[time-nuts] Re. DIY atomic "resonator"

Magnus Danielson magnus at rubidium.dyndns.org
Tue Apr 11 12:22:45 EDT 2017



On 04/11/2017 05:54 PM, Richard (Rick) Karlquist wrote:
>
>
> On 4/11/2017 12:31 AM, Andre wrote:
>
>> Has anyone else either built an atomic clock around a bare Rb lamp
>> module "core" or attempted
>>
>
> Not a DIY project, but I was the RF designer on the HP 10816 rubidium
> standard, which never made it to product introduction (a half dozen
> working pilot run units were built in 1982).  I would say this task is
> probably beyond the scope of a DIY project, at least for most
> time-nuts.

Probably right.

> The Rb lamp drive circuit (particularly getting the
> lamp to light up) is very challenging.

There would be none. It would be replaced by a laser. That has its own 
set of "problems" but different.

> The step recovery diode multiplier is very challenging.

Today you would not go the SRD route in synthesis.
Besides, SRDs can be hard to find these days.

> The photodetector and loop integrator is non trivial.

I'd expect the loop integrator to be done in digital, which eases up on 
some of the design problems.

> The synthesizer is the one thing that is easy in 2017.

Indeed-

 > The oven is also no simple thing to get
> low tempco.  Unlike a crystal, you have a lot of heat being
> generated by the lamp, etc. The lamp needs to be at a different
> temperature than the other cells.

Going down the laser-route, the balance of temperature between the cells 
is no longer a relevant problem. Further, the lamp and its heat is gone.

> You have to keep the tip
> off at the lowest temperature to keep the Rb in place and not
> "flood" the cell and block the light.  Etc., etc.

You still have this problem, but not as a lamp problem but only for the 
resonance cell part.

> This is in the category of projects where if you were qualified
> to do it, your time is far too valuable to do it for the amount
> of money you would save.

This is the type of project you do not to save any money, but to spend 
and learn.

Even if done in a much more modern fashion, avoiding several of the 
traditional problems, there is plenty of issues to solve and handle.

Cheers,
Magnus


More information about the time-nuts mailing list