[time-nuts] Frequency error -> Parabolic Time Error?
John Ackermann N8UR
jra at febo.com
Sat Apr 9 18:32:46 EDT 2005
Brooke Clarke wrote:
> Suppose that my reference oscillator is offset from 10.0 MHz by some
> small amount. If I plot the time interval between the reference
> oscillator and a GPS 1 PPS what will the plot look like over a long
> time period, a parabola?
> Have Fun,
If it's a stable offset, you should see a straight line with a linear
slope. If there's drift (ie, aging), then you'll get a parabola (and if
there are a couple of things going on, you may get a more complex
pattern yet). Usually, the frequency offset is by far the largest
component of the phase plot, so removing it makes any drift much more
visible; for that reason I recently updated my stable-stats program to
plot both the raw phase and the phase with linear offset removed.
More information about the time-nuts