Correct me if I'm wrong, but most valve amps don't use NFB because it's not plausible in the design. Tying the feedback loop to the output with an OPT would introduce phase shift and all sorts of other problems. Rod Elliott wrote about this a little bit. Maybe there are a few designs out there that manage to do so, but I'm not aware of them.
Also, since most valves are constricted to fairly low gain, trading it in with NFB doesn't make sense. There's no need to use NFB to stabilize the circuit like you would with an op-amp (infinite gain). Saying that valve amps need "adequate negative feedback" is a bit misleading methinks.
Yeah, I would say you are incorrect in the idea most valve amps don't use NFB. Most have a small amount of global NFB. You can't use too much or the circuit will oscillate due to phase shifting. Of course in the last 20 years many triodes SE, and PP have been put forward without global feedback due to the idea NFB is always bad. Even so I believe most you encounter do have NFB. The NFB used in tube amps is normally not a large amount. This will lower distortion a bit, extend bandwidth a bit, and lower output impedance some. It helps to have quality OPT's.
This page has pertinent formulas for designing tube feedback circuits. Very near the bottom of this long page are two graphs of response for an amp with and without feedback. In this case 12 dB feedback. One shows it without phase compensation in the feedback path and one with phase compensation.
Edited by esldude - 7/4/13 at 12:50pm