Pol, Glad to hear it! I can answer at least a few of your questions.
On Mon, Jul 2, 2018 at 3:32 AM Pol Henarejos <pol.henare...@cttc.cat> wrote: > Dear all, > > I have more news on that. After following the steps provided by Nick, I > finally succeed with the loopback. However, I have some observations: > 1- At the Addendum, it is described that set_tx_streamer shall be used > with E310. After many hours spent, I tried to set it also with X300 and > it worked with a simple flowgraph. So, my impression is that the > set_tx_streamer has to be also used in the case of X300, at least in my > case. > That's very strange to me, although I know they've updated the streamer logic since I've last merged from rfnoc-devel. > 2- I use the rfnoc-devel commit f3118cb6. > 3- I had to employ the 120 Msps master_clock_rate. For higher master > rates, the system does not work. No errors nor complains are displayed. > The leds just do not blink. > Can you be very specific about what the flowgraph looked like which would not work at 200Msps? > 4- In very very simple flowgraphs (radio RX -> radio TX without > queues/FIFOs) the loopback works at 200 Msps, but if some additional > blocks (splitters, adders, duc/ddc) are added, the system does not work > (no errors again, just leds off). > This is because RFNoC blocks have a total input/output bandwidth which is independent of the number of inputs/outputs. You get enough bandwidth to stream 1 input/1 output at 200Msps, but if you use a splitter (1 input, 2 outputs) or an addsub (2 inputs, 2 outputs) you can no longer stream at full rate. > 5- The presence of ddc/duc to reduce the clock_rate of intermediate > blocks seems that does not affect to the performance. All intermediate > blocks can work at 120 Msps but not at 200. > I'm assuming this is related to the 1in/1out thing. > 6- I cannot stop the workflow. This is very annoying because I have to > power off the USRP *every time* that I run the flowgraph. Even though I > send STREAM_MODE_STOP_CONTINUOUS to both rx radio and I set > set_tx/rx_streamer to false, the loopback still runs. I use GRC and I am > not sure wether is caused by gnuradio or RFNoC. I am going to work with > this. > Simplest thing to do is just run uhd_usrp_probe. I would have guessed sending a stream command which stops things would work as well but to be honest I haven't tried it because my use case doesn't demand it. Nick > > IMHO a warn/error/complain should be displayed if the RFNoC system > cannot work due to the master clock rate because there is no additional > information a priori to know what's wrong. My USRPs are placed in a > remote lab and I am not able to see wether tx/rx leds are blinking or > are off. My impression is that the blocks cannot meet the master clock > rate but I am pretty sure that you can provide more information on that. > > I will update this thread if more progresses are achieved. > > Thank you all. > > > Dr. Pol Henarejos > Researcher > Array and Multi-Sensor Processing Department, Communication Systems > Division > pol.henare...@cttc.cat > > Centre Tecnològic de Telecomunicacions de Catalunya (CTTC) > Av. Carl Friedrich Gauss, 7 > 08860 Castelldefels, Barcelona (Spain) > Tel: +34 93 645 29 00 Ext: 2177 <+34%20936%2045%2029%2000> > www.cttc.cat > > El 25/6/2018 a les 17:28, Nick Foster via USRP-users ha escrit: > > What is the behavior of the lights on the front panel? > > > > On Mon, Jun 25, 2018, 12:31 AM Pol Henarejos via USRP-users > > <usrp-users@lists.ettus.com <mailto:usrp-users@lists.ettus.com>> wrote: > > > > Dear Derek, > > > > I tried the link that you gave but no luck. > > 1) I disabled the use of timestamp and, dissecting the packets with > > wireshark, I cross-checked that the use_timestamp bit is 0. > > 2) I added the set_rx_streamer and issue_cmd to my flograph but > nothing > > happens. I am not sure if both commands work properly, but I guess > > they do. > > > > Do you have additional hints on that? > > > > Thank you. > > > > > > Dr. Pol Henarejos > > Researcher > > Array and Multi-Sensor Processing Department, Communication Systems > > Division > > pol.henare...