Maximum amount of negative feedback?!

[roberto]

Of course if you use frequency dependant NFB, like a presence or a depth control are, the amplitude of the frequencies involved will change. This is why the presence increases the mids and highs.

Those graphs are an indication of what happens: less NFB, more THD.
Guitar amps have higher THD that what is shown on those graphs.

[printer2]

But back to my initial comment, if you have NFB (lower output impedance) the sound is flatter from the speaker, less or no NFB (higher output impedance) and the bottom end and the highs get accented. It is just the amp acting more as a voltage source because of the NFB and more like a current source without NFB. When the amp manufactured making SS amps figure this out that the interaction between the amp and the speaker contributes to the 'tube' sound they started using current feedback on their SS amps.

[vibratoking]

There hasn't been a mention of stability. Too much feedback can cause the output section to go unstable. In many amps, you can't just blindly apply any amount of feedback that you want.

[tubeswell]

There hasn't been a mention of stability. Too much feedback can cause the output section to go unstable. In many amps, you can't just blindly apply any amount of feedback that you want.

Yup. Global NFB through several gain stages can experience phase-shift presented by interstage C-R networks (i.e.: coupling caps in combination with source and grid leak resistance). What's more interleaved OTs can be more prone to phase shift between Pr and core and then again between core and Sec. So different frequencies can present a spectrum, from negative to positive feedback.

You could limit the amount of feedback you apply, so that the amp remains within the stability margin, or you could add more feedback by hedging the amp's stability through 'slugging' the dominant feedback pole (with like 250pF from grid to ground on the output tubes) as Merlin Blencowe suggests.

[matt h]

(deleted)