After all this might be an expected behavior. Since the devices have different clocks, they sample the medium at different time instances,leading to this "distortion" of the signal. That's why this only happens when the one USRP transmits and the other receives, and not when the same usrp transmits and receives. This calls for some clock recovery function.
Thank you very much for your time Marcus Στις 14 Ιουν 2017 10:55 π.μ., ο χρήστης "George Vardakis" < vardakis....@gmail.com> έγραψε: Στις 13 Ιουν 2017 10:22 μ.μ., ο χρήστης <mle...@ripnet.com> έγραψε: In the two-USRPs scenario, one of the USRPs is transmitting, and one is receiving? Yes. That's the problematic scenario. Are they "through the air" or over a cable. If over a cable, how much attenuation has been added to the cable. They are through the air. On 2017-06-13 11:16, Manolis Surligas wrote: Ok lets explain a little bit more what actually you observe. Each burst is a training sequence that has a duration of 64 samples and then 1000 zero samples. This is generated using a vector source block repeatedly. As you can see in the ok.jpeg all sequences are exactly the same. In the other plot however the sequence constantly changes. To me, it seems that parts of the actual sequence are randomly cropped or re-arranged. This happens only when two different B210 are used. We also tried to transmit the sequence from a third device with the same results. And yes we have tried to receive from both devices each one working on a separate flowgraph. The result was exactly the same. On 06/13/2017 06:05 PM, mle...@ripnet.com wrote: I'm having a hard time seeing a significant functional difference between the two. Could you perhaps explain? Also, maybe a diagram showing the two configurations? Do you get any error messages during a run? Are you doing reception on the two USRPs within a single flow-graph, or two separate processes? On 2017-06-13 10:51, Manolis Surligas wrote: Hi Marcus, yeah we have updated the UHD on the latest version and the problem still persists. I have started to debug the UHD and we suspect some kind of inconsistency between the device pointers of the different devices. Perhaps a race condition is hidden somewhere. I also attach two plots. The first one is the problematic with two different USRPs and the second the expected signal transmitting and receiving using the same USRP. On 06/13/2017 05:30 PM, mle...@ripnet.com wrote: Thanks, George. Two things: Could you try updating to the most recent UHD version? It would be helpful if you provided your .dat files in the form of plots or graphs, highlighting the areas you consider "distortion". This will help us help you. On 2017-06-13 07:23, George Vardakis wrote: Good evening, I attach three captures to show you the problem, one with the reference signal i transmit, one with the received samples when using the same device for TX and RX and one with the received samples of the one device, when the second one transmits. I also attach the flowgraph i used for this capture. The sampling rate i use is 1MHz. The specific captures are done with the USRPs over USB-2 ports, but i have also tried over USB-3 and observed the same behavior. Thank you for your time same_tx_rx_device.dat <https://drive.google.com/file/d/0BxyASy57xNmqV2pLd2xUVWRRRU0/view?usp=drive_web> refference.dat <https://drive.google.com/file/d/0BxyASy57xNmqT1JQd2xfWjF6SjQ/view?usp=drive_web> different_tx_rx_device.dat <https://drive.google.com/file/d/0BxyASy57xNmqMVhqd29NTmdiMDA/view?usp=drive_web> On Tue, Jun 13, 2017 at 12:04 AM, Marcus D. Leech <mle...@ripnet.com> wrote: > On 06/12/2017 03:53 PM, George Vardakis wrote: > > Hello all, > > I have two USRP B210 devices which i use in my MIMO application. When i > use the two of them concurrently in a flowgraph - or different flowgraphs - > i notice that the signal received by the streams of one of the devices (the > one whose serial number is last in lexicographical order) is distorted. > Because i transmit a known training sequence, i know what to expect, and i > observe that sometimes the real or imaginary part of the signal has flipped > its sign (i see the signal mirrored compared to the known one), or other > times it is completely distorted. When i use each device separately, > everything works fine. The UHD driver i use is the v3.9.5 one.Any ideas > about it? > > Thank you! > > _______________________________________________ > Discuss-gnuradio mailing > listDiscuss-gnuradio@gnu.orghttps://lists.gnu.org/mailman/listinfo/discuss-gnuradio > > COuld you perhaps share a very-simple flow-graph that shows the issue? > What sample-rate are you using? Is this over USB-2 or USB-3? > _______________________________________________ Discuss-gnuradio mailing > list Discuss-gnuradio@gnu.org https://lists.gnu.org/mailman/ > listinfo/discuss-gnuradio _______________________________________________ Discuss-gnuradio mailing listDiscuss-gnuradio@gnu.orghttps://lists.gnu.org/mailman/listinfo/discuss-gnuradio -- /* Code is the Law */ _______________________________________________ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org https://lists.gnu.org/mailman/ listinfo/discuss-gnuradio -- /* Code is the Law */ _______________________________________________ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
_______________________________________________ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
