[time-nuts] Weird GPSDO behavior

Charles Steinmetz csteinmetz at yandex.com
Sat Sep 30 02:29:28 EDT 2017

Skip wrote:

> I have a NTGS50AA GPSDO (close cousin to the NTBW50AA and Thunderbolt)
> with the OCXO removed and a SRS PRS-10 rubidium oscillator in its place.
> Once per day (about 8am) something disturbs the system.  So, the GPSDO
> spends much of its time recovering and never gives me anywhere near
> the performance that this system is capable of.  I would think that it
> is not the PRS-10 as it has no knowledge of time.  I would also think
> that it is not the GPS system or receiver, since the GPS constellation
> repeats twice per day.

I'm interested in the temperature trace.  If I'm reading the plot 
correctly, it appears that it starts oscillating up and down about 0.5C 
6 or 7 hours from the beginning of the plot, then around the same time 
as the DAC voltage steps it shoots up 5C in 6 or 7 hours with a peak up 
to +6.5C, then falls back to near where it started over ~15 hours, at 
which point the cycle repeats.  According to what we can see, it appears 
that this may be a two-day cycle, with an amplitude of 2 or 3C on "odd" 
days and ~6C on "even" days, but there isn't enough data here to say 

Is this an environmental effect, or self-heating within the '50AA?  6.5C 
seems like an awfully big temperature shift for something indoors, 
unless it is in direct sun. I'm not sure how it relates to the DAC and 
PPS changes, but it seems to be on the same schedule so it appears to be 
related somehow.

In my experience with Tbolts, I have found that damping of 0.6 is WAY 
too low for good performance generally, and *especially* at time 
constants >500s.  I suggest you try damping of 10 or 20.  I'm not 
suggesting that this will fix your problem, but it may, and if not, even 
after you fix what's wrong I'd advise at least trying it.  You can 
always go back.

Finally, you may get better results by putting the OCXO back into the 
GPSDO and then feeding the GPSDO PPS into the PPS synch input of the 
PRS10, rather than having LH try to discipline the PRS10 directly.  I 
have had superb results that way, after playing around with the PRS10's 
internal disciplining routines.  The PRS10 manual gives the relevant 

Best regards,


More information about the time-nuts mailing list