Bill,
A linear class C amp is quite a trick.
Yes, it's tricky, but it can be done. Variable bias controlled by an
error amplifying loop is one way to do it.
REPLY:
If you're varying the bias to make it linear, it's no longer Class C.
I beg to differ. I don't see why varying (modulating) the bias turns it
into something else than class C. I maintain that as long as the bias
modulation I do is such that the active device conducts over less than
180 degrees of the RF cycle, it's still class C.
Imagine a very simplistic theoretical case: You have a transistor or
tube that has a straight transfer curve, and you modulate the bias
proportionally with the drive signal amplitude, starting from cutoff
bias, so that at any signal amplitude the device conducts for the exact
same 120 degrees. And you keep the drive level below the saturation
point. This amplifier would be linear, it would be class C, it would
have varying bias, and it would be somewhat more efficient than a class
AB amp.
Of course practical transistors and tubes don't have that straight
transfer curve, and also you want to drive the amp well into saturation
to further increase efficiency. This is where the bias modulation must
become a little more complex, for example by deriving the bias from an
error amplifier comparing input to output amplitude. The actual bias
voltage will then vary in a complex way with amplitude. But the
amplifier can still operate fully in class C.
The amplitude linearity of such a linearized class C amplifier can be
far better than that of a conventional class AB amp. Phase distortion is
another matter, though. Such amplifiers are best implemented with
devices whose capacitances don't change much (tubes), or devices that
have capacitances low enough that the phase modulation resulting from
capacitance fluctuations isn't too bad (UHF-capable LDMOSFETs at HF).
Manfred
========================
Visit my hobby homepage!
http://ludens.cl
========================
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|