[time-nuts] Thunderbolt performance in holdover mode

[SAIDJACK at aol.com]

Hi There,
 
this is quite a good correlation I would say. I am surprised that they use  
an external temp sensor, not internal to the OCXO.
 
4.5us drift over 24 hours is pretty good though.
 
bye,
Said
 
 
In a message dated 3/12/2009 12:07:43 Pacific Daylight Time,  
holrum at hotmail.com writes:


Here  is a clean plot of a Thunderbolt in holdover mode.  The graph is 1 hour 
 per major division.  You can clearly see the dependence of the DAC  voltage 
on temperature.  

You can also see that the unit holds the  PPS signal within 150 nS for the 
first eight hours then it starts going  wild.  Not shown on the plot is that 
after 11 hours the PPS signal was  off by 1 uS and at 24 hours it was off by 4.5 
uS.  The PPS error from 8  to 24 hours was a fairly straight line at 200-250  
nS/hour