What is wrong?

[lumanauw]

Hi, Mauro,

Long time no see

Yes, I notice that the more you put poles everywhere in the cct, the sound will become "dead". Maybe it is super stable, but not nice to listen to.

I think I recall that someone here is experimenting a design that has pole,zero,pole,zero in the whole loop. I don't remember who it is, but with this experience, maybe that kind of approach can be good sounding. Sure it is difficult to intentionally put zero anywhere, but wanting the whole amp to be stable. It's kind of standing on the edge of a cliff. You can see the whole beautiful scenary, but if you are not carefull, you will fall down.

[mauropenasa]

Hi, Lumanauw

Yes, more time...

The main point is doesn't concentrate only on the stability.
If you are aware that serve avoid the standard compensations, the way finds.
A lot of planners have developed a linear compensation techniques. Enough chooses that more applicable to the circuit.
For example, Graham has used techniques distinct from my, on his amp, but the instrumental result (DF linearity) can be similar...

Ciao

Mauro

[Miles Prower]

This assertion ("sounds better"... so repetitivly seen) is typical of what I call an inobjective comment.

Yes, it is "unobjective", but that alone does not make the matter unimportant in and of itself. In many cases, "unobjective" simply means that we don't yet have complete information concerning the phenomonon in question.

This is my biggest disagreement with Mr. Self: his attitude that if you can't measure it, then it's not important. He's making the same mistake as the "subjectivists", but from the opposite direction. Audio design (all electronic design in fact) is science + art. You need both or you will never get anywhere. You can measure DC bias levels, o'scope the output while running all sorts of waveforms through it, analyse distortion till the cows come home. In the final analysis, you have to connect the speeks and listen to the durn thing to see if all that scientific analysis has led you to what you were trying to accomplish in the first place. That, by definition, is subjective.

It is absolutely false that it sounds better, it is absolutely true you prefer the sound of Ollsen's amplifer (a huge lot of difference !)

A difference without a distinction. When I started the project, I certainly wasn't intending to use a quasi-complementary final. That whole topology just looks ugly right on the schematic. I decided to go with it for the main reason that one of the design objectives was that one could get everything needed to construct the circuit from the neighbourhood "Rat Shack". (And at the time, I had never heard of either Self or Ollsen.) They've been giving the do-it-yourselfer the shaft for years now, and they just don't have a decent selection of matched pairs of complementary power BJTs. So I tried doing it both ways to see if the quasi-comp would give acceptable results. Turned out that it worked better despite the nominally higher THD. There's probably an explanation for that out there waiting to be discovered.

I am not aware of IC's désigned for good sound and not having very good specs and a lot of feedback. And there usually are some IC's at the recording stage. Do they make so much harm to the perceived sound ?

And in the vast majority of cases those ICs are working into consistant, well characterized, loads. Powering up speakers is a whole 'nother story.