I notice that the GRC is using RFNoC blocks. Perhaps it would be a good test to implement standard UHD USRP blocks. I think that your functionality is possible with the USRP blocks. Rob
On Mon, Nov 19, 2018 at 7:25 AM Gwenhael Goavec-Merou via USRP-users < usrp-users@lists.ettus.com> wrote: > Hello, > > No ideas, advice or anything else to solve my problem? > > Thank > > Gwen > > On Tue, 6 Nov 2018 22:34:27 +0100 > Gwenhael Goavec-Merou via USRP-users <usrp-users@lists.ettus.com> wrote: > > > On Tue, 06 Nov 2018 15:42:10 -0500 > > "Marcus D. Leech" <patchvonbr...@gmail.com> wrote: > > > > > On 11/06/2018 12:23 PM, Gwenhael Goavec-Merou wrote: > > > > On Tue, 06 Nov 2018 12:09:02 -0500 > > > > "Marcus D. Leech" <patchvonbr...@gmail.com> wrote: > > > > > > > >> On 11/06/2018 10:24 AM, Gwenhael Goavec-Merou wrote: > > > >>> On Tue, 06 Nov 2018 09:40:09 -0500 > > > >>> "Marcus D. Leech via USRP-users" <usrp-users@lists.ettus.com> > wrote: > > > >>> > > > >>>> On 11/06/2018 04:12 AM, Gwenhael Goavec-Merou via USRP-users > wrote: > > > >>>>> Hi, > > > >>>>> > > > >>>>> My environment is: > > > >>>>> USRP: X310 with basicRX (currently one but 6 in a near future) > > > >>>>> UHD: 3.13.1.0-rc1 > > > >>>>> GNURadio: 3.7.13.4-r2 > > > >>>>> gr-ettus: master branch, up to date > > > >>>>> > > > >>>>> I need to sample 4 real signal / USRP and to demodulate / > decimate > > > >>>>> before transfer to the PC. > > > >>>>> > > > >>>>> I have created a GNURadio flow with: > > > >>>>> - 2 Radio (id 0 and 1), configured with 2 channels > > > >>>>> - 2 DDC (id 0 and 1), with 2 channels, with a center frequency > fixed > > > >>>>> to XX MHz > > > >>>>> - 4 complex To Real; > > > >>>>> - 1 QT Gui with 4 inputs. > > > >>>>> > > > >>>>> Each channel of each radio is connected to an DDC input (radio0 > > > >>>>> channel 0 to DDC0 input 0, radio0 channel 1 to DDC0 input 1, > etc.) > > > >>>>> > > > >>>>> With this setup and with an input configured as XX + YY MHz (to > have a > > > >>>>> beatnote) I'm able to see a sinusoid on the qt for all channels > (if I > > > >>>>> plug / unplug an input signal the corresponding plot is equal to > 0 or > > > >>>>> displays the signal). > > > >>>>> But: > > > >>>>> 1/ the console panel display continuouly messages about channels > > > >>>>> overrun 2/ the data flow is not continuous (visible in qt plot) > (I > > > >>>>> suppose is due to 1/ ) > > > >>>>> 3/ channels are not aligned (I suppose is due to 1/ and/or 2/ > > > >>>>> > > > >>>>> Theorically, by reading HDL files for the X310 firmware it's > seems > > > >>>>> possible to use this configuration. > > > >>>>> > > > >>>>> Then, how to fix my issue? > > > >>>>> - Is it a USRP/UHD limitation? > > > >>>>> - gr-ettus seems not heavily upgraded, is something must be > modified > > > >>>>> in this repository? > > > >>>>> - I am just wrong somewhere on my design? > > > >>>>> > > > >>>>> Thank you very much > > > >>>>> > > > >>>>> Gwen > > > >>>>> > > > >>>> What is the sample rate as delivered to the PC? What kind of > PC? > > > >>> The sample rate is 3MS/s for each channels after DDC (200MS/s > before due > > > >>> to the ADC). > > > >>> My PC is a Lenovo thinkpad x230 with a 1Gb ethernet card. > > > >>>> Overrun means that your PC isn't keeping up with the sample flow. > > > >>>> > > > >>> I've just done several test with only 2 channels : > > > >>> - 2 radio, 1 channel/radio and 2 DDC with 1 input (first on > radio0, a > > > >>> second radio1) : > > > >>> - data continuous (but not synchronized) > > > >>> - no message on display panel > > > >>> - 2 radio, 1 channel/radio and 1 DDC with 2 input : > > > >>> - console panel displays continuouly "Ooverrun on chan 0" > > > >>> - data not continuous > > > >>> - after a short time : freeze > > > >>> - 1 radio with 2 channels and 1 DDC with 2 input : > > > >>> - console panel displays continuouly "Ooverrun on chan 0" > and > > > >>> "Ooverrun on chan 1" > > > >>> - data not continuous. > > > >>> - 1 radio with 2 channels and 2 DDC with 1 input : > > > >>> - console panel displays continuouly "Ooverrun on chan 0" > > > >>> - data not continuous. > > > >>> > > > >>> So, the first test seems to show it's possible with my computer to > > > >>> receive 2 channel 3MS/s without overrun. Other tests seems to show > a > > > >>> problem for radio and/or for DDC to send or receive the full speed > ADC > > > >>> flow. > > > >>> > > > >>> Thank. > > > >>> > > > >>> Gwen > > > >> The only immediate comment I have is that 3Msps is not a valid rate > for > > > >> the DDC with a 200Msps input--rates must be an integer fraction > > > >> of the input sample rate. > > > > Tested with 2Msps this the same result... > > > >> It would be helpful if you shared your flow-graph for the cases that > > > >> don't work. > > > > joined > > > > > > > > Gwen > > > Also, in your Radio BLock, you have "E310: 2 channels" selected, which > > > very likely doesn't work with X310. > > > > > I've checked multiple time the flow-graph, top_block.py, do grep in > gr-ettus, > > uhd, gnuradio sources and I'm unable to find this option... Maybe we > have > > different version of one piece of code... > > > > I've forget to specify : to be able to select antennas according to > basicRX > > requierement (A or B) I've modified uhd_rfnoc_radio.xml to add this > entry (I > > must to sent a patch to fix this). > > > > Gwen > > > > > > _______________________________________________ > > 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
