Robert McGwier wrote:
Apologies. I know CREE is a leading company in GaN and I leaped without being careful. This is SiC technology.
Bob
We shall forever remember that Bob's GAN was [sic]'d (see http://en.wikipedia.org/wiki/Sic) :-D .
Seriously -- this looks great. I hope Marc can get a hardware version running soon. I am very impressed with the nearly parabolic Po vs Vd curve (fig 5). This sez to me that the "plate" source impedance is quite constant vs Vd, which in turn makes the HELAPS "plate modulation" application even easier.
From Fig 3, I gather that the predicted 3rd harmonic is more than 70 dB
below the fundamental. This is good because the 3rd harmonic of 145.x downlink is pretty close to the 435.y MHz uplink, and the fewer filters that are needed, the better! I wonder if perhaps some of the C5,5,7,8 capacitors could be open circuit ¼-wave @ 70 cm coax stubs to improve the isolation even more?
I'd like to suggest that the EAGLE Class-E "phase" PA be 3 "parallel" separate stages differing in phase by 120º, each feeding a separate dipole (which are physically oriented at 120º on the +Z antenna farm), running at Vd in the 15-25V range. This would afford 3x redundancy (remember how the AO-40 2M amplifier death was a single-point failure), and also puts the total PA DC efficiency > 90% (see Table I), and also saves the power losses in a 3-way high power divider. My gut feeling is that, for the relatively narrow signal bandwidth, all the amplifiers can share the same "plate" modulator.
Nice work, Marc!