Owen Duffy wrote:
It is no doubt an appealling explanation of why a PA may run hotter under
some circumstances, but it does not explain why for instance under some
circumstances, a PA may run cooler on a mismatched load.

What does explain it is the amount of destructive vs
constructive interference occurring at the source.
Assuming an unprotected source:

If the constructive interference is toward the
load, the source dissipation will decrease. If the
constructive interference is toward the source,
the source dissipation will increase. The conservation
of energy principle really does work to conserve the
ExH energy in a reflected wave.

Just because a special case results in zero dissipation
in a voltage source does not give us the permission
to make a magical leap of faith to "reflected power
doesn't exist" or "reflected power is *always*
dissipated in the load".
--
73, Cecil http://www.w5dxp.com