Unity Gain Stage

[R.G.]

I'm with pdf64. If you're going to go feedback, use as much internal gain as you can, and tinker the feedback components to get the gain and other characteristics you want.

I understand the idea of engineering economy - it works good enough without an other component to pay for - but the general concepts of feedback are to get as much gain and as little phase shift in the gain element as possible, and let the feedback components determine as much as possible.

I've been reading the thread and thinking in the back of my head "that's an application that cries out for a high-voltage opamp". Say what you will about opamps and tubes, where you want just a signal inversion, opamps do that with no muss, no fuss.

Applications like this really need something like a small, high voltage diffamp and gain stage to drive the tube that provides the output voltage.

[Ten Over]

An unbypassed cathode facilitates noise to enter the signal path. To avoid this occurring early into the signal path (ie where it causes the greatest degradation of the signal to noise ratio), it’s beneficial to fully bypass the cathode of an input stage.

I can certainly agree with this and I have always been a staunch advocate of shunting noise, particularly 50/60Hz noise, to ground. However, all three of my designs have input stages with fully bypassed cathode resistors. Design #2 has an un-bypassed cathode resistor on the local NFB stage, but the signal is at its largest preamp magnitude at this point so that degradation of the signal to noise ratio doesn't scare me.

Design #3 has an un-bypassed stage in the second position, but again the signal is larger than the input signal at this point. This stage has lower gain than the others so that a larger signal at the grid is necessary to get the same size signal at the tone stack. This stage also has an output impedance that is close to the original so that the tone stack is driven just about the same.

[Ten Over]

I'm with pdf64. If you're going to go feedback, use as much internal gain as you can, and tinker the feedback components to get the gain and other characteristics you want.

I understand the idea of engineering economy - it works good enough without an other component to pay for - but the general concepts of feedback are to get as much gain and as little phase shift in the gain element as possible, and let the feedback components determine as much as possible.

Let's not lose sight of the fact that this is just a simple little guitar amp that we were really only trying to get unity gain with phase inversion. Even so, the performance with or without a bypass capacitor is incredibly similar. The frequency response is dead flat either way. I don't have any equipment that is sensitive enough to pick up any difference in the signal-to-noise ratio. The phase shift at 5kHz is virtually identical either way and quite minimal, especially as compared to the phase shift on his other channel.

Nonetheless, if you feel better with more open loop gain, then change the 100K resistor in the feedback circuit to 33K and have fun soldering in the bypass capacitor.