[time-nuts] Thunderbolt reception problems

WarrenS warrensjmail-one at yahoo.com
Tue Nov 24 06:09:25 UTC 2009


Tom said

> I'd worry some about a TC that's too long. There's also the issue 
> of the correct damping factor to match the TC you use.

Too true, this started out to be just an antenna feed issue.
and it sounded like Said wanted to keep it simple.
I did not know this was going to be a Phase face off, 

The 500 TC AND a 1.2 damping AND a DAC_Gain that is not set correct
will very likely be too long for best phase long term error.  
It is more optimized to keep the short term Osc Freq noise down due to low signal levels.
WITH the correct Dac_gain and Damping, a TC of 500 should be fine, 
IF he has a good controlled temperature environment.  

Note the plot does not look like the typical Too Hi of TC drift.
On his plot, something is causing some pretty quick Phase jumps every 3 hrs or so, that are them  being pulled back in. 
I'm  thinking  temperature is the most likely cause, 
which a faster TC will lower the PEAK to Peak by turning some of the long term noise into lower level higher freq noise.

Damn always a trade off to made somewhere.

ws

**************

> Hi Tom,
>
> in progress now PRS-10 against a FireFly-IIA...
>
> The PRS-10 is set to a very long time-constant, more than
> 8 hours if I remember correctly.. The Thunderbolt is set to
> 500s TC.



That could be part of the problem. 8 hours seems really long
to me for a PRS10. I mean, it would not surprise me to see
a PRS10 time drift by 10's of ns over 1/3 to 1/2 of a day. So
I wonder what a TC of just 1 hour would look like. A one-day
run using PRS10 alone (no M12 gps) would also settle it one
way or another.

You might also try a run with the TBolt TC = 100s to see if
that makes a difference. The symptom of too large a TC is
excessive wander (because you are giving the OCXO too
much say over the GPS engine).

Sorry if this just makes more work for you. But the matrix
of settings to try and the references to use is bounded. So
even if each run takes a day, you'll have real data within a
week.

Unless you're independently sure of your TBolt OCXO and
your sky/volt/temp/air lab environment, I'd worry some about
a TC that's too long. There's also the issue of the correct
damping factor to match the TC you use. Warren can explain
that better.

/tvb


More information about the time-nuts mailing list