usrp-users@lists.ettus.com

Discussion and technical support related to USRP, UHD, RFNoC

View all threads

Simple way to clear RFNoC Replay input buffer

KR
kurz.reido@gmail.com
Fri, May 16, 2025 7:07 PM

Hello all,

A simplified diagram of my setup is as follows

Vector Source (Repeat) -> RFNoC Tx Streamer -> RFNoC Replay -> RFNoC Tx Radio

My vector source uses variables that are modified by the host laptop through XMLRPC Client, so the vector output can change if the user issues a change for a parameter through the GRC GUI.

My problem is that the vector source keeps sending the same data until the buffer is full, so when the user issues a change and presses the Record button for RFNoC Replay nothing seems to happen. Depending on buffer size the user may have to press the button several times in order to observe the vector source change on a Time Sink.

My question: Is there a simple way to clear the RFNoC Replay Block buffer using some GRC block or embedded python block code?

I have tried using embedded python blocks to make a splitter and a counter that would cut off the flow to the Replay block, but some samples seem to sneak through before the port change and other issues arise.

Hello all, A simplified diagram of my setup is as follows Vector Source (Repeat) -> RFNoC Tx Streamer -> RFNoC Replay -> RFNoC Tx Radio My vector source uses variables that are modified by the host laptop through XMLRPC Client, so the vector output can change if the user issues a change for a parameter through the GRC GUI. My problem is that the vector source keeps sending the same data until the buffer is full, so when the user issues a change and presses the Record button for RFNoC Replay nothing seems to happen. Depending on buffer size the user may have to press the button several times in order to observe the vector source change on a Time Sink. **My question:** Is there a simple way to clear the RFNoC Replay Block buffer using some GRC block or embedded python block code? I have tried using embedded python blocks to make a splitter and a counter that would cut off the flow to the Replay block, but some samples seem to sneak through before the port change and other issues arise.