[time-nuts] Why 9,192,631,770 ??

[Tom Van Baak]

> Thanks downloaded the paper. I assume the hyper fine transition happened to
> be the one nearest that frequency? You can tune to either side of the peak
> and get a lock on the next transition.
> Regards
> Paul.

Hi Paul,

The definition is the second specifies which hyperfine transition to use; it also requires zero temperature and zero magnetic field. Since you can't actually achieve the latter two in a beam tube, good cesium standards accurately compensate for the slight frequency shift with synthesizers.

Yes, when you apply the magnetic field you do get multiple peaks. You'll see this in most technical descriptions of how cesium clocks work; see this old one by hp:
http://leapsecond.com/museum/hp5062c/theory.htm

The center peak is the one to use because it's most immune from changes in magnetic field and also the tallest, sharpest peak. This is clear in the energy level diagram. You can also see each of the 7 peaks in detail here:
http://leapsecond.com/pages/cspeak/
Or the tall wall-poster(!) version:
http://leapsecond.com/pages/cspeak/image002.gif
John Miles joined the 9192631770 club:
http://www.ke5fx.com/cs.htm

Yes, older cesium standards would allow you to deliberately (or accidentally) lock to the wrong peak. In that case the clock runs many ppm fast or slow. Easy to detect.

The difference between the peaks is the "Zeeman frequency"; it allows one to cleverly indirectly calibrate the C-field so that the cesium standard more accurately ticks SI seconds. Modern cesium standards (e.g., 5071A) do this automatically. Here is a tall plot showing how much the other 6 peaks shift when the magnetic field is manually changed (e.g., hp 5061A):
http://leapsecond.com/images/cfield.gif
For more information on the Zeeman frequency and cesium beam tubes goggle: zeeman frequency site:febo.com

/tvb