Hi Jawad, look at the query_gpsdo_sensors example; we recently (3.9.2) overhauled that to show how to portably and reliably set the device time to GPS time.
The point is that "set_time_source" is a bit of a misleading name: it actually sets the timing *pulse* source, i.e. where the USRP gets its PPS from. The GNU Radio USRP blocks set the device time to 0, iirc. Best regards, Marcus On 11.01.2016 11:42, Jawad Seddar wrote: > Hi Derek, > > I set my USRP source block to use the GPSDO as a Time source, then set > the Sync parameter to "Unknown PPS". > All it does is reset the USRP tilme to 0 on the PPS front, it doesn't > set the time to the GPS time. > > Is there anything else I can try? > > In the meantime, I reverted to UHD 3.8.5 which works for me. > > > Regards, > Jawad > > 2016-01-08 20:10 GMT+01:00 Jawad Seddar <[email protected] > <mailto:[email protected]>>: > > Thanks Derek, I will try that when I get a chance. > > I didn't try setting the "Sync" parameter to "Unknown PPS" but was > guessing it might have something to do with that. > I will give it shot on Monday. > > Regards, > Jawad > > > 2016-01-08 20:07 GMT+01:00 Derek Kozel <[email protected] > <mailto:[email protected]>>: > > Hi Jawad, > > I believe that setting the USRP Sink block's setting to: > Clock: O\B GPSDO > Sync: Unknown PPS > > is all that's needed but I do not have the hardware in front > of me to verify this at the moment. I hope someone will chime > in with confirmation or the correct settings, but I will check > this when I get a chance, but it may not be until Monday. > > Regards, > Derek > > On Fri, Jan 8, 2016 at 10:40 AM, Jawad Seddar > <[email protected] <mailto:[email protected]>> wrote: > > Hello Derek, > > Thanks for your answer. > > Unfortunately, this is the method I was referring to when > I was talking about setting the time manually in a python > interpreter (the only difference is that I was using the > functions from gr-uhd that translate to the method you > mentioned). > > Is there a way of doing this directly from GRC? > It used to work out of the box when using a usrp_source or > usrp_sink block from gr-uhd. > > > Regards, > Jawad > > 2016-01-08 19:31 GMT+01:00 Derek Kozel > <[email protected] <mailto:[email protected]>>: > > Hello Jawad, > > The UHD 3.9 release changed the default GPSDO > behavior. UHD does not automatically synchronize > device time to GPS time on initialization because it's > not reliable during device start-up. I'm not 100% sure > about GNU Radio's handling of these changes so will > have to check before commenting on the behavior there. > > Here is a method of reliably setting the time in your > own code: > 1) Poll on usrp->get_mboard_sensor("gps_locked") > until it returns true > 2) Poll on usrp->get_time_last_pps() until a change > is seen. > 3) Sleep 200ms (allow NMEA string to propagate) > 4) Use > > "usrp->set_time_next_pps(uhd::time_spec_t(usrp->get_mboard_sensor("gps_time").to_int()+1));" > to set the time > 5) Poll on usrp->get_time_last_pps() until a change > is seen. > 6) Sleep 200ms (allow NMEA string to propagate) > 7) Verify that usrp->get_time_last_pps() and > usrp->get_mboard_sensor("gps_time") return the same time. > > Regards, > Derek > > > On Fri, Jan 8, 2016 at 4:50 AM, Jawad Seddar > <[email protected] > <mailto:[email protected]>> wrote: > > Hi all, > > Yesterday, I decided to update from GNU Radio > 3.7.7.1 + UHD 3.8.4 to GNU Radio 3.7.9 + UHD 3.9.1. > > I used pybombs and everything went well, my OOT > modules compile and my blocks run just like they > used to. > > My issue comes from the time that is set on the > USRP. It seems like the time on the device is not > set to use the GPSDO. After doing some digging, I > noticed that this commit might be responsible > : > https://github.com/EttusResearch/uhd/commit/f4f3ce2550aba933b5f4d310d82a42a41ce3f6a1 > . > > So I tried setting the time source as GPSDO on my > USRP source block (in GRC) but that doesn't seem > to be working. I still get something (in the tens > of thousands) that is way below what it should be > (I get the time on the USRP by looking at the > rx_time tag available at the beginning of a > received sample stream). > > If I run a python CLI and setup a usrp_source then > set the time manually to the gps_time, then the > time is somewhat correct. > > Am I missing a step when setting up my usrp_source > in GRC? > > Thanks, > Jawad > > _______________________________________________ > Discuss-gnuradio mailing list > [email protected] > <mailto:[email protected]> > https://lists.gnu.org/mailman/listinfo/discuss-gnuradio > > > > > > > > > _______________________________________________ > Discuss-gnuradio mailing list > [email protected] > https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
_______________________________________________ Discuss-gnuradio mailing list [email protected] https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
