Hi Leon thank you very much for the screenshots. From these screenshots it is clearly visible that the noise floor shape you plot is similar to the noise floor shape I plot. I did the measurements both with the version 3.15.0.0 and 4.4.0.0 and the shape does not change.
Thank you. Best regards Carlo Venier Il ven 23 giu 2023, 11:42 Leon Wabeke <lwab...@csir.co.za> ha scritto: > Hi Carlo > > > > I am not sure if this will help, but here is what I could quickly get for > you. I am accessing the X310 remotely and could either switch it to be > connected to Antennas (LP0410 Antenna > <https://www.ettus.com/all-products/lp0410/> 400MHz to 1GHz) sold by > Ettus or then connected to a test loopback path driven from a B210 via a > splitter. I didn’t run the B210s app, so that is as unconnected/terminated > as I could make it. > > > > I am sorry the remote exported screen cut of the highest bit of the > spectrum. This was FFT using 100000 samples at 100Msps, showing an average > over 50 such magnitudes to reduce the variance to give a more concrete view > of the spectrum shape. > > > > This was using UHD 4.2. I remember differences in the spectrum shape when > running UHD3.13 or 3.15. I mostly tested with decimation down to 20MSps. In > that situation the centre tuning frequency showed up as a leakage spur > (although from my understanding there should be no LO set to that with the > TwinRX architecture), but more so on the second channel. And with 2 TwinRxs > in one X310, channel 2 and 4 had that issue with the leakage being about 5 > dB higher than on the other 2 channels (although still present on the > others as well). This leakage seems to be improved now with UHD4.2 on my > hardware. > > > > Regards > > Leon > > > > *From: *Marcus D. Leech <patchvonbr...@gmail.com> > *Date: *Thursday, 22 June 2023 at 19:41 > *To: *Carlo Venier <venier.ca...@gmail.com> > *Cc: *usrp-users@lists.ettus.com <usrp-users@lists.ettus.com> > *Subject: *[USRP-users] Re: TwinRx Spectrum shape > > [The e-mail server of the sender could not be verified (SPF Record)] > > On 22/06/2023 12:03, Carlo Venier wrote: > > Hi Marcus, > > when I turn down the gain, the shape of the spectrum noisefloor remains > non-flat. Also at a gain of 0, the shape is wavy. > I'm not feeding any RF source to the channels: they are 50 Ohm terminated. > The plots of the spectrum were produced with a third party application. > However, I've recorded the samples by using the "rx_samples_to_file" tool > and then plotted them with Matlab but the shape remains the same (non-flat). > I am aware that the final spectrum is the convolution of all the filtering > stages, but I would like to know if this is the expected noisefloor > spectrum for the TwinRx (for example for UBX the spectrum noisefloor is > much more flat). > For this reason, I am asking if someone can provide a plot of the spectrum > of both channels at a sampling rate of 100 Msps. > > Thank you. > > Regards, > Carlo Venier > > I can't get to this myself until probably the weekend. > > It is conventional when measuring receiver/filter response to use a noise > source that produces enough power to be > considerably *above* the noise figure of the receiver. At max gain, the > TwinRx has a noise figure of about 4-5dB, > and a terminator is a noise equivalent of 3dB. A noise source of +15dB > ENR would be useful here. > > What I have noticed is that at narrower bandwidths (10MHz), the response > is non-flat (ignoring the inevitable edge roll-off) > at a level of about 1dB. That isn't unexpected when you have a long > filter chain. > > > > > > > > Il gio 22 giu 2023, 15:30 Marcus D. Leech <patchvonbr...@gmail.com> ha > scritto: > > On 22/06/2023 05:48, Carlo Venier wrote: > > Good morning, > > I have some questions regarding the TwinRX daughterboard. > > 1) With the following settings: > Master Clock Rate = 200 MHz > Tuning Frequency = 2450 MHz > Sample rate = 100 Msps > Gain = 93 > Every RF connector is 50 Ohm terminated > LO_OFFSET disabled > > I get a non-flat spectrum. In particular, the center portion of it is up > to 5 dB higher than at +/- 40 MHz from the tuning frequency. You can see > this behaviour in the attached figures. > > 2) The two channels have a different behaviour below 1.8 GHz (for example > at 450 MHz). The spectrum of channel :1 is a lot more wavy than the > spectrum of the channel :0. You can see this behaviour in the attached > figures. > In any case, also below 1.8 GHz the delta between min and max level is > about 5 dBs. > > 3) Sometimes, the DC component pops up. As the TwinRX channels are a > superheterodyne receivers, I don't expect that the DC component appears, > but it happens. > > Is it possible that the TwinRX I'm using is not working properly? Or am I > missing some settings? > > 4) Can somebody provide a screenshot of the spectrum of the TwinRx > channels with the same settings at 1) for the tuning frequencies of 450 MHz > and 2450 MHz? > > Thank you. > > > Best regards, > > Carlo Venier > > > > _______________________________________________ > > USRP-users mailing list -- usrp-users@lists.ettus.com > > To unsubscribe send an email to usrp-users-le...@lists.ettus.com > > What happens if you turn the gain down about 5dB? What are you using to > provide a perfectly-flat noise input? > > What are you using to collect the data that produced these plots? > > The analog bandwidth of the TwinRX is limited to 80MHz, and uses SAW > filters at the 1st IF (TA0824A and TA0851A, depending > on whether it's operating in "low band" or "high band"). The final IF > filter is a discrete multi-stage design, but I don't > know what it's filter shape is. > > IN ADDITION, there are RF input filters that are selected based on the > tuned RF frequency. The end result of all of these > filters is that the final filter "shape" is the convolution of all of > those individual filter shapes. > > I'm not sure what you mean by "sometimes the DC component shows up". I > would expect there to be spurs at some points-- > for example those that are exact multiples of internal clocks, etc. > > _______________________________________________ > USRP-users mailing list -- usrp-users@lists.ettus.com > To unsubscribe send an email to usrp-users-le...@lists.ettus.com > > >
_______________________________________________ USRP-users mailing list -- usrp-users@lists.ettus.com To unsubscribe send an email to usrp-users-le...@lists.ettus.com
