[time-nuts] Modern Rb atomic reference vs classic Cs

[ew]

A couple of years ago Time Tech was selling a OSA 8607 GPSDO with an AV of 8 E-14. AXTAL has one. Takes a lot of characterization, fully understand and constantly compensate the OCXO and continous GPS monitoring for more than 24 hours. W. Schaefer has the smarts. Does not have the aging  of a Maser.Not having their smarts I am trying to do something close using aging compensation along with pressure and temperature on a FRK/M100 Rb and use a 8607 for clean up. Will use multi day GPS averaging for fine control. Presently running 4 Rb to pick the best. Use HP5065A for AV and Datum RS 2000 for frequency ,does 24 hour GPS averaging. Have 5061B with new small tube.
Bert Kehren


In a message dated 3/17/2020 4:52:23 PM Eastern Standard Time, attila at kinali.ch writes:

On Sat, 14 Mar 2020 14:50:45 -0400
Bob kb8tq <kb8tq at n1k.org> wrote:

> A brand new 5071 with that tube appears to be over $90,000 these days. Based on 
> running one for a lot of years, After about 6 years of operation, you will be sending 
> it back for a new tube. By the time the unit is back in your lab, the refit will cost 
> roughly half the price the whole unit cost new …. gulp …..

At those prices, I'd rather go for a µQuans or SDS Rb clock.
Those don't lose atoms like the Cs beam does and thus don't need
a refill. Their lifetime is more likely in the decades than just
a few years. Weakest link, as far as I know, are the lasers.
And yes, after the second, at latest after the third Cs tube,
these Rb devices are cheaper. And they are as much a primary
standard as the 5071 is.


On Sat, 14 Mar 2020 13:57:21 -0700
Hal Murray <hmurray at megapathdsl.net> wrote:

> Can the physics-nuts calculate the Rb frequency relative to Cs?

Theoretically yes, practically no. While we "know" exactly what
each electron in the atom is doing and how it interacts with each
other electron and the nucleus, there are so many of them that there
is no closed form solution (c.f. three/many body problem). As far
as I am aware of, nobody has even done a complete numerical model.
So all the calculations we have today are approximations that
bunch most if not all inner electrons together and approximate them
by an average field.


> What's missing on a gas cell?  Is the problem theory or implementation?

Both. For one, you cannot control all parameters during production
well enough that a calulated shift would be any more accurate.
For another, there are quite a few shifts in the system, from
buffer gas shift (dependent on exact composition) to light shift
to RF power shift to RF field gradient shift to temperature gradient
shift to ..... Some of them nobody thought of until a PhD student
tried to figure out what the next limiting factor was.

If you would try to make the system as well characterizable as
a Cs beam standard, then you would end up with a product that is at
least as expensive, if not more. If you want to know more, have a look
at the publications, especially the dissertations, by the time/frequency
lab at UniNE: https://www.unine.ch/ltf/home/publications.html
They go into quite detailed analysis of what the different shifts are
that are affecting the long term stability of Rb vapor cells.


            Attila Kinali

-- 
<JaberWorky>    The bad part of Zurich is where the degenerates
                throw DARK chocolate at you.

_______________________________________________
time-nuts mailing list -- time-nuts at lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.