alhearn wrote:

If you read Chapter 13, "RF Power Amplifiers and Projects", in the 2004
ARRL Handbook, there are pages and pages of discussion of matching
plate or transistor output impedances to 50 ohms output and not once do
they ever refer to output networks dissipating or reflecting reflected
power. Not once. It's always a matter of simply matching one impedance
to another

Sorry, that isn't correct. Maybe the Handbook doesn't emphasize the
point clearly enough, but the design of output networks is specifically
*not* about impedance "matching".

The objective is to *transform* the impedance of the load at the
amplifier's output socket (usually the design value of 50R) into the
correct value of load impedance needed by the amplifier device.

As Al says, the correct value of load impedance for the amplifying
device is the one that will:

provide the highest power output consistent with required linearity,
while staying within the amplifiying device's ratings.


The whole point is to understand that the output network is not trying
to match any particular impedance. It's simply giving the amplifying
device the load impedance it needs in order to function correctly.

Those two objectives are not the same... and understanding the
difference is vital for this whole discussion.


--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek