[USRP-users] Sychnronize two USRP X310 with GPS

Ian Buckley ianb at ionconcepts.com
Fri Dec 19 13:26:53 EST 2014

A few questions that will help in diagnosing what you are seeing:

How long have you had the X300's continuously powered (with GPS antennas connected and a good view of the sky) before you make these measurements?

Can you describe how you determine they are out of sync? It sounds like you have them at 2 remote locations?

Are you using the latest in UHD and FPGA images? 

The fact that you see the same rx-time stream tags is a good sign that you are using the API correctly to perform the capture. The fact that you are within 20-100uS tells me that UHD has tried to initialize the internal VITA timestamp to a UTC related time, otherwise this value would be a count of 200MHz clock ticks since power was applied to each unit.


On Dec 19, 2014, at 5:10 AM, Simon Hartmann via USRP-users <usrp-users at lists.ettus.com> wrote:

> Hi all,
> the last weeks we tried to synchronize our two usrps X310 between two stations with two seperate GPS antenna, but the results weren't as good as we hoped. 
> For your information, every usrp is connected to his own pc on which the same code is running.
> Our first intent was to check if the gps antenna is locked (get_mboard_sensor('gps_locked')) and then start to receive samples with the same time given (set_start_time(uhd.time_spec_t(self.start_time))). We only want to receive samples, therefore we used set_start_time() instead of set_command_time().
> When we checked the sync, there was a difference about 20 - 100 us. According to the documentation only ~100 ns should be allowed.
> We also read out the rx-time stream tags for both measurements and compared them. However, they were identically.
> In the documentation of Ettus it's explained, if the GPS is locked to the satellites the internal VITA timestamp will be initialized to the GPS time, and the internal oscillator will be phase-locked to the 10MHz GPSDO reference. 
> So if their both locked they should have the same timestamp, or? Or do we need some further code to initialize the gps time to the usrp?
> What else could be the problem?
> Thanks,
> Simon
> _______________________________________________
> USRP-users mailing list
> USRP-users at lists.ettus.com
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

More information about the USRP-users mailing list