If one is serious enough about civilized PSK sidebands to consider linear
amplification-- which is no trivial matter at the "California kilowatt"
levels being discussed here lately-- why not simply take the next logical
step?
Class C, D, and E amplifiers are not that hard to modulate. Rather than do
anything so drastic as tinkering with PWM in the RF driver stages, apply PWM
methods to the power supply for the final!
The late Helge Granville of Motorola was a big proponent of Envelope
Elimination and Restoration, a technique in which the amplitude envelope of a
low level modulated (and filtered) signal is extracted and amplified, which
can be done with the same efficiency as the Class D or E RF final. The RF
component of the signal is hard limited and amplified through the regular RF
chain, preserving whatever phase changes are implicit in the modulation
technique being used.
If the amplitude component and RF phase/frequency components are reunited in
the final amplifier without excessive time difference, the result is a close
approximation to linear amplification, but at switching-mode efficiencies.
Furthermore, since the sharpest phase changes take place during those
instants when the supply voltage is at its lowest values, the strain on the
RF power transistors is minimal.
Admittedly, to implement EER for high-performance SSB or ISB transmitters,
spread spectrum, etc., is not a trivial matter either. The required dynamic
range and time delay matching can be pretty stringent. Still, amateurs have
built some very good 200-500W HF transmitters using the method, and the
requirements are not nearly so tight for relatively low speed data modes such
as we use on LF. It might be worth a try.
73,
John
|
