[USRP-users] B210 distributed MIMO with GPSDO
Hi All, I'm looking at using multiple B210s synchronized with GPSDO distributed over a few kilometers. The idea is to measure time-of-arrival of a signal source. I understand the limitations of the GPSDO 1pps & 10 MHz reference frequency/time stability. What I'm curious about is whether anybody has done this, and whether it worked without too much trouble. The only references I see of distributed MIMO, is where NI recommends using the N210: http://www.ni.com/tutorial/14705/en/ Any advice would be greatly appreciated. Stephan ___ USRP-users mailing list USRP-users@lists.ettus.com http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
[USRP-users] Tx_bursts
Hi Sumit, I am trying to send samples in bursts. How to compile this program tx_bursts. Because when compiled using command line arguments its showing error. /usr/local/src/uhd/host/examples$ ./tx_bursts --rate=1e6 --freq=100e6 --spb=8500 --repeat bash: ./tx_bursts: No such file or directory Can you please tell how where you to send burst samples. Regards, Sneha ___ USRP-users mailing list USRP-users@lists.ettus.com http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
Re: [USRP-users] Tx_bursts
Hi Sneha - If you're looking for the installed "tx_bursts" example, the default location should be in the directory "/usr/local/share/uhd/examples/". Cheers! - MLD On Tue, Oct 9, 2018, at 5:25 AM, Sneha vasan via USRP-users wrote: > Hi Sumit, > > I am trying to send samples in bursts. How to compile this program > tx_bursts.> Because when compiled using command line arguments its showing > error.> > /usr/local/src/uhd/host/examples$ ./tx_bursts --rate=1e6 --freq=100e6 > --spb=8500 --repeat> bash: ./tx_bursts: No such file or directory > Can you please tell how where you to send burst samples. > > Regards, > Sneha > _ > 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
Re: [USRP-users] FPGA compatibility
Hey Corey, yep, when updating your UHD you also need to update your fpga-src submodule; `git submodule update`. If you have changes you want to preserve: Standard git methods apply; just make a branch in the fpga- src submodule to contain your changes, then update the submodule, then merge in your branch. Best regards, Marcus On Mon, 2018-10-08 at 19:42 -0600, Corey Hahn via USRP-users wrote: > I did notice that UHD went from 4.0 to 3.14 is this a problem? Does > the getting started wiki need an extra step to choose the > branch/commit? > > On Mon, Oct 8, 2018 at 2:33 PM Rob Kossler wrote: > > Perhaps you are using fpga-src on 'master' rather than on commit > > 'ebf5eed' which is the one paired with UHD 'master'? > > > > Rob > > > > On Sun, Oct 7, 2018 at 10:22 PM Corey Hahn via USRP-users < > > usrp-users@lists.ettus.com> wrote: > > > Been using uhd/rfnoc for 9 months now with no comparability > > > problems i couldn't fix. > > > My last devel version was from the beginning of June. > > > Upgraded rfnoc and rebuilt from scratch the pybombs rfnoc > > > packages via the getting started wiki. > > > I have an x310 USRP > > > Redid the place and route with the new rfnoc build. Now i get > > > the dreaded comparability mismatch. In the past this was fixed > > > because i built the bit file or ran with uhd of two different > > > versions. This is on one machine and should not have this > > > problem. Where do i start debugging this problem? > > > > > > root@optical-cuda1604-20181007-srn21:~/rfnoc# uhd_usrp_probe > > > [INFO] [UHD] linux; GNU C++ version 5.4.0 20160609; Boost_105800; > > > UHD_3.14.0.0-110-g6af6ac32 > > > [INFO] [X300] X300 initialization sequence... > > > Error: RuntimeError: Expected FPGA compatibility number 35, but > > > got 36: > > > The FPGA image on your device is not compatible with this host > > > code build. > > > Download the appropriate FPGA images for this version of UHD. > > > Please run: > > > > > > "/root/rfnoc/lib/uhd/utils/uhd_images_downloader.py" > > > > > > Then burn a new image to the on-board flash storage of your > > > USRP X3xx device using the image loader utility. Use this > > > command: > > > > > > "/root/rfnoc/bin/uhd_image_loader" -- > > > args="type=x300,addr=192.168.40.2" > > > > > > For more information, refer to the UHD manual: > > > > > > http://files.ettus.com/manual/page_usrp_x3x0.html#x3x0_flash > > > root@optical-cuda1604-20181007-srn21:~/rfnoc# > > > > > > ___ > > > 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 ___ USRP-users mailing list USRP-users@lists.ettus.com http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
