[time-nuts] GPSDO for a week

[Bruce Griffiths]

Tom Van Baak wrote:
>> Thanks a lot for taking the time to do these measurements Tom.
>>
>> If I am interpreting the data correctly the ADEV stays below
>> 3E-(8+tau) out to tau=1000s and is above that for tau=10000s. So the
>> Thunderbold is indeed not suitable for VLBI at 40-50MHz with
>> integration times longer than 1000s.
>>     
>
> I'm not sure what you mean by the expression: 3E-(8+tau). For
> example, if you plug in tau = 1000s into that equation, you get
> 3e-1008, which isn't right.
>
> Also how does this compare with your earlier requirement of
> 3 ns at 1 second (which would be 3e-9). I'll help, but I'm a bit
> confused what your performance goal is and how that relates
> to 50 MHz.
>
> /tvb
>
>   
Tom

Try 3E-(8 + log_10 (Tau)).

The requirement should be something like:

1) The rms phase error between independent LO pairs should be < 1 radian
for the duration of the signal integration.
This requirement is akin to avoiding carrier collapse of signal when the
sideband noise power exceeds the carrier power.

2) Ideally the LO phase noise should be less than the ionospheric
contribution (however only have one data point for the ionosphere  ADEV
~8E-11@ 10s).
Presumably this is the contribution at night as the ionosphere TEC rises
significantly during the day as does the ionospheric delay and its
fluctuations).

If the delay through the ionosphere is plotted as a function of time,
the dawn wave where the TEC starts to rise can clearly be seen.
Periodic disturbances of TEC can also occur at this time.

Bruce