[time-nuts] Rubidium Drift
John Ackermann N8UR
jra at febo.com
Fri Nov 3 09:04:34 EST 2006
Tom Van Baak wrote:
>> I've recently acquired a Rubidium Osc ( Efratom FRS-C 10Mhz) as my
>> first venture into this field. It's working as per the spec given
>> ,checked against Trimble GPS with 1pps output.
>> Can you point me to somewhere where there is a description of the
>> physics behind the drift that occurs?Drift is speced as <5x10^-11 per
>> month....but why should there be any "drift" at all?
>> "Wandering" about nominal I can expect ,but consistent "drift" ?
> Hi Peter,
> You're off to a good start. With enough days of data
> you should be able to measure the drift quite nicely.
> To answer your good question; we all have come to
> know that cesium is accurate and rubidium drifts. But
> there's much more to the story...
[ Tom's good stuff snipped ]
For what it's worth, I have plots with (at the moment) 98 days worth of
data showing two HP 5061As and an HP 5065A against GPS (and each other)
at http://www.febo.com/time-freq/plots. The drift of the 5065A is hard
to discern because what I presume are other environmental effects hide
it; in other words, it's pretty darn low.
You'll see a whole bunch of files there with hopefully reasonably
explanatory names like "cs1-gps-hourly.html" which shows CS1 vs. GPS
with hourly averages. The files with "-daily" are 1 day averages, while
the ones with "-tail" show the seven most recent days of data with 10
Each web page shows the raw phase, the drift-removed phase (using a
linear calculation of drift) and ADEV as well as some other statistics.
The pages are automagically updated every 15 minutes.
I'm hoping to keep the experiment going for a while longer; I'd like to
get out to at least 120 days of data. I can make the raw data available
if anyone is interested.
More information about the time-nuts