Re: [USRP-users] B210 distributed MIMO with GPSDO
Works! Don't forget that PPS accuracy of a GPS receiver is theoretical best- case some 50 ns; in reality, things can get a little worse than that. Best regards, Marcus On Tue, 2018-10-09 at 09:32 +0200, Stephan Esterhuizen via USRP-users wrote: > Hi All, > > I'm looking at using multiple B210s synchronized with GPSDO > distributed over a few kilometers. The idea is to measure time-of- > arrival of a signal source. I understand the limitations of the GPSDO > 1pps & 10 MHz reference frequency/time stability. > > What I'm curious about is whether anybody has done this, and whether > it worked without too much trouble. The only references I see of > distributed MIMO, is where NI recommends using the N210: > > http://www.ni.com/tutorial/14705/en/ > > Any advice would be greatly appreciated. > > Stephan > ___ > 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
Re: [USRP-users] B210 distributed MIMO with GPSDO
On 10/09/2018 01:24 PM, Marcus Müller via USRP-users wrote: Works! Don't forget that PPS accuracy of a GPS receiver is theoretical best- case some 50 ns; in reality, things can get a little worse than that. Best regards, Marcus It wasn't clear to me whether that was a *single* GPSDO whose outputs were being distributed, or whether this was individual B210s with on-board GPSDOs distributed over several kilometers. So, Stephan, which case is it? On Tue, 2018-10-09 at 09:32 +0200, Stephan Esterhuizen via USRP-users wrote: Hi All, I'm looking at using multiple B210s synchronized with GPSDO distributed over a few kilometers. The idea is to measure time-of- arrival of a signal source. I understand the limitations of the GPSDO 1pps & 10 MHz reference frequency/time stability. What I'm curious about is whether anybody has done this, and whether it worked without too much trouble. The only references I see of distributed MIMO, is where NI recommends using the N210: http://www.ni.com/tutorial/14705/en/ Any advice would be greatly appreciated. Stephan ___ 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 ___ USRP-users mailing list USRP-users@lists.ettus.com http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
Re: [USRP-users] FPGA compatibility
They match. commit 6af6ac32 cah@snoke:~/rfnoc/src/uhd$ git log -1 commit 6af6ac32c30d2dc0efa6eab61e4a3920649e3e62 Author: Trung Tran Date: Fri Aug 3 09:12:55 2018 -0700 examples: add rfnoc_radio_loopback This example will allow an RF->RF loopback using RFNoC devices. cah@snoke:~/rfnoc/src/uhd$ cd fpga-src/ cah@snoke:~/rfnoc/src/uhd/fpga-src$ git log -1 commit 6af6ac32c30d2dc0efa6eab61e4a3920649e3e62 Author: Trung Tran Date: Fri Aug 3 09:12:55 2018 -0700 examples: add rfnoc_radio_loopback This example will allow an RF->RF loopback using RFNoC devices. I want to reiterate i have done nothing but rm -rf the old rfnoc and .pybombs folders and rerun the pybombs installer. Corey On Tue, Oct 9, 2018 at 11:23 AM Marcus Müller wrote: > Hey Corey, > > yep, when updating your UHD you also need to update your fpga-src > submodule; `git submodule update`. If you have changes you want to > preserve: Standard git methods apply; just make a branch in the fpga- > src submodule to contain your changes, then update the submodule, then > merge in your branch. > > Best regards, > Marcus > On Mon, 2018-10-08 at 19:42 -0600, Corey Hahn via USRP-users wrote: > > I did notice that UHD went from 4.0 to 3.14 is this a problem? Does > > the getting started wiki need an extra step to choose the > > branch/commit? > > > > On Mon, Oct 8, 2018 at 2:33 PM Rob Kossler wrote: > > > Perhaps you are using fpga-src on 'master' rather than on commit > > > 'ebf5eed' which is the one paired with UHD 'master'? > > > > > > Rob > > > > > > On Sun, Oct 7, 