[USRP-users] GPSDO loss of lock
stoledo at tau.ac.il
Sun Apr 14 14:07:14 EDT 2013
Sounds good. I also assumed the the GPS frequency lock has a slow loop
filter, but it's good to get a verification of this. Thanks a lot Marcus.
On Sun, Apr 14, 2013 at 8:55 PM, Marcus D. Leech <mleech at ripnet.com> wrote:
> On 04/14/2013 01:36 PM, Sivan Toledo wrote:
> The UHD version is UHD_003.005.001. Not sure about firmware version
> because we did not write down which version we uploaded to the N200. Is
> there a way to tell? Should we upgrade to the latest firmware/FPGA image?
> Testing those values so frequently is going to cause unnecessarily-high
> control traffic. Unless things are *broken*, there's absolutely no reason
> to test them so frequently. Every few seconds perhaps, but every 64
> packets? That's a very high rate. And the GPS_LOCK sensor mechanism
> within the GPSDO probably doesn't even operate over such short time
> intervals--like I said, it's taking a "long view".
> The VCTCXO is your typical GPSDO is designed to provide fairly-long
> "holdover" times, and the servo mechanism tends to be able to cope with
> only-occasional "good" fixes from the satellite constellation. Again, I
> can't comment on why your GPSDO is reporting loss-of-lock so frequently,
> but it likely won't have a huge impact.
> On Thu, Apr 11, 2013 at 6:14 PM, Marcus D. Leech <mleech at ripnet.com>wrote:
>> We are trying to accurately measure time-of-arrival in two identical
>>> USRPs (N200 with WBX), both with the built-in GPSDO. We see differences in
>>> time-of-arrival between the two receivers, even at excellent SNRs, and we
>>> are trying to track the reason for this. Both receivers are connected to
>>> the same antenna through a T connector with identical cables.
>>> One thing that we checked is the GPS lock, the LO lock, and the REF
>>> lock. We test for all 3 every 64 packets of samples we get from UHD. We
>>> found that the LO and REF are locked all the time, but the gps_locked
>>> sensor reports loss of lock every 2 to 3 seconds. It then locks and unlocks
>>> a couple of times (usually 4) in quick succession, then locks for 2-3 more
>>> seconds, and so on.
>>> The GPS antennas seem fine and the GGA sentence reports that the
>>> receiver sees 11 satellites. So I don't think that flaky GPS reception is
>>> the problem. This happens to both units.
>>> We have 2 questions:
>>> 1. Why does the gps_locked reports no lock for a short period every few
>>> 2. How does this affect the 10MHz clock to which the sampling clock and
>>> the LO are locked? Can GPS loss of lock cause a significant drift or
>>> increased phase noise in the 10MHz reference and the clocks that are locked
>>> to it?
>>> Thanks, Sivan
>> I can't tell you why the gps_locked sensor is coming and going.
>> But the oscillator on board the GPSDO is of high quality, and the servo
>> in the GPSDO takes a "long view", so loss-of-lock won't necessarily cause
>> any really bad issues with clock stability, as long as it's getting
>> good timing data on a regular basis.
>> Which version of UHD and firmware are you using?
>> Marcus Leech
>> Principal Investigator
>> Shirleys Bay Radio Astronomy Consortium
>> USRP-users mailing list
>> USRP-users at lists.ettus.com
> Marcus Leech
> Principal Investigator
> Shirleys Bay Radio Astronomy Consortiumhttp://www.sbrac.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the USRP-users