[time-nuts] stabilities of atomic clocks (was: Beginner's Atomic Clock)

[Attila Kinali]

On Tue, 17 Sep 2019 07:03:15 -0700
"Richard (Rick) Karlquist" <richard at karlquist.com> wrote:

> What I was always told was that Rb has low enough noise coming off
> the atoms such that it can use a relatively fast loop to lock
> the OCXO and clean it up.  OTOH, the 5071 Cs has to use a very
> long time constant loop to control the 10811, hence it doesn't
> clean it up except very close to the carrier.  Even the high
> performance version of the 5071 doesn't compare with Rb.

This is bascially the big point about why atomic clocks perform
differently. The SNR per interogation is the driving limit for
short taus. Rb vapor cell standards have a LARGE number of atoms
that are interrogated (IIRC millions) with a relative high detection
peak (depending on the exact built type, between 1% and 30% above
background noise level).

On the other extreme are single ion optical clocks, which interrogate
a single atom only. The SNR is so low that anything below minutes of
interrogation is not going to do anything. That's why optical clocks
are so tied to the development of high stability cavities (crystaline,
low expansion spacer, cryogenically cooled with special, low thermal
movement, low creep mirrors, in vacuum) to get a laser that is stable
over these interrogation times.

Somewhere in-between are dual-resonance (ie same principle as Rb vapor cells)
Hg Ion clocks (i.e. the JPL Hg ion clock, or DSAC). They interrogate a fairly
large number of atoms (10'000s of atoms), but have such a low photon efficiency
(i.e. probability of interaction with an atom), that the detection SNR is very
low. Hence the need for interrogation times in the 10s to 100s of seconds.

The one atomic clock that does not fit in this categorization scheeme
are active hydrogen masers. These do not interrogate atoms, but cause
atoms to emit coherent photons. Or in other words, they make the atoms
themselves into oscillators. That's why the noise of AHMs goes down
with tau (at short taus) while for all other, passvie atomic clocks
it goes down with square-root of tau.


At longer taus, the stability of the environment of the atomic clocks
becomes the dominant factor. This is pretty bad for Rb vapor cells, 
as they are sensitive to temperature and barometric pressure changes.
AHMs suffer from changes in the wall (where atoms bounce off) and 
of pressure deformation of the large and resonant cavity. Other types
of atomic clocks can be made pretty insensitive to enviroment by
being in vacuum and by suspending the elements such, that pressure
changes do not move them (much).

			Attila Kinali
-- 
It is upon moral qualities that a society is ultimately founded. All 
the prosperity and technological sophistication in the world is of no 
use without that foundation.
                 -- Miss Matheson, The Diamond Age, Neal Stephenson