2018 at 10:22 PM Corey Hahn via USRP-users < > > > usrp-users@lists.ettus.com> wrote: > > > > Been using uhd/rfnoc for 9 months now with no comparability > > > > problems i couldn't fix. > > > > My last devel version was from the beginning of June. > > > > Upgraded rfnoc and rebuilt from scratch the pybombs rfnoc > > > > packages via the getting started wiki. > > > > I have an x310 USRP > > > > Redid the place and route with the new rfnoc build. Now i get > > > > the dreaded comparability mismatch. In the past this was fixed > > > > because i built the bit file or ran with uhd of two different > > > > versions. This is on one machine and should not have this > > > > problem. Where do i start debugging this problem? > > > > > > > > root@optical-cuda1604-20181007-srn21:~/rfnoc# uhd_usrp_probe > > > > [INFO] [UHD] linux; GNU C++ version 5.4.0 20160609; Boost_105800; > > > > UHD_3.14.0.0-110-g6af6ac32 > > > > [INFO] [X300] X300 initialization sequence... > > > > Error: RuntimeError: Expected FPGA compatibility number 35, but > > > > got 36: > > > > The FPGA image on your device is not compatible with this host > > > > code build. > > > > Download the appropriate FPGA images for this version of UHD. > > > > Please run: > > > > > > > > "/root/rfnoc/lib/uhd/utils/uhd_images_downloader.py" > > > > > > > > Then burn a new image to the on-board flash storage of your > > > > USRP X3xx device using the image loader utility. Use this > > > > command: > > > > > > > > "/root/rfnoc/bin/uhd_image_loader" -- > > > > args="type=x300,addr=192.168.40.2" > > > > > > > > For more information, refer to the UHD manual: > > > > > > > > http://files.ettus.com/manual/page_usrp_x3x0.html#x3x0_flash > > > > root@optical-cuda1604-20181007-srn21:~/rfnoc# > > > > > > > > ___ > > > > 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 > > ___ USRP-users mailing list USRP-users@lists.ettus.com http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
Re: [USRP-users] FPGA compatibility
Ok, I get it now. They are NOT suppose to match. Using the commands you laid out put me on the right branch for fpga-src. Place and Route ing again. Corey On Tue, Oct 9, 2018 at 9:27 PM Corey Hahn wrote: > They match. commit 6af6ac32 > > cah@snoke:~/rfnoc/src/uhd$ git log -1 > commit 6af6ac32c30d2dc0efa6eab61e4a3920649e3e62 > Author: Trung Tran > Date: Fri Aug 3 09:12:55 2018 -0700 > > examples: add rfnoc_radio_loopback > > This example will allow an RF->RF loopback using RFNoC devices. > cah@snoke:~/rfnoc/src/uhd$ cd fpga-src/ > cah@snoke:~/rfnoc/src/uhd/fpga-src$ git log -1 > commit 6af6ac32c30d2dc0efa6eab61e4a3920649e3e62 > Author: Trung Tran > Date: Fri Aug 3 09:12:55 2018 -0700 > > examples: add rfnoc_radio_loopback > > This example will allow an RF->RF loopback using RFNoC devices. > > I want to reiterate i have done nothing but rm -rf the old rfnoc and > .pybombs folders and rerun the pybombs installer. > > Corey > > > On Tue, Oct 9, 2018 at 11:23 AM Marcus Müller > wrote: > >> Hey Corey, >> >> yep, when updating your UHD you also need to update your fpga-src >> submodule; `git submodule update`. If you have changes you want to >> preserve: Standard git methods apply; just make a branch in the fpga- >> src submodule to contain your changes, then update the submodule, then >> merge in your branch. >> >> Best regards, >> Marcus >> On Mon, 2018-10-08 at 19:42 -0600, Corey Hahn via USRP-users wrote: >> > I did notice that UHD went from 4.0 to 3.14 is this a problem? Does >> > the getting started wiki need an extra step to choose the >> > branch/commit? >> > >> > On Mon, Oct 8, 2018 at 2:33 PM Rob Kossler wrote: >> > > Perhaps you are using fpga-src on 'master' rather than on commit >> > > 'ebf5eed' which is the one paired with UHD 'master'? >> > > >> > > Rob >> > > >> > > On Sun, Oct 7, 2018 at 10:22 PM Corey Hahn via USRP-users < >> > > usrp-users@lists.ettus.com> wrote: >> > > > Been using uhd/rfnoc for 9 months now with no comparability >> > > > problems i couldn't fix. >> > > > My last devel version was from the beginning of June. >> > > > Upgraded rfnoc and rebuilt from scratch the pybombs rfnoc >> > > > packages via the getting started