NFB applied directly to Rtail

[EtherealWidow]

So I know that this isn't normally done, but is there a reason? All I have on hand is a 1M pot. Amp in question is a Spitfire (47k tail). Any reason this shouldn't be done?

[EtherealWidow]

Also, I apologize in advance if this question is rooted in ignorance. I still have a lot I'm currently reading on NFB. Poles and zeros and whatnot.

[JMFahey]

Post schematic

[martin manning]

You need to develop some NFB signal voltage at the non-inverting triode's grid, so you need to add a resistor between the junction of the 47k and the 0.01uF cap and ground.

Do you have the 22k-3k3 line out off of the 8 ohm output? If so you could lift the ground from the 47k-0.01uF junction and connect it to the line-out for a quick listen to what NFB will do for you. You could put a pot across the line out to vary the NFB, and/or add a cap for some presence effect.

Be aware that the phase of the output signal may be wrong for NFB applied in this way, so you may have to swap the OT primary leads.

[EtherealWidow]

Martin, is that why they always have the much smaller resistor "below" the tail? That would make a lot more sense now.

[martin manning]

Yes, that's right.

[EtherealWidow]

So is NFB more of a current signal, which is why you want it to flow across the large resistance of the tail?

[martin manning]

It's more about the AC voltage that appears at the second triode's grid, but you need some current and resistance to develop that voltage. In fact the small resistor might just as well be inserted between the cap connected to the second grid and ground. The FB resistor would go from the speaker output to the top of the small resistor as before for voltage division, leaving the tail resistor out of it altogether.

[EtherealWidow]

So the NFB signal appears at the second triode's grid? How is this possible if the grid is grounded at AC?

[martin manning]

In the usual LTP with GNFB, the second triode's grid is AC grounded though a cap and the small resistor at the bottom of the tail. If an AC voltage is developed across that small resistor, it will appear on the grid.

[EtherealWidow]

Hey Martin. Thanks for the reply and sorry for the delay. To be honest, GNFB is still confusing for me and I've been trying to research it more, learning about the function of the tail resistor and how it emulates a CSS. Also trying to understand how adding a single signal to the cathodes of both triodes doesn't boost the gain of one and subtract from the other.

[pdf64]

Also trying to understand how adding a single signal to the cathodes of both triodes doesn't boost the gain of one and subtract from the other

Yes, a tiny bit of the feedback signal probably does end up at the cathodes of the regular LTP, but I think it's probably insignificant / greatly attenuated compared to the feedback signal level at the 'other' grid.
Seen the Aiken and Merlin pages on the LTP?