Re: [Discuss-gnuradio] Python synchronisation of 2xUSRP N200 with SBX in one UHD container.

[John Shields]

On 18/09/17 21:27, Marcus D. Leech wrote:
> On 09/18/2017 05:21 PM, John Shields wrote:
> > On 18/09/17 20:52, Marcus D. Leech wrote:
> > > On 09/18/2017 03:49 PM, John Shields wrote:
> > > > Thanks,
> > > >             I have attached the GRC. I should note that before I 
> > > > run, I wait until GPS is locked as evinced by the 
> > > > query_gpsdo_sensor utility:
> > > >
> > > >             GPS Locked
> > > >             Trying to align the device time to GPS time...
> > > >             GPS and UHD Device time are aligned.
> > > >             last_pps: 1505763709 vs gps: 1505763709.
> > > >             Printing available NMEA strings:
> > > >             GPS_GPGGA: 
> > > > $GPGGA,194149.00,4331.1729,S,17234.2461,E,2,09,1.5,32.2,M,8.5,M,,*7B
> > > >             GPS_GPRMC: 
> > > > $GPRMC,194149.00,A,4331.1729,S,17234.2461,E,0.2,0.0,180917,,*20
> > > >             GPS Epoch time at last PPS: 1505763710.00000 seconds
> > > >             UHD Device time last PPS:   1505763710.00000 seconds
> > > >             UHD Device time right now:  1505763710.01897 seconds
> > > >             PC Clock time:              1505763710 seconds
> > > >
> > > >             Done!
> > > >
> > > >             Kind Regards,
> > > >
> > > >                    John
> > > >
> > > > p.s. In Jan 2016, Nigel wrote that one of the steps needed to align 
> > > > included 'Use integer_N_ mode for PLL' when I tried to use the 
> > > > method you recommended, I got an error so ceased that line of 
> > > > investigation for the moment - is integer_N_mode tuning required?
> > > Integer-N will help, but my understanding of the ADF4351 is that the 
> > > re-synch feature, enabled through timed commands, will also work.
> > >
> > > Also, the "uhd_fft" command has a --phase-relations option that will 
> > > plot relative-phase between two input channels.  You might look at 
> > > that.
> > Thanks Marcus,
> >
> >                         I can see that uhd_fft has a phase-relations 
> > option but, as far as I can tell from looking at the source, it 
> > doesn't synch the channels so, just to make sure I understand the 
> > point of your e-mail, uhd_fft uses a QT widget to display relative 
> > phase and I could investigate modifying my FG to use that method to 
> > display relative phase between the two channels?
> Yes, or add features to uhd_fft to do full phase-synch dancing.
>
> It already has --clock-source --time-source and --sync   options. You 
> could modify it to add the timed-commands thing.
Thanks, sorry for being so thick!

          Slainte,

                      John