[time-nuts] GPSDO TC & Damping

Magnus Danielson magnus at rubidium.dyndns.org
Sat Jan 10 09:49:00 UTC 2009


Dick,

> Sorry, Magnus, I didn't mean to put words in your mouth.

No worries, I just did not want it to be raised to "truth".

> I was remembering last Fall, when you suggested that people look at your  
> ADEV plot and to be mindful of the of the bounding slopes of the  
> curves. If I've mis-remembered the emphasis or the facts, I do  
> apologize. I thought your argument then, as I remember it, was strong  
> and valuable. Seems like it gave a good range of possible values to  
> use for Tau in the measurements.

Do not recall the particular discussion. A ref into the archives or date 
would be apprechiated.

> I probably won't go far beyond the capabilities of the TBolt, such as  
> implementing a PID controller with dynamic control of variables using  
> a microcontroller and LPGAys and writing my own software, but I love  
> it when you talk that way.

Sounds dangerous. :)

I have been looking further into the variance, ADEV, MDEV and TDEV with 
respect to the effect of various dampings. It is becomming clear to me 
that the gain effect which can be attributed to damping will ripple over 
to all those measures. I have not had the time to convert this to a 
complete analysis, but I think I can do that.

However, to give a partial answer to Bruce, I would like to point to a 
passage in Gardiner where the optimum damping for minimum flicker noise 
is being found, and that is for 1.14 rather than 0.5 which would be 
minimum noise bandwidth. Gardiner does not go into ADEV and friends, but 
does provide the refernces and gives an indication of convergance 
problems for flicker noise on variance, also indicating that this is a 
problem for further analysis. Using ADEV and friends in replacement 
analysis could be made analogously for other noise types and as it 
happends also work in cooperation with investigating the effect on 
various noisetypes.

Regardless, it is an interesting little problem and I get to exercise 
the little grey and refresh my always failing abilities in integration.

Cheers,
Magnus



More information about the time-nuts mailing list