[time-nuts] How Rubidiums make their frequency

[Poul-Henning Kamp]

In message <ad9f78230604191255m6deb020arf3bb3b583d385782 at mail.gmail.com>, "Matt
 Ettus" writes:

>Since we can now make DDS's with arbitrary frequency resolution, could
>you make an Rb oscillator without the magnetic field adjustment? 
>Wouldn't that reduce a source of error in frequency?  Then we'd be
>left with the ideal resonance frequency, right?
>
>Are there any other influences on the resonance frequency?  I assume
>temperature and density don't matter.

In fact density/pressure does matter and is one of the major reasons
why rubidiums drift:  Rubidium is absorbed into the glas container
and as the pressure drops the frequency pulls.  The absorption
also makes the glas darker and darker, being a major wear-out
mechanism for Rb units.

As far as I know, this is why Rb is never classified as a primary
standard:  A drift-free unit has yet to be constructed.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.