Single B210 with GNU Radio: TX and RX synchronisation

[keukenrol]

Dear

I am tasked with implementing a time-of flight based system. The system sends a pulse on frequency A and receives it after a delay (microsecond range) on frequency B (using correlation). I have setup the pulse generation in GNU Radio Companion. To test the latency of the setup I connected RF TX A to RF RX A on the B210. The output is then visualized in a QT GUI Time Sink.

However, every time when starting the flow the visualized delay is different. I want to make sure my transmitter and receiver are synchronized. This means they should start at exactly the same time and continue at the same rate, only then the time between sending the data to the USRP sink and receiving it from the USRP source can be used for the application.

How do I need to configure both USRP Source (RX) and USRP sink (TX) in GNU Radio Companion to make them synchronized?

Furthermore, how would you suggest handling overflows or underflows? These have a big influence in measuring the time difference.

Best regards

[keukenrol]

Solved by varying the start time for the sink and source blocks. 6.55 microseconds earlier for TX then RX on my unit.

[ShayAbby]

Hello keukenrol,

Thank you for your update in this problem. I have the same issue. Can you provide more details. How did you get 6.55 microseconds? The delay time is different everytime. Can you elaborate on whether/how it solves the problem?

[keukenrol]

Hello

the flow consisted of a USRP sink block and a USRP source block. A pulse generator was connected to the sink block. The TX port was connected to the RX port using an SMA cable. The same frequency was used.

The USRP source output and the generated pulsed signal were both fed trough a peak detector and then piped to a custom python block, which determined the true delta time between the two detected pulses in terms of samples. Keep in mind that signal processing might alter the delays.