wiki. >> > > > I have an x310 USRP >> > > > Redid the place and route with the new rfnoc build. Now i get >> > > > the dreaded comparability mismatch. In the past this was fixed >> > > > because i built the bit file or ran with uhd of two different >> > > > versions. This is on one machine and should not have this >> > > > problem. Where do i start debugging this problem? >> > > > >> > > > root@optical-cuda1604-20181007-srn21:~/rfnoc# uhd_usrp_probe >> > > > [INFO] [UHD] linux; GNU C++ version 5.4.0 20160609; Boost_105800; >> > > > UHD_3.14.0.0-110-g6af6ac32 >> > > > [INFO] [X300] X300 initialization sequence... >> > > > Error: RuntimeError: Expected FPGA compatibility number 35, but >> > > > got 36: >> > > > The FPGA image on your device is not compatible with this host >> > > > code build. >> > > > Download the appropriate FPGA images for this version of UHD. >> > > > Please run: >> > > > >> > > > "/root/rfnoc/lib/uhd/utils/uhd_images_downloader.py" >> > > > >> > > > Then burn a new image to the on-board flash storage of your >> > > > USRP X3xx device using the image loader utility. Use this >> > > > command: >> > > > >> > > > "/root/rfnoc/bin/uhd_image_loader" -- >> > > > args="type=x300,addr=192.168.40.2" >> > > > >> > > > For more information, refer to the UHD manual: >> > > > >> > > > http://files.ettus.com/manual/page_usrp_x3x0.html#x3x0_flash >> > > > root@optical-cuda1604-20181007-srn21:~/rfnoc# >> > > > >> > > > ___ >> > > > 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 >> >> ___ USRP-users mailing list USRP-users@lists.ettus.com http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
Re: [USRP-users] B210 distributed MIMO with GPSDO
Each B210 will have a GPSDO, they'll be separated by a few kilometers. My understanding is as follows: 1) When each B210 syncs, the 1PPS pulse between each USRP would be aligned to +-50 nsec. Since they're only separated by a few km, I suspect the alignment might be better because of common errors (Ionosphere, orbits, clocks, etc) in the GPS solution 2) The initial sync to 1PPS aligns the internal FPGA real time clock to, say, 50 nsec 3) From here on, time will be aligned to 50 nsec +- the 10 MHz frequency stability integrated over your observation time. 4) With the Ettus GPSDO allan deviation spec of 1e-11 @ 1 second, this means for any 1 second interval the relative clock offset between two B210s should be around sqrt(2)*1e-11 seconds. My specific applications doesn't necessarily need the LOs to have the same phase, just the time of each sample be know. As a backup, since the B210 has 2 RX chains, I can do my application with RX1 and record GPS L1 signals on RX2, then post-process the GPS data to solve for clock offset. Regards Stephan On Wed, Oct 10, 2018 at 1:26 AM Marcus D. Leech via USRP-users < usrp-users@lists.ettus.com> wrote: > On 10/09/2018 01:24 PM, Marcus Müller via USRP-users wrote: > > Works! > > > > Don't forget that PPS accuracy of a GPS receiver is theoretical best- > > case some 50 ns; in reality, things can get a little worse than that. > > > > Best regards, > > Marcus > It wasn't clear to me whether that was a *single* GPSDO whose outputs > were being distributed, or whether this >was individual B210s with on-board GPSDOs distributed over several > kilometers. > > So, Stephan, which case is it? > > > > > > On Tue, 2018-10-09 at 09:32 +0200, Stephan Esterhuizen via USRP-users > > wrote: > >> Hi All, > >> > >> I'm looking at using multiple B210s synchronized with GPSDO > >> distributed over a few kilometers. The idea is to measure time-of- > >> arrival of a signal source. I understand the limitations of the GPSDO > >> 1pps & 10 MHz reference frequency/time stability. > >> > >> What I'm curious about is whether anybody has done this, and whether > >> it worked without too much trouble. The only references I see of > >> distributed MIMO, is where NI recommends using the N210: > >> > >> http://www.ni.com/tutorial/14705/en/ > >> > >> Any advice would be greatly appreciated. > >> > >> Stephan > >> ___ > >> 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 > > > ___ > 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
