[time-nuts] Primary standard again

Magnus Danielson magnus at rubidium.dyndns.org
Wed Feb 24 20:02:47 UTC 2010


Rick Karlquist wrote:
> If you want to get technical, the frequency of a cesium standard
> also depends on the gravitational acceleration, but for relativistic
> reasons, not newtonian physics.  Any decent cesium is accurate
> enough that it will noticably speed up at NIST in Boulder.  NIST's
> best clocks speed up noticibly if you move them from the ground
> floor to upstairs.
> 
> The pendulum is also not primary because it's length has to be
> calibrated.  A cesium kit does not require any calibrated parts,
> including the microwave cavity, which only has to be symmetrical.

One important aspect of a primary standard is that you want to achieve 
repeatability of realization. This involves avoiding calibrating to some 
arbitrary point. Caesium clocks have been realized in many very 
different forms, but achieves this to various degrees.

Another aspect is the accuracy. For a clock to be meaningfull in a 
calibration chain, it needs to realize an accuracy such that it is 
meaningfull to calibrate other clocks to it.

A further important aspect is the stability, as it will affect the 
calibration measures taken and usefullness for the purpose.

Then, on top of that, it is a gentlemens agreement on what is wise to 
use for common reference for such a realization. Many primary standards 
definitions would be possible, but in the end having a common one that 
everyone agrees upon is also important.

I think it would be challanging to define it more precisely than that. 
We know what it is when we see it. :)

Cheers,
Magnus




More information about the Time-nuts_lists.febo.com mailing list