, it's going to increase the gain in some region, but the circuit is designed kind of wacko.
The thing about schematics like this, that are posted in forums, is that they may not be so great an idea. It's OK o try it but I'm not certain of the results.

[Steve Conner]

Yep, definitely something weird about that circuit.

[soundguruman]

It kind of reminds me of a variation of the boogie positive feedback circuit, that was stolen from "how to hotrod fender amplifiers" book, originally designed by Dan Armstrong. IMHO, it was Armstrong that designed the circuit, and Randell copied it, and claimed he invented it, which of course he did not.
The input signal is out of phase from the plate of V1a, then back in phase at the plate of V1B. This in phase signal is fed back to the V1A cathode, and causes a gain boost, but not very much of one....circuits like this are kind of squirrely, and you usually have to use two tubes instead of one to reduce the oscillation....

[Merlinb]

Actually it is negative feedback (the two stages are inverting, remember), but the feedback factor is so small (<0.0009) that I doubt it has any significant effect on the circuit.

[Dagan8]

Actually it is negative feedback (the two stages are inverting, remember), but the feedback factor is so small (<0.0009) that I doubt it has any significant effect on the circuit.

Thanks. Other amp designs from the same designer(hint: D-Style) have a different NFB design were he takes the output from the second stage and feeds it back into the input of the same stage through a large 40M resistance. This would also have a very tiny feedback factor. I'm curious why he seems to always use some sort of tiny NFB on the preamp stage.

[Gingertube]

With your second example (40M from anode back to grid of the same stage) the amount of feedback will depend upon the grid leak resistor. Should it be the "typical" 1M then the gain of the stage will be controlled back to 40. Given a free run 12AX7 triode stage with full cathode bypassing will normally give you a gain of around 52 to 55 then that is a gain reduction of 20%. There will be a corresponding decrease in output impedance from the anode of 20% which may better drive a tone stack or similar.

That is probably what was intended with your first example as well. MerlinB is definitely right about it being NEGATIVE feedback. The feedback signal applied at the cathode is in phase with the grid signal and therefore SUBTRACTS from it. A corresponding reduction in output impedance from that first tube results, intended (most likely) to better drive the tone circuits. It was just very badly done. You often see this sort of feedback connection in HiFi Power Amps where the feedback connects from output tube anode to driver tube cathode. In that case it is designed to better drive the Miller Capacitance at the output tube grid.

So why do it? Driving tone stacks from a lower impedance gives them a more predictable response. This is why Marshall used cathode follower drive of their tone stacks.
Having said that many people prefer the interactive nature of the tone stack driven straight off the anode as Fender did. That (additional interaction) is due to the tone stack settings actually modifiy the load on the driving tube and so you get a level of interaction in addition to the B.M.T controls interacting with each other.

When I first started playing guitar through a valve amp with the usual B,M T controls I had some difficulty learning to drive that tone stack, until I learned enough to treat the Treble control like a volume control and recognize that it also determines how much signal is passed to the bass and mid controls, high treble setting => limited bass and mid control, low treble setting => more bass and mid control.

One other thing to remember when designing these sort of feedback circuits is that you MUST NOT let the stage(s) within the feedback loop saturate. The feedback mechanism relies on the gain within the loop. When a stage saturates it doesn't have any gain and the whole thing goes to "custard".

A bit of a "rave" - hope there is something useful in it.

Cheers,
Ian

[mooreamps]

I'm curious why he seems to always use some sort of tiny NFB on the preamp stage.

Now that would be a good question.... What would be the properties of NFB on an amplifier gain stage ??? hmph ???
Clearly the "resident magician" has no clue.....

-g

[soundguruman]

Now that would be a good question.... What would be the properties of NFB on an amplifier gain stage ??? hmph ???
Clearly the "resident magician" has no clue.....

-g

I continue to insist that this is positive feedback, and it will increase the gain, although there's not a heck of a lot of it.
When the feedback is in phase with the input, it will boost the gain, not reduce it. To reduce it, it needs to be out of phase.
NFB on a stage? It will cause an attenuation of the signal, depending on what NFB frequencies are fed back into the input.
If it was NFB, it would probably be intended to prevent oscillations or ringing of the preamp tube...but I don't see it as such.
Resident Magician? who's that?

[Gingertube]

Sorry Mate,
The "resident magician" and I are correct in saying that it is NEGATIVE feedback.

positive going at V1A grid => negative going at V1a anode.
negative going at V1a anode => negative going at V1b grid.
negative going at V1b grid => positive going at V1b anode.
positive going at V1b anode => positive going at v1a cathode (the feedback signal).

Input to V1a is the grid volts minus the cathode volts. Both are the same phase - hence the signal to V1a is REDUCED, overall gain is reduced, that is it is negative feedback.
the "trick" here is to recognize that the drive signal for V1a is grid volts minus cathode volts.

Of-course with that "pissy" feedback factor who will ever notice.

Cheers,
Ian

[Chuck H]

positive going at V1b anode => positive going at v1a cathode

Yup... And the anode of V1a is out of phase with the cathode of V1a. So the cathode of V1a is positive. The AC being fed back is positive. It is a positive feedback loop.

"Grid volts minus cathode volts". Yup. But if that in itself qualifies a gain reduction then any kind of feedback, positive or negative, should reduce gain. Not the case. Because of the phase.

It's a positive feedback loop.

EDIT: Gary, don't be small about this. Have some grace and dignity.

[Dagan8]

Thanks for the open discussion guys. I am learning a lot reading through it.

Although there is not much to learn from this comment:

Now that would be a good question.... What would be the properties of NFB on an amplifier gain stage ??? hmph ???
Clearly the "resident magician" has no clue.....
-g

Cheers

[teemuk]

It's positive feedback in phase sense but inserted to a cathode like that it will still decrease gain because it adds on to the cathode's own "degenerative feedback". In fact, inserting negative feedback to a cathode like that even triggers oscillation in some amps. If you would want to make that thing works as positive feedback in the usual sense of the term you'd insert the feedback to the grid.

I agree on the note that the circuit doesn't seem overly sensible. The whole tonestack and even the volume control are wrapped inside the feedback loop so the feedback will basically try to linearize (read: minimize) their effects. The good thing considering such a botched design is that the dividance ratio of the feedback is 1M / 100 (so practically you can consider it about zero) and therefore the feedback loop basically has no mentionable effect.

[Steve Conner]

The original poster implied that this is a D****e circuit. The Rock/Jazz switch certainly bears that out.

Supposedly the 40M resistor in the original circuit was meant to add "compression", and I think it did this by messing with the DC conditions of the stage, adding a long time constant.

But the circuit under discussion here has AC coupled feedback, so it can't affect the DC conditions. It must do something different. Or nothing at all

Bearing in mind the mystique around D****e amps, it's quite possible the schematic could be wrong, either accidentally or deliberately.

[Chuck H]

I'm not sure if it's adding degenerative or additive feedback. The signal being fed back isn't a mirror image of whats on the V1a plate. There are phase shifts within the loop. In that case the cathode is acting like an input to some frequencies when phase alignment is right and possibly degenerative on other frequencies. Depending on factors like gain and control settings.