Just playing around with the upper end on the UHF band. Thought I'd share it with you guys. The antiparallel oscillations can easily be multiplied to generate a 2GHz frequency. Left out the filtering to not ruin the symmetry. It is beautiful :D
P.S. There are also some parasitic 5GHz oscillations. If you adjust the time frame to 100ps/s you can see them rise, but for some reason they are not stable and begin to give in to the fundamental resonance in a minute or so.
For efficiency decrease the bias level on the transistors. It seems to work all the way down to 1-2mA (aa long as the devices work in the linear region). This high idle current I've chosen to give rise to the 5GHz oscillations I mentioned. They fade away when biased low enough.
