Re: [Discuss-gnuradio] Is that possible combining usrp_fft.py and usrp_spectrum_sense.py to do spectrum analysis and observe in GUI in a wider frequency range?

[Brook Lin]

Thanks for your reply and kind reminder, Eric. I think usrp_spectrum_sense.py
works fine on my computer without segmentation fault. 

I observe that there is a set_next_freq() function in usrp_spectrum_sense.py
file, which is used to tune the center freq used in set_freq() function,
isn't it? The target_freq is used to set the center freq of the USRP to
determine the value of DDC.

Here is the build graph part that I changed and I thought it fit my project
purpose: observing fft in larger bandwidth. 
========================================
                 # build the graph

                self.u = 
usrp.source_c(which=options.which,decim_rate=options.decim)
                if options.rx_subdev_spec is None:
                        options.rx_subdev_spec = pick_subdevice(self.u)
                self.u.set_mux(usrp.determine_rx_mux_value(self.u,
options.rx_subdev_spec))

                if options.width_8:
                        width = 8
                        shift = 8
                        format = self.u.make_format(width, shift)
                        print "format =", hex(format)
                        r = self.u.set_format(format)
                        print "set_format =", r
            
                # determine the daughterboard subdevice we're using
                self.subdev = usrp.selected_subdev(self.u, 
options.rx_subdev_spec)

                #input_rate = self.u.adc_freq() / self.u.decim_rate()
                input_rate = self.u.adc_freq() / options.decim

                #************************************************************
                # Set the freq_step to 75% of the actual data throughput.
        # This allows us to discard the bins on both ends of the spectrum.
                
                self.min_freq = options.freq_min
                self.max_freq = options.freq_max
                self.freq_step = 0.75*input_rate
                self.min_center_freq = self.min_freq + self.freq_step/2
                nsteps = math.ceil((self.max_freq - self.min_freq) / 
self.freq_step)
                print "nsteps is ", nsteps
                self.max_center_freq = self.min_center_freq + (nsteps * 
self.freq_step)
                for i in range(nsteps):
                        if i == 0:
                                self.next_freq = self.min_center_freq
                        else:
                                self.next_freq = self.next_freq + self.freq_step
                        
                        print self.next_freq, "***", i, "***"

                        if not(self.set_freq(self.next_freq)):
                                print "Failed to set frequency to", 
self.next_freq

                        self.set_freq(self.next_freq)

                #************************************************************
                        self.scope = fftsink2.fft_sink_c (panel, title="Signal
Detection",fft_size=1024, sample_rate=input_rate,
baseband_freq=self.next_freq)

                        self.connect(self.u, self.scope)

                        self._build_gui(vbox)

                        # set initial values
                        .........................
                        .........................
==============================================
However, by given the lower bound and the upper bound of the freq range,
next_freq can be get and "Failed to set frequency to", self.next_freq" is
also get. and all the plots are the same, seems
self.set_freq(self.next_freq) doesn't work. Do you know why it failed to
set?

The 2nd question is, I know I call self._build_gui(vbox) within for loop. So
in GUI window, I will have several fft plots regarding the deferent center
freq. How can I get one GUI with one plot and show all the spectrum, say all
FM Stations?

I am new to Gnuradio, so any reply is helpful for me.

Thanks,
Brook


Eric Blossom wrote:
> 
> Yes.  Note however that at this moment there is a known problem with
> usrp_spectrum_sense.py (actually believed to be in gr_feval_dd) that
> causes a segmentation fault.  See ticket:199
> http://gnuradio.org/trac/ticket/199
> 
> Eric
> 
> _______________________________________________
> Discuss-gnuradio mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
> 
> 

-- 
View this message in context: 
http://www.nabble.com/Is-that-possible-combining-usrp_fft.py-and-usrp_spectrum_sense.py-to-do-spectrum-analysis-and-observe-in-GUI-in-a-wider-frequency-range--tp14183248p14260983.html
Sent from the GnuRadio mailing list archive at Nabble.com.