On Mon, 26 Jan 2004 19:55:30 +0000, Paul Burridge
wrote:

Hi guys,

I've just been doing some calculations out of sheer curiosity. It
turns out that to make a tank circuit for 450Mhz (which isn't *that*
high a frequency by today's standards) would take 5pF || 25nF. These
are both *very* low values. 5pF is getting perilously close to being
seriously affected by circuit board stray capacitance and 25nF isn't
much better; little more than a couple of turns of wire, I guess.
What does one do in such circumstances? Should one be thinking in
terms of etching these values out of the PCB by the time one gets of
to these frequencies, or is it still acceptable to make them up out of
discrete components?

p

You can buy surface-mount caps below 1 pF and inductors below 1 nH, so
a discrete LC is perfectly feasible at 450 MHz. Even with a couple of
pF of circuit stuff - say, the gate of a gaasfet or an IC - you should
still be able to make a resonator or a filter up there. A small
trimmer cap or varicap would help nail the resonant frequency.

PCB capacitance for 0805 or 0603-size parts is pretty small, and you
can keep connection inductance in the 1-2 nH range.

Helical resonators are nice, too, as are coaxial ceramic resonators.
450 MHz is sort of in the transition region between discrete and
distributed resonator domains.

John