[time-nuts] GPSDO question

Bob Camp kb8tq at n1k.org
Sun Jul 26 14:12:02 UTC 2015


Hi

Commercial GPSDO’s are manufactured to a specific customer requirement. The application in the system dictates how it behaves
as it it powered up or power cycled. There is no single “always correct” approach. 

In a Time Nut environment, people seem to be quite happy setting up temperature controlled areas that are stable to a small fraction 
of a C. They also seem to be able to set up power that never ever goes out. Both are quite different than the constraints on a 
commercial system. A GPSDO optimized for this sort of environment may be a very different beast. 

Survey performance varies from module to module. It also is highly dependent on your antenna and sky view. A survey with an antenna
on top of a 300 M tall tower will progress very differently than a survey on an antenna at the bottom of a well. Commercial systems 
are designed with a good antenna location (or not) as part of the spec. Time Nuts seem to often have antennas at the bottom of a well. 
A full survey on an older GPS may take weeks with a poor antenna location. The same process on a new(er) GPS and a good antenna
can complete in under 10 minutes. Thats a wide range.

The answer to the survey issue in both cases (commercial and Time Nut) is to lock the GPS in a fixed location mode. Do the survey 
however you can. Get the best location possible. Load that location into the GPS each time it boots (or save it in the module). That 
way there is no survey process to mess things up. You also *should* have a much better location this way. The auto-survey process 
rarely gives a really good location. 


> On Jul 26, 2015, at 12:27 AM, Bob Stewart <bob at evoria.net> wrote:
> 
> Speaking with Tom recently impressed on me just how much I don't know about what's "normal" for a GPSDO.  The options on the LEA-6T are formidable and have consequences for my GPSDO.  The question I'm struggling with now is what's normal during the survey process?  The only example of a commercial GPSDO I have is the KS-24361.

It’s a good example of an older GPSDO design that was done back when the GPS signal was being deliberately degraded by the DOD.

>  I believe it leaves the oscillator output available and both the FLL and PLL are functioning.

At least on the ones I have looked at, it continues to play with the output quite a bit for the first 12 hours and still is swinging things around 
over the next couple of days. 

>  I've set the bit in the LEA-6T so that it doesn't emit the PPS when the time isn't valid.  So, with the other restraints I've put on it (TDOP, TAcc, etc), there's no PPS during survey.  That means there's no FLL or PLL, and the DAC is frozen, so I might as well shut off the oscillator output.

That depends a lot on when you shut it down and how good a DAC setting you have stored in memory. 

>  On the one hand, a survey implies the OCXO has just been turned on and all the nastiness of early retrace.  

Or (more likely in a commercial environment) there has been a brief power outage and things are trying to get going again. 

> And the first several hours of these 34310-T OCXOs are horrible.

…. and how were those OCXO’s treated before you got them ….

>  OTOH, there's the KS which, by default anyway, ignores all that and does the best it can, and 
> does it right quickly.  I suppose it wouldn't be all that hard to add still more controls to allow it to work or not work as the user desires.  Any input on the state of the art would be appreciated.

Well, there are probably a dozen or so “controls” that you can set in a commercial environment. Time Nuts seem to like to fiddle. That would 
probably up the number into the several dozen range. There are a *lot* of corner cases (I can only see one sat a day for 10 minutes and want to do an auto survey …).
If you include them all, there really is no upper bound other than exhaustion on the part of the designer. Consider that (as others have 
noted), every time you add a user settable control, you also add the likelihood it will be set incorrectly. Each time it is set wrong, you
get a support call. 

So much fun….

Bob


> 
> Bob - AE6RV
> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.




More information about the Time-nuts_lists.febo.com mailing list