Hasan, Are the USRPs sharing any kind of timing / clocking signals? Typically we would recommend that you share a reference clock and PPS clock across all devices, and from there you'd just need to coordinate timed commands. With timed commands in UHD, you can tell each USRP to reset its internal sense of time to 0 on the next rising PPS edge. Once they have agreed upon this time, you should see them stay locked together as a shared reference clock with prevent drift. One obstacle is ensuring that each terminal (are these separate computers?) needs to send the "time reset" command to its USRPs within a single PPS cycle. If your reset commands came on different PPS cycles, you'd see your radios offset by 1 second. Here is the manual page with the list of timed commands that you can use for this:
https://files.ettus.com/manual/classuhd_1_1usrp_1_1multi__usrp.html Can you also let us know what USRPs and version of UHD you're using? Sam Reiter Ettus Research On Wed, Dec 4, 2019 at 4:03 AM Hasan Can Yildirim via USRP-users < usrp-users@lists.ettus.com> wrote: > Hi all, > > *Here is the summary of my experimental setup:* > > 4 data receiving channels with 2 USRPs. Sampling rates are 100 Msps. > 1 data transmitter with 1 USRP. Sampling rate is 100 Msps. > 1 calibration transmitter with 1 USRP. Sampling rate is 50 Msps (or > smaller, say 33.3 Msps). > > All the daughterboards are UBX160. > The USRPs are connected through the 10Gbit/s ethernet connections. > I am using a modified version of the C/C++ code with the uhd libraries. > So, no gnu-radio etc. > > *The question in one sentence:* > > How can I start transmitting at the same time with 2 USRPs, that has two > different sampling rates, and 'invoked' at two different terminals? > > *Here is the detailed explanation about what I want to achieve with this > setup:* > > I want to use the calibration signal, to estimate the unknown > delays/phases (introduced by the hardwares) on the receiver side, then use > this calibration to apply angle-of-arrival algorithms on the received data. > I know that I could transmit the calibration signal and the data signal, > at the same time. Then apply a high-pass filter to separate the calibration > signal, do the calibration, and so on... > Instead, I would like to synchronize the data and calibration > transmitters, so that they start transmitting at the same time. Receiver > will receive continuously, for a long enough duration. > In other words, the calibration signal will be a signal of N samples and M > zeros. Mean while, the data will be N zeros, M data samples. So, during the > first N samples, I will receive only the calibration signal, then during > the next M samples, I will receive the data signal. To achieve this, the > two transmitters have to start transmitting at the same time (maybe a small > error with a few samples is acceptable). How can I achieve this? > > > Thanks a lot for your advices. > > Cheers, > Hasan > > > > _______________________________________________ > USRP-users mailing list > USRP-users@lists.ettus.com > http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com >
_______________________________________________ USRP-users mailing list USRP-users@lists.ettus.com http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
