[time-nuts] comparing two clocks - Racal 1992

Bill Hawkins bill at iaxs.net
Sun Feb 23 15:51:37 EST 2014


Most of this discussion of comparing clocks has been about time
interval.

The Racal 1992 has a function called "Phase A rel B" that reads the
phase angle in degrees regardless of frequency.

At one time I had two HP Z3801 receivers and three 1992 counters, so it
was natural to compare the two receivers with phase. There is no concern
about the minimum interval between A and B. It does not matter which
comes first. There is no sign with the 0-359 measurement.

So what you do is record the phase angle reading with time (manually in
my case, but there is a connector for GPIB). The rate of change gives
the error, where one count is 1/36th of a nanosecond for 10 MHz
oscillators. Cable length matters if you want the absolute difference,
but a relatively steady phase (it isn't noiseless) over an adequate
amount of time yields satisfactory results for adjusting rubidium and
cesium oscillators.

You don't need a precision external reference when you are dealing with
one part in 360. The noise is about 5 parts in 360.

Comparing instantaneous clock display times is another matter
altogether.

Bill Hawkins

Age disclaimer: My memory isn't what it used to be.




More information about the time-nuts mailing list