[time-nuts] You can build a fountain from the things you find at home...

Magnus Danielson magnus at rubidium.dyndns.org
Tue Sep 27 20:11:57 UTC 2011

On 27/09/11 21:46, Poul-Henning Kamp wrote:
> In message<CAP6i9MmntLTia=XXUafbyrp+e3=koc1TRSEvL=g05emepQ4v-w at mail.gmail.com>
> , brent evers writes:
>> So at its most basic, I'm wondering what type of clock would make the
>> most sense to consider - cesium fountain, or hydrogen maser?
> Based on what I can gather, a fountain is probably easier to make work
> than any of the other, from a pure _mechanical_ point of view.
> The important detail to remember here, is that the people who have
> built fountains so far, have all tried to get better performance
> than very mature hydrogen masers and high performance cesiums.
> If your goal is simply to make "something that works" you probably don't
> need 4 separate shields, niobium surface etc.
> That said, it's probably not a one weekend task.

A small starter would be a rubidium. Getting the glas-work with rubidium 
isotopes might be the show-stopper unless you work on it.

A cesium beam (or any other beam) would be possible, but require a bit 
more work and preparations. If we talk about getting it to work level 
then getting the selector magnets working properly might be a bit 
troublesome. Most other things should be doable for a handy mechanics 
with a good toolshop. Getting good performance would meet many 
challenges, but a "working" beam should be possible at least.

A hydrogen maser would be possible if you can tool such large 
structures. The glaswork might be a challenge and also the magnet assembly.

A fountain would be a nice challenge, and the physical package would be 
possible. It would take some effort to build the optical stuff and get 
it to work well. The RF section should be simpler than the beam designs.

A variant which might be of interest is ion traps. Should not need that 
complex mechanics even if it has some challenges.

Trying to approach any of these would require reading up on a large 
number of articles and patents to kind of "learn" the field. In alla of 
these, getting full performance would involved much effort.

Oh, would be fun to tinker with physical packages if someone would cook 
them up.


More information about the time-nuts mailing list