Steve:
I have also built several amplifiers based on the "Granberg" doctrine,
and my results agree very well with yours. I have tried a number of bias
schemes ranging from dirt-simple to over-the-top, but it always works out
about the same. You need to get just enough IDD flowing to get the
transistors into their linear range.
The generation of low-order RF harmonics will always be a problem,
unless you are prepared to run IDD way up, at which point you have a
combination amplifier/shack heater. My preference is to do what I have always
done anyway - a decent lowpass filter on the output. This doesn't solve the
entire ploblem, so I have built myself both 5-10MHZ low pass filters, and 2
or 3 section PI filters. The latter are a quick and easy way to get another
10-20db of suppression.
For my next amp I have decided the best topology to be not broadband -
too many compromises. Trying to get an output transformer optimized for use
from 1.8 to 30MHZ is pretty much futile. So I have decided to just bite the
bullet and use a design with either totally switchable output circuits
(including harmonic suppression) or to subdivide into three bands -
low,mid,high (pretty original, eh?) This may involve a little foray into the
design of mechanical switches, possibly motor driven. I have devised schemes
(on paper, of course, so far) that are really more elegant overall than a
bank of expensive vacuum relays.
I think that the government should pass a new law that every radio ham
whose name ends with ASCII code $60-$7B will receive a yearly allotment of 8
MRF150's.
Keep your emitters emitting and your collector collecting.
Eric von Valtier K8LV
|