# Variable propagation delays **URL:** \ **Category:** GNU Radio\ **Created:** [March 12, 2014, 8:52pm UTC](https://www.ruby-forum.com/t/variable-propagation-delays/239744 "2014-03-12T20:52:26Z")\ **Posts on this page:** 1\ **Page:** 1
**Author:** ![musicdenotation](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/m/ea666f/32.png) [@musicdenotation](https://www.ruby-forum.com/u/musicdenotation)\ **Post date:** [March 12, 2014, 8:52pm UTC](https://www.ruby-forum.com/t/variable-propagation-delays/239744/1 "2014-03-12T20:52:26Z")
Hello, My application requires very consistent propagation delays from TX to RX between measurements (within 1 sample or so). I am seeing some odd behavior and could use a little help. My flowgraph is as follows: Vector source -\> Frequency mod -\> USRP sink USRP source -\> Head -\> File sink (The reason for the vector source is that it has a rewind() method that I need) To take consecutive measurements without restarting the script (thus restarting the USRP PLL’s), I stop the flowgraph, rewind() the vector source, reset() the head, and schedule both usrp\_source and usrp\_sink using set\_start\_time(), then start the flowgraph again. When I TX and RX on the same frequency, I get consistent propagation delays between measurements (within 1 sample). When I TX and RX on different frequencies (with an external frequency shifter (hardware) in the loop), I get propagation delays varying from ~2 to 10 samples between measurements. Could these variations be due to residual samples in some blocks’ buffers when the flowgraph is stopped and restarted? But then why are the delays consistent when TX and RX are the same frequency? Could it somehow be due to USRP clock phase variation when the flowgraph is restarted? Is there any way to reset all the buffers and streams (so that each time the flowgraph is stopped then started, everything is “fresh”) so that multiple measurements can be taken without restarting the script (thus the USRP PLL’s)?