[linrad] ATF33143 and K.

[wb9uwa@xxxxxxxxxxx]

HI Leif,

I have learned much experimenting with the simulated ATF33143 amp in Serenade SV.
The referance amp from Lionel uses 3.2nh inductance on each source lead. I did a simulated
output impedance measurement and found it to be 15 ohms. The overall results are much
better when using 14.5nh on each source lead. In that case, the output impedance is 55 ohms.
I find that I still must throw away some power in resistors, so I add 10 ohm in series. This 
makes 65 ohms, so I use 65 ohm resistors in a diplexor. A 220 ohm load resistor makes
S22 look like 50 ohm at all frequencies. Odd thing. For the most part, the drain needs series
resistance for a good K rather then some loss by adding a load. In this case when I remove
the 220 ohm load resistor, K drops from 1.13 to .99 near 144 Mhz.

I did try using less source inductance to lower the impedance. This created a moving target.
I could never get K back over 1. Even more resistance had to be used in series with the drain.
K was never really very good at any source inductance setting.  I had to use a feedback
network to turn it into a VHF transistor. Then K is very high except at 144 where it can be
just over one as long as some series resistance is used.
I think conventional circuits throw away a lot of gain and power. The lossless feedback
circuit throws away a similar amount of gain, but perhaps throws away less power.
I was able to bring the impedance down to 50 ohms, but K is dangerously close to 1 then.

I am anxious to make some p1 measurements.

K is a funny animal. I am still trying to figure it out. It is some sort of a gain limit.
I have out of band K under control. It is impossible to have large amounts of gain
and a good 144 Mhz K at the same time.
The preamps that I have modeled (designed by others) all have a K below 1 at some
frequency or another. Usually it is poor near 2M and above 5 gig. My preamp has a 
K in the thousands above 1Ghz.

73, Jim.

LINRADDARNIL