[time-nuts] Trueposition behavior

[David G. McGaw]

I think it would be a dead-band.

David N1HAC

On 3/10/20 2:41 PM, djl wrote:
> Is this effectively dither on the control voltage?
>
> On 2020-03-10 10:30, Skip Withrow wrote:
>> Hello Time-Nuts,
>> A couple of weeks ago I posted regarding some weird behavior of the
>> Trueposition GPSDO that I was seeing.  I have now been able to get
>> back to the problem and have further results to report.
>>
>> I have several of these GPSDOs that are version 12.0.1 (I believe the
>> latest is 12.1.1) and they all seem to behave the same.
>>
>> I have attached two Lady Heather plots.  One is the GPSDO output of
>> the control voltage, the other is the 10MHz output monitored by
>> another NTBW GPSDO.  The bumps in the control voltage are what was
>> bothering me.
>>
>> It appears that the way this GPSDO operates is to let the oscillator
>> free run until there is 50ns of accumulated time error (LH reports
>> 'Normal' during this phase),  Then it goes into control mode, you can
>> see more noise on the DAC signal during this time plus the big spike,
>> and yanks the phase back to zero (LH reports 'Acquiring' during this
>> phase).
>>
>> The unit must model the OCXO drift as the DAC does change during the
>> drift period.  So, the longer the unit runs the longer between upsets
>> (hopefully).  These plots were taken shortly after start up, I am in
>> the process of letting it run for a while now.  BTW, the very periodic
>> dropping and acquisition of the WAAS bird (PRN138) is still present.
>>
>> This behavior is a good news/bad news situation.  On one hand, during
>> the drift period the output of the GPSDO is not influenced by GPS and
>> getting yanked around second by second (like the Thunderbolt). On the
>> other hand, if the 50ns yank happens when you are making a measurement
>> it kind of trashes things.  I haven't tried seeing if LH can disable
>> disciplining (for use during measurement periods).
>>
>> Regards,
>> Skip Withrow
>>
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at lists.febo.com
>> To unsubscribe, go to
>> https://nam12.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists.febo.com%2Fmailman%2Flistinfo%2Ftime-nuts_lists.febo.com&data=02%7C01%7Cdavid.g.mcgaw%40dartmouth.edu%7Cbb722bbdb01b4db9fe0308d7c522c000%7C995b093648d640e5a31ebf689ec9446f%7C0%7C0%7C637194625354501923&sdata=Bx2snhE2Wbc0Fo7MNo9kcsQ61JNeycC%2FEt90tEA657k%3D&reserved=0 
>>
>> and follow the instructions there.
>