Hi Matthias,
Thank you for your help! Setting both parameters now works for me. When setting the MasterClock manually to twice the sample rate I basically get the same performance results like when I achieve oversampling through interpolation/decimation functions within Matlab: For my pure USRP-USRP connection the performance is improved significantly but for the mixed Hardware cases ,using either a Signal Generator (TX) or Spectrum Analyzer (RX) with one USRP, the oversampling does not result in any improvement of my received waveform. Up to now, I cannot really explain this effect. What I have also noticed is that the signal quality is not constant over several transmissions when sending with the B200 even though I do not change any parameters. I observed the same effect when looking at the spectrum sending two sine signals several times. For constant TX gain the IP3 value varies for each transmission which may explain this effect. This inconsistency makes a definite performance characterization rather difficult. Whenever I send with the signal generator and only receive with the USRP, the signal quality is more or less constant which also supports the assumption that the inconsistency is caused by the TX amplifier. I might be able to do similar performance test with the USRP X-Series and a UBX-160 Daughterboard soon and hope that I will see improvements in the signal quality. Best regards, Felix Date: Wed, 2 May 2018 10:20:31 +0200 From: Matthias Br?ndli < <mailto:matth...@mpb.li> matth...@mpb.li> To: <mailto:usrp-users@lists.ettus.com> usrp-users@lists.ettus.com Subject: Re: [USRP-users] Optimizing B200 Performance Message-ID: < <mailto:e0ea50a0-6ed4-ffd6-a9ff-dbf6183e5...@mpb.li> e0ea50a0-6ed4-ffd6-a9ff-dbf6183e5...@mpb.li> Content-Type: text/plain; charset=utf-8 Hi Felix, the following works for me (newlines added for readability), and I see in the startup messages that it properly sets the master clock rate: /usr/local/lib/uhd/examples/rx_samples_to_file --args "serial=309B1E7,master_clock_rate=32768000" --rate 2048000 --freq 200000000 --file /dev/null mpb On 02/05/18 10:11, Felix via USRP-users wrote: > Dear Matthias, > > Thank you for your feedback! I actually already take some headroom for the > digital signal into account by setting my USRP rx gain to a value where the > max amplitude (normalized to 1) of my received signal is always between 0.7 > and 0.9 to prevent clipping of the signal. > > Regarding the master clock rate setting: I am working with signals with a > bandwidth of 20MHz or higher so my interpolation/decimation factor is > limited to values 1 and 2 using the B200. It seems that oversampling is > necessary to prevent filtering of the spectrum edges by the hardware. I > have tried setting the master clock with the --args argument within the UHD > functions 'tx_samples_from_file' and 'rx_samples_from_file' to achieve a > factor of 2 between master clock and sample rate. However, as I am connected > to two different B200 devices I have used the --args function argument > before to specify the serial number of the device I want to use. It seems > like I am only able to either specify my clock rate or my device but not > both parameters at the same time. > > Do you know if there is a way to specify both parameters (device serial > number and master clock rate) within one function call? > > Best regards, > Felix
_______________________________________________ USRP-users mailing list USRP-users@lists.ettus.com http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
