Output stage optimal bias

[hpl]

Hi,

I'm currently reading "Audio Power Amplifier Design Handbook" by D. Self and am having trouble understanding some parts. Maybe some of you guys can enlighten me (it is probably pretty basic).

How is what the author calls "optimal bias" determined? Is it done empirically or can the optimal value be derived (calculated) from the output device parameters? If so, how?

I have tried to simulate several of the circuits in Chapter 5 and experiment with Vbias to see if I can duplicate the values the author presents in table 5.2, but I'm unable to do so. Ofcourse, this could be due to that i'm probably not using the same models that he is, but I figure I'm doing something wrong.

[P.Lacombe]

Hi hpl,

Things are not simple at all.

Mr Self don't say importants things, such that determined optimal bias (any manner...) cannot be stable : variations of dissipated power in final devices (amplifiers are not for measure, but for music listening !) causes important variations of Vbe and Hfe.

But Vbe and Hfe are the most important parameters to determine correct bias (which causes minimal distorsion)...

Methods for compensatig this annoying phenomenon are described in many, many patents...

Instead of this, I use lateral mos transistors, which exhibits quasi null temperature coefficient and stable transconductance at optimal bias.

[Shaun]

Originally posted by hpl
How is what the author calls "optimal bias" determined? Is it done empirically or can the optimal value be derived (calculated) from the output device parameters? If so, how?

IIRC, he determined his optimally biased class B empirically, and thereafter defined it as a certain voltage drop across the Re resistors of the output transistors (as opposed to bias current).