Joel Kolstad wrote:
On a reciever, I'm using a MMIC LNA at VHF (2m), and while it's quite clear to
me why I need to match at the amplifier's input -- to minimize VSWR and get as
much of the signal from the antenna into the system -- it's not quite to clear
to me just how important output matching is. The S22 parameters for this LNA
are very close to an open circuit with a little capacitance -- I expect the
output looks back into a drain or something else that's a reasonably decent
current source at RF, and you see that shunted by parasitic capacitances.

So... how important is it that I bother to build a matching network at the
output? (I'll be "outputting" to another ~50 ohm input.) It doesn't really
seem that important, and building something like an L-match with de-Q-ing
resistor drops the gain a couple dB anyway.

I know that for a transmitter output matching is important so that reflections
from poor or non-existant terminations don't blow up the output amplifier, but
for low-level signals as in a receiver it wouldn't seem to matter so much?

Thanks,
---Joel Kolstad

OK, since you called it an LNA, I assume you want to take advantage of
the low noise. In that case, you do NOT necessarily want to match the
input for low SWR. You want to match to the amplifier's input noise
resistance: the equivalent input noise voltage divided by the
equivalent input noise current. This in general is NOT the same as the
input impedance. Adjust the input match for the lowest noise figure,
not the lowest SWR, if it's low noise you want.

Output matching will transfer the greatest power to the load. Assuming
S12 is very low, adjusting the output matching will not materially
affect the input power, so maximizing the output power will maximize
the gain. You may or may not have a need to do that. And you need to
pay attention to amplifier stability: is it unconditionally stable, or
must you keep the load within some bounds to keep it stable? Have a
look for articles about "maximum available gain" and "maximum usable
gain." For example, see
http://www.microwaves101.com/encyclo...factor.cfm#mag.

Cheers,
Tom