[time-nuts] Altitude Effect on Cesium Standard?

Tom Van Baak tvb at LeapSecond.com
Wed May 7 23:08:59 EDT 2008

> Hi Tom:
> I've been making some measurements of the Datum 4065 vs. GPS where the 4065 is 
> setting all the parameters (i.e. I have zeroed the offset and it's doing it's 
> thing).  After some days the t/T plot looks like a straight line at about 4E13 
> with a couple of ns bobble.


> The 4065 is at about 910 feet (280 meters) elevation.  Your web page:
> http://www.leapsecond.com/great2005/  says about 1.1E-16 per meter so that's 
> 3E-14 about an order of magnitude smaller than what I'm seeing.

Makes sense. Internal frequency offsets for that vintage cesium
standard is likely far greater than relativistic blue shift corrections.

> The 4065C manual says the accuracy is 1E-12 or 5E-13 depending on if the tube 
> is rated for lifetime or high performance.  I assume this means without any 
> adjustments, i.e. power up and that's what you get.


> Do you have a feel how well your Cesium standards set themselves?
> -- 
> Have Fun,
> Brooke Clarke


I wish I had a nice table showing you my results, but I don't
have the data handy. It very much depends on the model of
cesium standard. The new ones, like the 5071A, can truly
be called primary frequency standards because great care
was taken in their design and manufacturing; if the spec is
that they power up to 1e-13 then that's what they will do, every

The old ones, like the 5060 or 5061, and various FTS cesium
standards were happy to be accurate stable to 1e-11 or 1e-12
or so. Most of them had Zeeman inputs so you could tune it
closer in accuracy.

The C-field adjustment for the old 5061 were very coarse so
that the user could tune the output frequency to match some
other national lab standard. Note also in those pre-leap second
days, when there was both atomic and astronomical second,
one needed to adjust the output frequency by multiples of
50 or 100 x 1e-10 every now and then. So a coarse adjustment
was needed. Later model 5061, and I think all FTS models,
went with a fine C-field adjustment only.

Two experiments you could try with your 4065C -- one is to
power cycle it and measure the offset (this will take some days
to hours against GPS). Repeat this a few times and see the
distribution. You will learn a lot about its accuracy, stability,
and retrace.

The other experiment is to calibrate it using the Zeeman input
and see what that gives you for a frequency offset.


More information about the time-nuts mailing list