[USRP-users] Ettus UBX-160 vs SBX-120
marcus.mueller at ettus.com
Tue Jul 4 04:36:50 EDT 2017
So, if your ADC runs at 120 MHz, the unambiguously representable
baseband is -60 MHz to +60MHz.
Now you have a baseband signal at +65 MHz that (barely, but still)
passes through the analog anti-aliasing filter, what happens?
Exactly, it shows up at -55 MHz. If you had useful signal at -55 MHz,
bad luck, it's now superimposed with the alias of the 65 MHz signal.
Effectively, that means that if you use an anti-aliasing filter with a
passband that is exactly as wide as the Nyquist bandwidth of the ADC
you're dealing with, you can't really use the upper and lower part of
the observed spectrum, because that is overlayed with the lower and
upper transition width of the filter, respectively.
"Can't use", "Passband", "Transition Width" are all relative to what
distortions you're able to accept, by the way.
On 07/04/2017 10:26 AM, Simon Olvhammar via USRP-users wrote:
> Thank you Keyong Su Shin,
> However I suppose you mean aliasing just at the band edges?
> I get around 100 MHz of good data when using 120 MS/s on the SBX-120
> On Mon, Jul 3, 2017 at 3:18 PM, Kyeong Su Shin <ksshin at uw.edu
> <mailto:ksshin at uw.edu>> wrote:
> Hello Simon Olvhammar:
> Just one more note:
> A relevant e-mail thread achieve, since you are apparently using
> ADC sampling rate of 120MS/s on SBX-120 daughterboards:
> To get better data quality, you may want to sample faster and then
> do digital filtering -> decimation (USRP will do that for you, if
> you are fine with integer decimators). This may prevent you from
> using the 120MHz bandwidth, but improve the overall data quality.
> Kyeong Su Shin
> On Mon, Jul 3, 2017 at 1:35 AM, Simon Olvhammar via USRP-users
> <usrp-users at lists.ettus.com <mailto:usrp-users at lists.ettus.com>>
> I am not sure if I understand correctly.
> How are the ADC to slow if I use a master clock rate of 120
> MHz and sampling rate of 120 MHz on a UBX-160? I don't need
> 160 MHz bandwidth only 120 MHz.
> On 06/28/2017 04:12 PM, Sylvain Munaut wrote:
> We currently running a system with a Ettus x310 and a
> We also a have UBX-160 daughter board.
> For several reasons, mainly due to computer
> performance limitations, we can
> only run with a complex sampling rate of 120 MHz.
> My question is if I will see any improvement in
> performance, e.g. filter
> characteristics, if I instead use the UBX-160 and run
> it at 120 MHz compared
> to SBX-120 at 120 MHz or are they similar?
> You can't do that.
> If you have the master clock-rate at 120 MHz and use a 160
> MHz wide
> daugtherboard, you will get aliasing because the ADC are
> too slow.
> You'd have to go to a 100 MHz samplerate for it to work.
> (200 MHz
> sampled by the ADC then HW decimated by 2 to 100 Msps).
> USRP-users mailing list
> USRP-users at lists.ettus.com <mailto:USRP-users at lists.ettus.com>
> USRP-users mailing list
> USRP-users at lists.ettus.com
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the USRP-users