@cttc.cat <mailto:pol.henare...@cttc.cat> > > > > Centre Tecnològic de Telecomunicacions de Catalunya (CTTC) > > Av. Carl Friedrich Gauss, 7 > > 08860 Castelldefels, Barcelona (Spain) > > Tel: +34 93 645 29 00 Ext: 2177 <+34%20936%2045%2029%2000> > > www.cttc.cat <http://www.cttc.cat> > > > > El 22/6/2018 a les 14:21, Pol Henarejos via USRP-users ha escrit: > > > Thank you, Derek. I will take a look. > > > I hope that some progress will be achieved in a near future. > > > > > > Best regards. > > > > > > > > > Dr. Pol Henarejos > > > Researcher > > > Array and Multi-Sensor Processing Department, Communication > Systems > > > Division > > > pol.henare...@cttc.cat <mailto:pol.henare...@cttc.cat> > > > > > > Centre Tecnològic de Telecomunicacions de Catalunya (CTTC) > > > Av. Carl Friedrich Gauss, 7 > > > 08860 Castelldefels, Barcelona (Spain) > > > Tel: +34 93 645 29 00 Ext: 2177 <+34%20936%2045%2029%2000> > > > www.cttc.cat <http://www.cttc.cat> > > > > > > El 22/6/2018 a les 13:13, Derek Kozel ha escrit: > > >> Hello Dr Henarejos, > > >> > > >> Currently an RFNoC only flowgraph is not fully supported. What is > > >> missing is the issuing of the commands to start streaming and a > > >> mechanism to handle timestamps. Some modifications are needed to > > make > > >> these work. > > >> Nick Foster has done a write up on a few different ways of > handling > > >> the necessary modifications. > > >> > > >> https://corvid.io/2017/04/22/stupid-rfnoc-tricks-loopback/ > > >> > > >> Regards, > > >> Derek > > >> > > >> On Fri, Jun 22, 2018 at 10:10 AM, Pol Henarejos via USRP-users > > >> <usrp-users@lists.ettus.com <mailto:usrp-users@lists.ettus.com> > > <mailto:usrp-users@lists.ettus.com > > <mailto:usrp-users@lists.ettus.com>>> wrote: > > >> > > >> Dear all, > > >> > > >> I am playing with RFNoC on a USRP X300. I built a custom > > RFNoC FPGA > > >> image with the blocks that I need successfully and > > everything goes > > >> smooth. However, if I want to use a flowgraph with only > > RFNoC blocks > > >> (RFNoC Radio RX --> DmaFIFO --> RFNoC Radio TX), it does not > > work. > > >> The USRP does nothing (no blinking leds, no error messages, > > >> nothing). If I replace the DmaFIFO block with a Copy block, > > then it > > >> works but, as expected, this is not efficient since all the > > samples > > >> go to the host and go back to the USRP. I would want to > > perform a > > >> relay by using only RFNoC blocks. My impression is that a > > command to > > >> the USRP might be needed (a streaming command or so), but I > > have no > > >> idea what is wrong. > > >> > > >> I attach the flowgraph. > > >> > > >> (only RFNoC blocks does not work) > > >> > > >> (flowing samples to host and going back to USRP does work > > but it is > > >> extremely inefficient) > > >> > > >> Thank you. > > >> > > >> > > >> -- > > >> Dr. Pol Henarejos > > >> Researcher > > >> Array and Multi-Sensor Processing Department, Communication > > >> Systems Division > > >> pol.henare...@cttc.cat <mailto:pol.henare...@cttc.cat> > > <mailto:pol.henare...@cttc.cat <mailto:pol.henare...@cttc.cat>> > > >> > > >> Centre Tecnològic de Telecomunicacions de Catalunya (CTTC) > > >> Av. Carl Friedrich Gauss, 7 > > >> 08860 Castelldefels, Barcelona (Spain) > > >> Tel: +34 93 645 29 00 Ext: 2177 <+34%20936%2045%2029%2000> > > >> www.cttc.cat <http://www.cttc.cat> <http://www.cttc.cat> > > >> > > >> > > >> _______________________________________________ > > >> USRP-users mailing list > > >> USRP-users@lists.ettus.com <mailto:USRP-users@lists.ettus.com> > > <mailto:USRP-users@lists.ettus.com <mailto: > USRP-users@lists.ettus.com>> > > >> > http://lists.ettus.com/mailman/listinfo/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 <mailto: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 <mailto: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 > > > >
_______________________________________________ USRP-users mailing list USRP-users@lists.ettus.com http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
