[time-nuts] Trimble Thunderbolt setup optimizationand addinga10811 OCXO

[Arnold Tibus]

John,

>>
>> I put in as K0 what I measured several times with the 10811
>> as 'gain' with -0.74 Hz/V. Is that not the right way?

>Yes, that's fine.  Because the varactor is nonlinear and a small part of the
>overall tank circuit C, the tuning gain depends on the adjustment-screw
>setting to some extent.  Specifying it within a factor of 2x is reasonably
>OK, but it's always better to measure the actual gain by adjusting the
>tuning-DAC parameter manually with the output centered at 10.000... MHz.


I had checked for the 0 offset point  for 10.0 MHz down to 1 µV and 
stepped from this point some µV and mV in both directions away and did  
calculate the gain in Hz/V. Yes, I saw the unlinearity at higher values.
It is a bit difficult to find the correct value because the fluctuations,
I could define finally the Kosc as -0.74025V.

>> I tried to
>> use several times also higher values. The result was not impressiv,
>> did not give clear signs either.
>>
>> A factory reset I did not perform yet, but I am trying at the moment
>> the proposed 0.2 or even a bit lower values. Tomorrow I will
>> check the result, but up to now I do not have a good feeling.

>I just pulled my 10811-equipped TBolt out of storage after several months,
>and it powered up in a rather confused state.  Before I did a factory reset
>on mine, the reported PPS/10 MHz deviations were entirely fictional.  It was
>reporting reasonable values, but in reality, it wasn't steering the
>oscillator closer than a few parts in 1E-8.  The factory reset cycle fixed
>the odd behavior (note that you have to re-enter the tuning-sensitivity
>parameter after doing that).

I made a reset with the Trimble program to the factory values and 
made then a new self survey. The result does look similar as before 
with the Trimble std. parameters, the oadev values better as with 
the original OCXO. But all the values do look worse as the oscillator 
alone - looking specially for tau between 1 and 500. 
Then I corrected the Kosc. to -0.74 Hz/V and increased the time const. 
to 600 and lowered the damping factor somewhat. 

Looks like the values are coming closer to the origin osc. amounts 
but it is very slow and the pps offset does not seem to go down 
(shifting manually with the delay input is not the right way).

How can one find out the influence of the correct antenna position 
with that movements of EFC (250 µV in 15 min and sudden jumps 
of 100 µV due to a SAT change) ?
Perhaps the Thunderbolt is not designed to gain from better 
OCXOs?
Does the TB perhaps ' learn', but very slow (what I cannot imagine)?

I think that any influences due to antenna types and small position 
changes of the antenna are not of practical importance under these 
conditions. I never saw a difference when run in original version 
changing between special Irvine, HP,  QF-Helix or cheap Patch Antennas.

How do you realize theset influences?
What are your results for the TB with the 10811 on EFC, PPS and
Osc.? 
What setting are you applying?
I am keen on to see the values from other TBs, there must already be 
hundreds around. running ...

Regards
Arnold