Use "unknown PPS" on both of them. The MIMO cable shares both REFCLOCK and 1PPS signals, so they will both be synchronized to the same time.
On 2017-09-12 16:13, John Shields wrote:
Thanks Derek, No, I hadn't been power cycling between the runs - good point, obviously, I should have.
In terms of the 10 MHz and 1 pps references, in the configuration I was testing, I don't believe so in that I had the MIMO cable disconnected. My strategy was to have 2 USRPs with no MIMO - expecting little synchonisation. Then I was going to add the devices into the same container and connect the MIMO cable and expected things to improve and lastly, I was going to hand-code the SBX phase synch code.
In terms of the version of UHD, the fg shows: <<< Welcome to GNU Radio Companion 3.7.11.1 >>>
Thanks Marcus, I will implement your way of measuring the running phase offset and also thanks for correcting my understanding of O/B GPS .
In terms of getting the devices in the container to be the best synch they can be, I presume for the device which has the GPS, for the clock source and time source, I would put O/B GPS for the device which has it and for the other, I would put MIMO cable for both but in terms of the 'Sync' field, where the options are PC Clock, Unknown PPS and Don't Sync, which option should I select?
Thanks again for your help.
Kind Regards,
John
On 11/09/17 09:00, Derek Kozel wrote:
Hi John,
Are you power cycling the USRPs between tests or just rerunning the GRC flowgraph? Do you have shared 10 MHz and 1 PPS references? What version of UHD is printed in the output?
Thanks,
Derek
|