[time-nuts] Some results of Tbolt + LPRO case

Bruce Griffiths bruce.griffiths at xtra.co.nz
Fri Jan 30 00:10:18 UTC 2009


Esa Heikkinen wrote:
> Hi!
>
> Although this will not be my final setup the measurements had already 
> started with "Tbolt OCXO replaced by LPRO" and are now done. I think 
> some of you might be interested of the results.
>
> Here's some Lady Heather logfiles and couple of pictures.
>
> Unmodified Thunderbolt in my place (many trees at the yard etc):
> http://www.amigazone.fi/files/gpsdo/tbolt-default.log
> http://www.amigazone.fi/files/gpsdo/tbolt-normal.jpg
> (Includes short holdover test at the end)
>
> Unmodified Thunderbolt in holdover mode:
> http://www.amigazone.fi/files/gpsdo/tbolt-holdover.log
>
> Unmodified Thunderbolt OCXO performance (disciplining off):
> http://www.amigazone.fi/files/gpsdo/tbolt-ocxo-freerun.log
>
> OCXO now replaced with LPRO, normal run:
> http://www.amigazone.fi/files/gpsdo/tbolt+lpro.log
>
> OCXO replaced with LPRO, disciplining turned off:
> http://www.amigazone.fi/files/gpsdo/tbolt+lpro-freerun.log
> http://www.amigazone.fi/files/gpsdo/tbolt+lpro-freerun.jpg
>
> This final test simulates my planned setup (without final cleaning loop 
> however).
>
> Next thing is to try to get some ADEV charts with Ulrich's Plotter. For 
> some reason it shows only e-1 values for now, the actual shape of the 
> trace looks ok. Have to learn about the settings...
>
>   

Esa

The first thing you have to do when using PLOTTER is to scale the phase
values (they are in nanoseconds).
Simply set the coefficient of y to 1.000E-9 in the scale dialog.


With only 20,000 or so phase values the estimator OADEV degrades
considerably in accuracy for tau > ~2000 sec or so,
TOTDEV is better in this respect.
THEO_1 and its variants are even better estimators in that they make
more effective use of the available data.

Bruce



More information about the time-nuts mailing list