On Mon, 21 Apr 2008 17:04:02 +0100, Richard wrote:

I need some information on an appropriate input circuit for a small RF
power amplifier using a 2E26 operating in the HF region. Output would be
about 27 watts. Drive input power about 0.12 watts.

I want the input circuit of the PA to present a 50 or 75 Ohm load to the
driver (the exciter - probably a CW QRP rig - will supply about 0.12
watts). Perhaps (I muse) it would be good, (certainly it would be
convenient) if the input circuit was NOT a tuned input circuit, but see
below.

Questions (that incidentally show my thinking):

Can anyone point me to a website that gives the design data for an
appropriate circuit given the things I have mentioned?

Is it okay to have an untuned grid circuit? I muse no problem if the
exciter has a pi-network output circuit. I just wonder that if the
exciter uses a pi network, if that would represent both the exciter
anode/collector tank circuit, and the grid tank circuit. (So, really
there would be a tuned input circuit - the output pi-network of the
exciter). (Just shows I know nothin', I'm not even certain whether solid
stage rigs use Pi-networks to match with the load).

Any thoughts on biasing arrangement - fixed bias, drive bias, or
cathode bias? (Perhaps fixed bias is best - no need for a clamp valve).
Grid bias needs to be about - 45 volts. Muse that a simple RF choke
supplying the bias to grid would suffice.

So, basically, I want to be able to plug a QRP rig into the PA, tune up
the QRP rig and the plate circuit of PA, and that's it.

Elsewhere you say that a 2E26 needs 170mW, here you're going to give it
120?

If you want to minimize the drive power you're going to have to use a
tuned grid circuit; having such a power-efficient design is going to make
your circuit susceptible to oscillation, as it'll take less than 1% of
the output power to come back in to the grid circuit and cause problems.
So a 120 (or 170) milliwatt input implies both a tuned grid circuit and
neutralization.

If you want to go broad-band, you need some dissipative elements in the
grid circuit to broaden the response. Something like a plain old
resistor would do quite well. With resistive loading, your grid circuit
will both be broader band _and_ your need for neutralization will be far
less. I'd be tempted to think along the lines of a 1W QRP rig and
resistor in the grid circuit that dissipates 800mW. I'll leave the
calculations to you, but this will broaden your grid circuit response
considerably, ease your bias network design problem, and reduce (or
eliminate) the need for neutralization and/or obsessively high degrees of
isolation in your circuit layout.

Have fun -- take everything I just said with a grain of salt, because
I've only ever designed transistor PA stages.

--
Tim Wescott
Control systems and communications consulting
http://www.wescottdesign.com

Need to learn how to apply control theory in your embedded system?
"Applied Control Theory for Embedded Systems" by Tim Wescott
Elsevier/Newnes, http://www.wescottdesign.com/actfes/actfes.html