[time-nuts] Re: Phase coherence with 2X GPSDO

[Bob kb8tq]

HI

> On Mar 8, 2022, at 2:31 AM, ed breya <eb at telight.com> wrote:
> 
> Just an interesting observation. Last night, shortly after I wrote about my dual GPSDO system, I decided to take a look at it, with the usual scope setup looking at both 10 MHz outputs. It looked like one of those reasonably good times, where the phase between them seemed very stable during fairly long observation time. They were nearly in quadrature, drifting very slightly back and forth, over maybe an hour of occasional observation. This is akin to the proverbial "watching paint dry" - nothing new to report, over a long time, blah blah blah.
> 
> I went inside for at most a couple minutes to look at something, and when I returned to the garage, I noticed they were nearly in-phase, all of a sudden. Over the next half hour or so, I could see the phase drift back to about where it was before. During that brief moment when I was gone, I had missed what was probably a DAC value update, when one of the units decided it was time to make a change.
> 
> The change of one or some LSBs must have had a relatively instantaneous effect in the time scale involved, and I missed it, but I did see the aftermath. So, you can add that sort of effect in the time/frequency situation. The temperatures and drifts of various parts, and the SW figuring out what to do about it, tend to go quite slowly, but once the decision is made to change a value, there's a step function involved.
> 
> As I understand, the tuning DAC may be updated once a second at most (if comparing to the 1PPS), in small increments, as the DO part tries to keep everything right according to the GPS part.

In the case of most GPSDO’s running with most firmware: 

The updates are indeed once a second. The math does it’s thing on each PPS
and comes up with a “new guess” each time. The delta with most DAC’s is in the
“several to many” LSB's range. A lot depends on what’s going on with the oscillator
and the rest of the “stuff”. 

While we tend to blame everything on the poor OCXO in the typical OEM module,
that may not be the problem. The Vref or even the tempco of the DAC could easily
be getting into the mix. There are some well known GPSDO’s that a lot of us have
where this is the case. 

Since I’m sure “why” will be the next question …. 

If the device is targeted at a very long lifespan (10, 20, ….. years ) the trim range is 
going to be non-trivial. Just how large it is depends a bit on how conservative the
folks who did the oscillator / GPSDO felt when they did it all up.  Could you use 
a $50 Vref and a $90 DAC? That’s really going to destroy your budget ….

Bob


> With my simple setup and observations, I can't tell really how big, or how often these adjustments are made, but whatever this was, it was quite obvious. The most interesting things tend to happen when you're not looking.
> 
> Ed
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com -- To unsubscribe send an email to time-nuts-leave at lists.febo.com
> To unsubscribe, go to and follow the instructions there.