[time-nuts] Exact Rubidium frequency
Tom Clark, K3IO
tom.k3io at gmail.com
Mon Jul 2 01:34:48 EDT 2007
rlutwak at comcast.net wrote:
To the best of my knowledge, the best measurement of the Rb ground state hyperfi
ne frequency comes from the LPTF group in Paris, 6 834 682 610.904 333 Hz, with
an accuracy of a few parts in 10^15, by direct comparison of a Rb fountain to a
The Rubidium hyperfine transition at 6.834... GHz is sensitive to the
magnetic field -- more so than cesium or hydrogen. The builders of Rb
standards make use of this to fine-tune the output frequency of the Rb.
The "C" field is in fact used to tweak the frequency of a Rb. It is
because of this sensitivity that Rb is not considered to be a "primary"
As usual, Tom van Baak has a nice write-up on this at
http://www.leapsecond.com/pages/hp-5065a-cfield . For a 1 Gauss
magnetic field change (roughly that of the earth where it is said that
one gauss is as good as another), the frequency is shifted by ~1:10e10.
Back in the era before NCOs (numerically controlled oscillators) and
digital synthesizers, the reference oscillators made use of easily
generated frequencies. Manufacturers often picked frequencies that
could be easily generated with multipliers and small dividers. As I
recall, the C-field was often chosen to make the tail-end digits of the
Rb frequency ....610.904333 Hz end up at something related to (as I
recall) 7/16kHz = 687.50000 Hz because they could generate a fixed 7/16
generator with non-programmable logic.
BTW -- I stumble on a pretty good (albeit with commercials) description
of how different atomic standards work here
1. mailto:rlutwak at comcast.net
More information about the time-nuts