[time-nuts] Sensitivity to g in atomic clocks

Rick Karlquist richard at karlquist.com
Tue Jan 11 18:40:46 UTC 2011

Bob Camp wrote:
> Hi
> There are three things people refer to as G sensitivity on a crystal
> oscillator:
> 1) 2 g tip sensitivity (flip it 180 degrees on the bench).
> 2) Vibration sensitivity (vibration translates into phase noise)
> 3) Relativistic effects (take the oscillator to the moon and it's
> frequency
> changes).
> The first does not apply to an atomic clock. Since (unlike quartz) they
> are
> sensitive to magnetic field, they will change when you flip them around,
> but
> for other reasons. That would make actually proving a "no sensitivity"
> spec
> difficult.

Actually, some cesium beam tubes are sensitive to acceleration
because of beam ballistics.  The 5071 CBT is specifically engineered
to make this effect negligible.  In Rb standards, inverting the
clock might change the temperature distribution and affect the
frequency that way (after a delay).

Regarding relativity:  the state of the art clocks at NIST supposedly
are so precise that they could detect the relativistic shift between
the 1st and 2nd floors of the building.  Not that they fit in a

Rick Karlquist N6RK

More information about the time-nuts mailing list