On Apr 30, 8:54 am, huiliu wrote:
Hi, all, my current project requires a high phase stability (1 mill
degree @ 10 MHz / oC) resonant surface coil. I am trying to use a LC
resonant circuit to drive this surface coil, but its phase stability
is really poor (about 40-80 mill degree @ 10 MHz / oC ). I guess the
main problem comes from the used capacitor whose value is easily
changed by temperature. so my question is that can the crystal
oscillator help me more? or is there any other method to achieve this
requirement?

Thanks a lot.

If it is as you suspect, that the capacitor is causing the trouble,
why not use a capacitor with low temperature coefficient? C0G
dielectric ceramic capacitors are rated at +/-30ppm/C maximum, and I
have found some to be very much better than that.

I have no idea what the circuit is that you are driving, but I would
expect that if you are driving a high-Q resonant tank, the phase shift
will be very rapid with frequency, and that implies the need to use a
tank circuit with very good temperature stability. If I'm not
mistaken, the rate of phase change at resonance of the impedance of a
series RLC, expressed as radians per fractional frequency change, is
2*Q. So for a milliradian change, you need to keep the resonant
frequency within f(center)/(2000*Q). A milliradian is 57
millidegrees, so to keep such a tank within 1 millidegree if the Q is
only 10, you need to hold the resonant frequency (relative to the
excitation) to a little better than one part per million. I'd say
that's not a very reasonable goal if the inductor dimensions change
with temperature. It will take a decent oscillator just to maintain
the excitation frequency to that sort of accuracy. You might actually
find a capacitor that will give you that sort of stability, by
selecting from C0G caps (or possibly by making your own by plating
electrodes on fused quartz...). But that doesn't solve the problem of
the inductor itself.

Could you characterize the tank circuit frequency shift versus
temperature and compensate by causing the excitation frequency to
change synchronously with that?

Cheers,
Tom