On 03/07/2018 06:51 PM, Inkyu Bang via USRP-users wrote: > Hi, I am inkyu. > I am interested in synchronizing two USRP (each of them connected to a > different laptop) using pre-defined signals. > > Let's we have USRP A and B. > We want to send the signal from USRP A and B which should arrive the > destination receiver at the same time. > > We think of using pre-defined signal (let's say it preamble) before > sending data packets from A and B. > > For example, A sends preamble and B receives it. > Then, A sends data packets after "pre-defined time interval A" after it > sends preamble. > Also, B sends data packets after "pre-defined time interval B" after it > receives a preamble from A. > > Of course, we need to carefully choose pre-defined time interval A and B > to satisfy arrival time at the destination. > > Here are my questions. > > (1) Is there any related (this kind of synchronization) techniques which > are applicable in USRP?
Sample times are timestamped, so you can match your synchronization to that timebase. It's a wide question though, so it's hard to give you a more specific answer. In a nutshell, on device B, you run your correlation algorithm (or whatever it is that you do for synchronization), and when you have identified a sample for which you can say it's the first sample from your burst, you calculate the time based on the timestamps. > (2) Can we measure or predict processing time budget when we running USRP? Yeah, you can do that. From where to where are you measuring? Are you interested in end-to-end latency, or some other metric? -- M _______________________________________________ USRP-users mailing list [email protected] http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
