Bob, The 84MHz bandwidth constraint is because of the analog bandpass filter [1] on the UBX's RX signal path [2]. I'd guess that UHD will yell at you if you feed in an invalid bandwidth, but I've never tried it. If I remember correctly, you can sample at rates that aren't an even division of the MCR, you'll just end up using a CIC filter that causes rolloff in your spectrum [3]. - Sam [1] https://www.mouser.com/datasheet/2/412/55916-1504717.pdf [2] https://files.ettus.com/schematics/ubx/UBX-160_revE.pdf [3] https://witestlab.poly.edu/blog/why-does-my-received-spectrum-droop-at-the-edges/ On Wed, Apr 8, 2020 at 4:49 PM Tillson, Bob (US) via USRP-users < usrp-users@lists.ettus.com> wrote: > so with the UBX-160 on an X310, there is the following caveat: > > > > * The UBX 160 transmitter path has 160 MHz of bandwidth throughout the > full frequency range of the device; the receiver path has 84 MHz of > bandwidth for center frequencies from 10 MHz to 500 MHz. > > > > I guess my question is how does this manifest itself? > > > > If I ask for 100 MHz of BW, do I get 84 or does it fail? > > > > How would I get 84 given the requirement of sample rate be an even divisor > of 200 MHz clock? > > > > If I wanted 100 in that range, would there be any way to get it from a > single channel in that band? Most other cards don’t seem to have the BW in > that range. > > > > Thanks, > _______________________________________________ > USRP-users mailing list > USRP-users@lists.ettus.com > http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com >

 so with the UBX-160 on an X310, there is the following caveat:____

 __ __

 * The UBX 160 transmitter path has 160 MHz of bandwidth throughout
 the full frequency range of the device; the receiver path has 84 MHz
 of bandwidth for center frequencies from 10 MHz to 500 MHz.____

 __ __

 I guess my question is how does this manifest itself?____

 __ __

 If I ask for 100 MHz of BW, do I get 84 or does it fail?____

 __ __

 How would I get 84 given the requirement of sample rate be an even
 divisor of 200 MHz clock?____

 __ __

 If I wanted 100 in that range, would there be any way to get it from
 a single channel in that band?  Most other cards don’t seem to have
 the BW in that range.____

 __ __

 Thanks,____

 _______________________________________________
 USRP-users mailing list
 USRP-users@lists.ettus.com <mailto:USRP-users@lists.ettus.com>
 http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

Hi Bob, hi Sam, top of my head, UBX-160 doesn't even have adjustable bandwidth (basically, of the modern devices, only AD9xxx-based systems have). So, yeah, you'll always get a two-sided84 MHz analog bandwidth (that's how you get 160 MHz in complex baseband). You'll notice when oversampling at 200 MS/s that analog filters are quite a bit worse than what's relatively easy to do in digital filtering in an FPGA at these rate. On classic Gen-3 DDC chain (I do think that's the same for the newer RFNoC chain), yes, there's two 2-decimating half-band FIRs with upwards of 40 taps, and a CIC to do the rest of the decimation. So, if you use a sampling rate that divides the master clock rate by 4, you get two HBs, if it doesn't divide by 4, but by 2, you still get one of these nicer HBs, and for odd factors, you'll fully have to rely on the CIC. That's not necessarily terrible, but CICs do have sinc²-y shape, so that's a bit of rolloff at the edges. Cheers, Marcus On 10.04.20 05:44, Sam Reiter via USRP-users wrote: > Bob, > > The 84MHz bandwidth constraint is because of the analog bandpass filter > [1] on the UBX's RX signal path [2]. I'd guess that UHD will yell at you > if you feed in an invalid bandwidth, but I've never tried it. If I > remember correctly, you can sample at rates that aren't an even division > of the MCR, you'll just end up using a CIC filter that causes rolloff in > your spectrum [3]. > > - Sam > > [1] https://www.mouser.com/datasheet/2/412/55916-1504717.pdf > [2] https://files.ettus.com/schematics/ubx/UBX-160_revE.pdf > [3] https://witestlab.poly.edu/blog/why-does-my-received-spectrum-droop-at-the-edges/ > > On Wed, Apr 8, 2020 at 4:49 PM Tillson, Bob (US) via USRP-users > <usrp-users@lists.ettus.com <mailto:usrp-users@lists.ettus.com>> wrote: > > so with the UBX-160 on an X310, there is the following caveat:____ > > __ __ > > * The UBX 160 transmitter path has 160 MHz of bandwidth throughout > the full frequency range of the device; the receiver path has 84 MHz > of bandwidth for center frequencies from 10 MHz to 500 MHz.____ > > __ __ > > I guess my question is how does this manifest itself?____ > > __ __ > > If I ask for 100 MHz of BW, do I get 84 or does it fail?____ > > __ __ > > How would I get 84 given the requirement of sample rate be an even > divisor of 200 MHz clock?____ > > __ __ > > If I wanted 100 in that range, would there be any way to get it from > a single channel in that band?  Most other cards don’t seem to have > the BW in that range.____ > > __ __ > > Thanks,____ > > _______________________________________________ > USRP-users mailing list > USRP-users@lists.ettus.com <mailto: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 >