George - I greatly enjoyed this particular post from the link you provided. Made sense to me. Or so I think
---------------------
1st and foremost a tube (the triode) is the most linear amplifying device invented by man.
All else follows from that.
A typical SS Amp (order of magnitude figures) will have an open loop gain of 100,000 with a feedback factor of 5,000 giving a closed loop gain of 20.
A typical tube amp will have an open loop gain of 50 with a feedback factor of 2.5 giving a closed loop gain of 20.
I'm talking about a combination of local and global feedback here.
Any feedback signal is time delayed and smears time coherency. The more stages the feedback is applied over the worse that problem.
Tube amps therefore tend to sound more "correct" in terms of pace rhythm and attack, they are just (for want of a better adjective) more fluid and "breathe" with the music.
This tends to influence things like stereo imaging or "sound stage" and ambience.
Higher amounts of feedback tend to counter soft compression effects and make clipping harder and more abrupt. In engineering terms we say the transfer function graph (signal voltage out vs signal voltage in) is less smooth with more abrupt corners near the limits. The more abrupt a change in the transfer function the more higher order harmonics are generated. Anyone who has studied musical theory will know that the second harmonic of (for example) a C note is a C one octave above and 3rd harmonic is still musically related BUT by the time you get to 7th, 9th , 11th harmonic the "note" produced is musically unrelated to the fundamental, we call this musically "quint". While 1% 2nd harmonic distortion is "euphonic", about 0.01% of 7th harmonic is seriously objectionable and just grates like fingernails on the blackboard.
The more feedback, the more harmonic multiplication, that is the harmonic distortion products which are in the feedback signal are subject to further harmonic distortions which are then fed back to produce even more.
There are other things at work as well BUT I believe that the level of feedback is the major difference in the sound between SS and tube amps and that teh time coherency is at least or more important than the harmonic distortion profiles.
I design tube amps with local feedback around each tube stage, seldom over 2 or more stages, never over 3 or more stages, and in particular don't use "global" (from output back to input stage) feedback at all.
Just one mans opinion, but an opinion based upon designing and building both SS and tube amps over many years (I've given up the SS Amps now).
Does that mean I don't use SS at all - NO, my tube amp designs use transistor current sources in the "tail" of triode tube differential amplifiers on the input, I often use MOSFET Source Followers with transistor current source loads to drive the output tubes etc. BUT all of the gain stages are tubes. Also I borrow ideas from modern SS amp design to implement things like balanced shunt feedback loops (local feedback around only 2 stages).
-------------------------
