Thanks for the response.
I am using a USRP2 N210 with the shortwave receiver daughterboard. When
I
was playing around, I was getting an error when I set the sample rate on
the USRP source to 30MHz, and was told it had something to do with
dividing
100 by 30 and not getting an integer (you get 3.33333), so it was
recommended I use 25 or 50, with 25 giving more resolution on the
results.
If I am setting the rate to 50MHz on the source and 30MHz on the FFT
sink,
and I am actually seeing (and ultimately recording with a file sink) the
whole shortwave spectrum, I will be satisfied.
Thank you.
Paul B. Huter
Andrew D. [email protected] wrote:
Hello, first I think you meant sample rate, not size. Who said you cant
use
30Mhz, what kind of hardware are you using? Lastly the FFT algorithms do
not deal with frequency in that way, you are still seeing 0-50Mhz
spectrum,
just the label now says 50Mhz.
Andrew
On Sun, Nov 17, 2013 at 8:51 AM, Paul B. Huter [email protected]
wrote:
Thank you.
Paul B. Huter
So you set the sampling rate to 50 Msps, which produces complex 8-bit
samples, and so it represents 50 MHz of spectrum.
When you pass this to the FFT sink, you will see all 50 MHz. You’re at
complex baseband, so you’ll see -25 to +25 MHz. Even if you set the
sampling rate of the sink to 30 Msps, you are still seeing 50 MHz. The
sampling rate to the sink does nothing but set the x-axis. So if you
set it to 30, the x-axis will represent -15 to +15 MHz, but a signal
at -15 MHz will actually be at -25 MHz.
You have to go through a sample rate changing block, like a decimating
FIR filter (which only gives you integer decimation rates), the
rational resampler block (which gives you a rate change of
interp/decim where interp and decim are integers), or the fractional
resampler or PFB arbitrary resampler (where the rate you set is a real
number). See the manual for how to use any of these blocks
(GNU Radio Manual and C++ API Reference: Main Page).
Tom