Is this NFB in a common cathode gain stage?

[Willabe]

I got this from this thread, here's the link. Full schemo can be found there.

http://www.el34world.com/Forum/index.php?topic=15906.0

On V2b the grid leak/return R is standing on V2a cathode R.  

Question is;

1. Is this NFB?

2. Does it cause compression/limiting in V2b?

(The more signal that V2a's plate generates the more signal it's K generates and is fed to V2b's grid. K is 180 degrees out of phase with it's plate. So K signal fed to V2b's grid phase cancels with the signal taken from V2a's plate? The more signal the more it cancels, so automatic volume/limiting = compression?)

3. Or are they just changing V2b's bias point?

4. Or is it both?

I don't recall seeing a common cathode gain stage wired up this way before.


                     Brad        

Edit; Sorry some how V2b moved in editing the drawing, should be to the right of V2a.

[John]

I am looking at it as a simple attenuator, IOW just look at those 470K resistors as a 1M pot turned halfway up, with the grid coming off the wiper.

[DummyLoad]

local NFB, grid leak for V2.2, ~950K load for V2.1, and signal divider.

--pete

[Willabe]

Thanks Pete.

But does it introduce any compression/limiting effects?


                       Brad   
 

[John]

But does it introduce any compression

Well sure, if you push down on it. 

[tubenit]

I just see an ordinary gain stage & no NFB.   Like John said ....... sort of a 1M pot turned to 470k/470k ground.

It's just drawn weird. 

(my new favorite excuse) 

With respect, Tubenit

[tubeswell]

I just see an ordinary gain stage & no NFB.   Like John said ....... sort of a 1M pot turned to 470k/470k ground.

It's just drawn weird. 

(my new favorite excuse) 

With respect, Tubenit

The two 470k in series makes 940k (1M) that is AC-tapped off the plate of the stage and fed back to the un-bypassed cathode of the same stage (where the signal is out of phase with the plate). As it happens, just the fact that the cathode is un-bypassed will be adding cathode current feedback into the stage. The plate-to-cathode local NFB dampens the effect of the cathode current feedback somewhat.

[Willabe]

The two 470k in series makes 940k (1M) that is AC-tapped off the plate of the stage and fed back to the un-bypassed cathode of the same stage (where the signal is out of phase with the plate).

Ok, I missed that before when DL said "local FB" it just struck me now. The local FB is from plate to cathode on V2a.

But the 2x470K junction is still tied to the grid of V2b. They saved a coupling cap doing it that way but did they wire it up that way for another reason too?

I was thinking (AC) NFB injected into the grid (along with the AC plate signal of V2a) of V2b because V2a's cathode is un-bypassed.

They should cancel each other out somewhat. The more the plate signal developed then more cathode signal developed and that's gonna act like a compression circuit.

Well depending on how much the plate signal swamps the cathode signal. The more the plate swamps the cathode the less the compression effect down to not happing at all.

As it happens, just the fact that the cathode is un-bypassed will be adding cathode current feedback into the stage.

Which stage, V2a or V2b?

The plate-to-cathode local NFB dampens the effect of the cathode current feedback somewhat.

So the plate-to-cathode local NFB dampens the effect of the cathode current feedback somewhat, to V2b?

Why "current" FB and not voltage FB? Since the cathode of V2a is un-bypassed there's AC signal voltage. Or is it both at low a Z?

And isn't the bias point of V2b changed a little by standing V2b's grid leak on top of V1a's K R? That K R has got to develop a few acv.   



                      Brad     

[Willabe]

Like John said ....... sort of a 1M pot turned to 470k/470k ground.

But that's the point, it's not going to ground directly.

It's taped into/from the cathode which is 180 degrees out of phase with it's own plate.


               Brad         

 

[tubenit]

There are probably different ways to approach local negative feedback, I'd guess?   

This is what I typically think of in that regard.   From plate to grid on a 12AX7

With respect, Tubenit

[Willabe]

There are probably different ways to approach local negative feedback, I'd guess?  

Yes I think you're right.

I'm wondering why they didn't wire it up like you just showed? It saves them a coupling cap but that's it.

I still think it performs a little different because it's also tied to the next stages grid, so it effects V2a and V2b. Now how much difference this makes?      

Maybe the engineer who designed the amp did it just for the sake of doing something a little different?    

KOC has said that putting a small value cap across the plate R or running it from plate to K makes no difference in effect/tone.

But again this is different because it's tied to the next stages grid, not just it's own K or it's own grid.    



                              Brad        

[PRR]

If it is NFB, then we have 940K:4.7K or 200:1 loss, around a stage with gain less than 50.

That's essentially zero NFB.

I think it is an accident that worked. The second tube gets a DC bias which increases its current and shifts its plate voltage lower. How much? How does it sound? I don't know, it's not obvious, which is why I suspect it "just happened" on the prototype and happened to work satisfactorily.

Yes, it is also an attenuator. You often do this between high-gain stages which are intended to get clipped. The clipping action is even harder to predict by thumb.

[Willabe]

That's essentially zero NFB.

I think it is an accident that worked.

Ok, thanks PRR, round file can then.    

Like my favorite scene from Rudolf the Red Nose Reindeer where Yukon Cornelius lookin for gold, yells "Wahoo" then throws his pick axe in the air and when it comes back down after sticking in the snow on the ground, he takes it out, sniffs it, licks/tastes it and tastes it and says.............        "Nothing".

http://youtu.be/tFxzohbXLSo  


                   Brad