John Larkin wrote...
Why do these FETs have such high DC gate currents?
Thanks,
- Win
(email: use hill_at_rowland-dot-org for now)
I think gaas, and compound semiconductors in general, have very high
defect densities; plus, these parts have schottky gates, and schottky
diodes are always leaky for some reason.
Another interesting property of gaas is "trapping states." The high
crystal dislocation density creates places where charge can get
trapped under gates, so the transconductance and pinchoff voltages of
mesfets is a nasty function of temperature and bias history. Under
constant-bias conditions, Id can change 10% overnight. I managed to
make money off this effect, by replacing a piece of very expensive
gear that depended on mesfets to behave, and they wouldn't.
Given what crummy parts gaasfets *appear* to be, they can hit noise
figures below 1 dB as RF amps.
I haven't tried any of the phemts as regards bias stability, but I'd
expect similar behavior. They do seem to be a lot more predictable
than mesfets, though. NE25339 (see the NEC.COM site, or buy from
Mouser) is a dual-gate mesfet.
John