Back your baseband magnitude down to 0.9 and try again. Sent from my iPhone
> On Nov 10, 2020, at 10:02 PM, Lukas Haase <[email protected]> wrote: > > *** (So sorry for sending this twice, I forgot to attach the screenshot) *** > > Hi Marcus, > >>> On 11/10/2020 07:06 PM, Lukas Haase via USRP-users wrote: >>> >>> The reason why I am asking about this is because I would like to >>> cross-check my measurements. >>> >>> My original question was about RX gain but it actually also relates to TX >>> gain. >>> >>> With an X310 and UBX-160, TX Gain=0 and sending out a full-scale sinusoid >>> at 915MHz, I measure approx. -2.5dBm. >>> >>> This is not consistent at all with the file above (assuming I interpret how >>> "gain" is defined correctly, hence my original question). >>> >>> Would anyone with an UBX-160 be so kind to confirm/disconfirm which power >>> level I would expect to see? >>> >>> (Preferably at 915MHz with TXgain=0 and full-scale sinusoid but I'm happy >>> with any other configuration that I can x-check). > >> So, according to the charts, at 900Mhz, you'd expect about -9dBm with a >> 0dB gain setting, with a full-scale baseband signal. >> >> So, there's a discrepancy of about 6.5dB. > > Exactly. That's a bit too much for my taste and hence maybe my measurement is > wrong. > >> How did you measure the power >> level? I could easily expect a couple of >> dB discrepancy just due to component batch-to-batch variability. >> So, how was the power measured? > > I know, ideally you'd use a powermeter for that. But I don't have one. > Instead I have a VNA (Agilent E5080A) which is freshly calibrated and hence > the power measurement is very accurate (<<0.5dB). > > I set the measurement as "B Source Port 1" which gives directly the received > signal in dBm. I pick a "reasonable" IF bandwidth (50kHz), 100x averaging, > save the dBm/f trace to a CSV file and import into MATLAB. Then I integrate > over the frequency: > > deltaF = frequencies(2)-frequencies(1); > Ptot = 10*log10(trapz(10.^(PSD/10)/RBW)*deltaF); > > I can verify that the integrated power is roughly the same as the marker on > the frequency. > > Having said that, I just realized something very, very weird: With the > UBX-160 I set: fcenter=915MHz, samp_rate=5e6, TXgain=0 and send a full scale, > complex sinusoid with 1MHz. Hence I would expect to see one tone at 916MHz > and a small one (due to I/Q mismatch) at 914MHz. However, instead I get the > spectrum attached. There is another tone at 910 MHz and the spectrum is > centered at 913 MHz (should be 915). > > That makes no sense! What could be going on here? > > Thanks, > Lukas > > <E5080A_UBX160_Output.PNG> _______________________________________________ USRP-users mailing list [email protected] http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
