[time-nuts] 5061A and Zeeman freq question
magnus at rubidium.dyndns.org
Sat Jul 19 19:21:49 EDT 2008
Predrag Dukic wrote:
> I am not a quantum physicist, but my understanding of Zeeman
> splitting is a bit different.
> With applied homogenous DC magnetic field to the atoms on their path
> You get splitting. Without it You have no Ramsey fringes.
> There is a formula to calculate mid-fringe offset from defined cesium
> resonance knowing splitting frequency.
Not only that, there is a formulat to describe the shifts of all 7 Rabi
pedistals. The "sidebands" depends on the square of the field. The
relation between the sidebands separation on the C field and its
relation to the shift in frequency of the mid fringe is used on modern
Cs beams to create a C field lock loop and autonmous triming, causing a
much reduced systematic drift effect. See FTS 4065 manuan for instance.
The Zeeman splitting is an arbitrary design aspect and the ease of
designing the frequency synthesis controls the selection of nominal C
field and thus Zeeman splitting in early designs. With the modern high
resolution DDS chips, a more complex method for controling the system
So seek these magic numbers elsewhere than the pure physics, althought
they do relate to it.
> So You should always be able to tell the right frequency from the
> splitting frequency, and that has nothing to do with definition of
> the cesium as primary standard.
> The correct field strength and correct splitting frequency, by my
> opinion have some importance to other parameters; stability, mid
> fringe width etc,
> while resonance frequency is tied to split frequency by an exact
> mathematical equation.
With the ideal case of exactly 0 as C-field.
> All of this under assumption that the field is homogenous. If not,
> than You have some other problems.
More information about the time-nuts