This is what I did on my SDF AA 764 champ to reduce the NFB loop .

[catnine]

I placed a 100k ohm linear pot in series with the normal 2.7k NFB resister. There is a 47 ohm resistor to ground so i didn't change that because it seems part of the bias for the second gain stage of the 12ax7 tube. I have the stock 2.7k ohm resister on a dpdt slide switch so I can dial,in more resistance than the stock 2.7k NFB resistor from the OT secondary in one position and if I want the stock 2.7k I flip, the switch and it bypasses the 100k pot and is a direct connection to the 2.7k ohm NFB on the OT secondary. Yet I really hear a difference was this ths way to do this mod? In the title I said SDF champ I meant SF AA764,Champ build.

I can't seem to find out where I found this mod yet it seems to work . Can anyone tell me if this is right? It seems the 2.7 k NFB and the 47 ohm to ground are in parallel because the OT secondary to ground makes the circuit parallel .

What I was trying to do was add more resistance to the stock 2.7k NFB resistor so in affect to make the amp more like one without a NBF at all.

I didn't want to just have a switch to disconnect the NFB I wanted it adjustable and as far as I can tell when the slide switch is adding the 100k linaer pot to the 2.7k in series there is a definate differance in the way the champ sounds , it is loser and more open with 100 k added to the 2.7k .

What I did seems to work but I still can't figure out why .

Can anyone explain this too me so it makes sense?

[Enzo]

OK, my thoughts on this.

If you like the results, then it is a good mod.

Adding resistance in series with the 2.7k will reduce the amount of NFB. Making it variable will allow you to adjust the amount of NFB. There really isn't a need for a switch, turning the 100k pot to zero is the same as shorting across it with a switch. Of course with your switch, you can leave the pot set one way and just flip that on and off. Up to you.

100k seems large to me. As you start at the zero ohms end and turn the pot upwards, is there a point, maybe half way along, where any further turning makes little differnce? In other words maybe only the first half of the adjustment range does anything? If the whole turning range makes useful adjustment, then leave it. However, if the pot only seems to make any changes up to a point, set it at that last point, then disconnect it and measure how much resistance that is. For my example, lets say that up to 50k, the pot makes a nice steady adjustment, but after 50k it just sounds the same. That means only the first 50k matters, and you could replace the 100k with a 50k. WHy do that? Because if the 100k is only half useful, all your adjustment range is crammed into half the pot travel. By making the pot 50k, then the WHOLE pot travel is useful adjustment range. Kinda like an amp you can't turn over 4 because it is too loud and all your adjustments come out just a touch over 1. Woudln't it be cooler if all the way up amounted to the 4, then it would be so much easier to adjust for half that.

Now for the 47 ohm. Yes it is involved with the bias of that tube, but just barely. The tube current passing through that 1500 ohm and the 47 ohm bias the tube. But that 47 ohms is less that 1% of the 1500 ohms. In other words if you got rid of the 47 ohm entirely, you'd never hear the difference. (ignoring the NFB effect for the moment) The 47 ohm is not really there to provide any bias.

The 47 ohm and the 2.7k are not in parallel, they seem like it because of the low DC resistance of the output transformer secondary. The DC resistance of that winding is not what the circuit is about. The whole point of this, or any, amp is to put a voltage on the speaker to make it produce sound. Negative feedback takes this voltage and sends a sample of it back to the circuit. This is a low power amp, so the voltage will be low ther as well, but just to make the math easy (don't worry), let us say the output is 10v of signal at full blast. Probably less than that, but I am making the example up. 10v there at the output. That 10v appears across the OT winding regardless of its DC resistance. Now that 10v goes to the speaker, but it also goes to that 2.7k resistor. The 2.7k and the 47 ohm form a voltage divider. Since the total resistance ther is 2747 ohms, at the point where the two resistors join, there will be 47/2747 times the 10v. That is about 0.17v. We are injecting 0.17v of signal into the cathode of the preamp tube. And because of the number of stages and configuration, that voltage is out of phase with the voltages already in that cathode leg. That is why it is negative feedback. SO whatever positive change in voltage through that cathode will be diminished by that small NFB voltage.

SO what if we raise the 2.7k? What if we only add 2000 ohms. That makes 4700 total plus the 47, or 4747. So now the voltage at the cathode junction is 47/4747 x 10v, or about 0.10v. Our NFB voltage has dropped from 0.17 to 0.10. Some more examples:
10k --- 10k+2.7k = 12,700. 12,700+47=12747, so 47/12747 x 10 = 0.036v
50k --- 50k+2.7k = 52,700. 52,700+47=52747, so 47/52747 x 10 = 0.009v

SO the NFB gets smaller as the resistor grows.

It really isn't about the resistor, it is about the amount of signal being returned out of phase to the preamp. IIf we replaced the 47 and 2700 with 94 and 5400, the exact same ratio would exist.
47/2747 x 10 = 0.17
94/5494 x 10 = 0.17
SO if we used them instead, the NFB would be exactly the same. Because the ration of resistors stayed the same, so does the ratio of voltages. When we change the ratio of resistors, we change the ratio of voltages.

Since that NFB voltage is out of phase, it bucks the signal voltage already there. That is why it reduces the volume some. Without the NFB, the sound should be a little louder, but also a little less clean, a little more raw. SO the more you reduce the NFB the more it gets like no NFB at all.

[catnine]

I was told on another site that the 2.7k NFB resister was in parallel with the 47 ohm simple because each end of the 2.7k ended up grounded yet this didn't sense because the 2.7k is wired in series with the 47ohm to ground and the OT secondary produces a a/c voltage to drive the speaker . I knew it was voltage for the OT sent back through a resister out of phase with the second gain stage of the 12ax7 . Now if I measure across the 2.7k resister I get 46 ohms when I switch the slide swich to bypass the 100k pot , if I switch to the pot and dial the pot up full then I read 2.7k ohms . this is the part that baffles me. If I recall if I disconnected the 2.7k from the OT tap then I could read the 2.7k value with an ohn meter. this was well before I installed the pot. Is it that the 2.7k has a low resistance to ground through the OT that I can't read the 2.7k ?

[Enzo]

Yes, to your meter, the transformer winding disappears, it has a VERY low DC resistance. DC resistance is what your meter measures. SO to your meter, the 2.7k is parallel to the 47 ohm. But the circuit doesn;'t work like DC resistance. It works on signal voltages. One always has to be aware of false current paths when taking in-circuit resistance readings.

With the switch open and the 100k inserted, you may be able to measure teh 2.7k while it is connected, but ionly while the pot is set at 100k. Turn the pot way down and yo are back to being unable to read the resistor.