Thanks all for the comments. I am waiting for samples from Cree. Once I get them, I will build a unit. I just got the Modelithics models for all the capacitors and inductors today. I will run another simulation with those models that seem to be better than the ones I used.
In any case, SiC devices seem to perform just fine down on 2 meters. GaN technology from Cree so far is only one internally matched device good for 3 to 3.5 GHz, and it is only 15 W. Eudyna has the devices we need, but they will not give away samples under any circumstance. Nitronex devices are also narrowband. And, more important, Cree is the only company so far that has a bunch of talented engineers who developed some reliable models. If GaN devices become available, we can always try them out in a similar circuit.
Regarding this statement:
From Fig 3, I gather that the predicted 3rd
harmonic is more than 70 dB below the fundamental.
The harmonics have been effectively attenuated by the 5 element Chebyshev filter and the matching network. The output of the amplifier without the filter has the harmonics at very high levels since it is a switching amplifier.
Usually, if you can simulate somthing good, you have a good chance of getting it to work close enough. The opposite could be true, too, but it is much less common.
I'll be out in the field this weekend for the 10 GHz contest, and then off to the EME conference in Germany. Will do more work on this after that. I am sure there will be more question and suggestions for improvement in the meantime.
73, Marc N2UO
__________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com