Lab memo lab 1 - ISY: Communication Systems
Contents
1. subdev self u selected_subdev subdev_spec self u set_mux self u determine_tx_mux_value subdev_spec self u tune 0 subdev carrier subdev set_gain 0 self connect self src self scale self fir self uscale self f2c self u In this code the line blks2 rational_resampler_fff fir_interp fir_decim produces a ra tional resampler block that also lowpass filters the resulting signal with normalized cut off fre quency 0 4 relative to the resulting sampling frequency The parameters are the interpolation and decimation rates of the rational resampling The rational resampler is followed by a scaling by 215 and conversion to complex necessary because the USRP accepts complex samples in the range 21 215 1 When the USRP is instantiated the interpolation rate and number of transmitter paths are set The next five lines obtain a reference to the daughterboard connection at side A configures the FPGA mux to use its DAC path for the incoming samples tunes the carrier frequency and sets the gain to 0dB which makes sample value 2 correspond to 1 V output It is also possible to use 1 0 as argument to usrp selected_subdev in order to use a daughterboard at side B Connect the I and Q transmitter connectors TX A and TX B of the LFTX daugherboard to the two inputs of the oscilloscope Study the signal on the oscilloscope and verify that it has amplitude 0 1 V and frequency close to the expected 440 Hz You may see the tone on2. Software Defined Radio Lab 1 Getting in Touch with GNU Radio and the USRP Version 0 3 Anton Blad and Mikael Olofsson September 19 2011 1 Introduction The purpose of this lab is to give a first impression of signal processing using GNU Radio and the USRP It covers the basics of the GNU Radio flowgraph descriptions and gives an understanding of the transmission and reception paths of the USRP You will start with a very simple flowgraph that produces a sine wave alters the amplitude and plays the signal on the speakers You will then alter the flowgraph to instead send the signal to the USRP and view the output using an oscilloscope When you have a basic grasp of the processing that happens in the transmission chain you will write a new flowgraph that receives a signal using the USRP and shows its FFT on the screen You will also play the received signal on the speakers Finally you will replace the sine wave source with a wave file reader and add configurable filters at the receive side with which it is possible to change the bass and treble Along the way you will also see how you can present graphs of signals using a GUI 2 Preparations Those preparations are all related to the receiver that we construct in Sections 4 3 and 4 5 Choose the interpolation rate usrp_interp of the USRP Based on that compute the interpolation rate fir_interp and the decimation rate fir_decim of the rational resampler Notice that all involved s
3. ampling frequencies are integers and all interpolation and decimation rates have to be integers 3 Setup The following hardware is needed for this lab A computer with USB 2 0 Speakers connected to the line out A USRP with an LFTX LFRX daughterboard pair at the A side e e e e An oscilloscope with two channels You need GNU Radio installed The lab is tested and ensured to work with GNU Radio 3 3 0 and assumes that it is installed in usr local gnuradio 3 3 0 3 1 Initialization If you are doing this lab in CommSys research lab Log in to one of the lab computers as the user SDR lab user Your lab teacher will give you the password Your lab computer is prepared with all needed files in the directory home sdrlabuser labfolder lab1 Open a terminal and cd to that directory You do not need to worry about making changes to files there That directory will be updated for the next occation If you are not following these instructions in CommSys research lab Open a terminal Create a directory labfolder somewhere and cd into it Download and extract the accompanying lab files by executing the following commands wget http www commsys isy liu se SDR labs gr lab1l tar gz tar zxf labi tar gz cd labi wget http www commsys isy liu se SDR labs gr test wav RAR SF All files needed in this lab are now in the directory lab1 that you just created 4 First Steps Open the file base py which contains the Python code for th
4. ates The rational resampler has to do the correct thing to ensure that the samling rates match Even if you have done everything correctly you may still see spurios printouts of that kind in the terminal The reason for that is that the clocks of the USRP and the audio sink are not locked Eventually that leads to overruns or underruns in the software or hardware buffers downsampled signal 48kHz a gr fir_filter_fff bass gr fir_filter_fff mid a gr fir_filter_fff treble gr multiply_const_ff gr multiply_const_ff gr multiply_const_ff bassgain midgain treblegain TT a gr add_ff add audio sink sink Figure 2 USRP audio player with equalizer which is what the printouts are reporting Building a rate adaptor block to compensate for this is possible but outside the scope of this introductory lab 4 6 A USRP equalizer In the last part of this lab you will change the sound source from a simple sinus signal to a music sample read from a wave file Begin by copying your usrp_rx py to wav_rx py Notice the code starting with tb start That code provides a commandline interface for setting equalizer parameters Use the file wav_tx py and insert code from your usrp_tx py into wav_tx py Modify wav_tx py to use gr wavfile_source instead of gr sig_source_f self src gr wavfile_source test wav True gr wavfile_source produces float samples so no type con
5. both of the channels even though the imaginary part of the source signal is zero This is because the CORDIC in the AD9862 chip imposes a complex phase shift of the signal Unfortunately it is not possible to clear the phase accumulator register other than by restarting the USRP so tuning to 0 Hz will produce a complex output even if the input is real So if you do this as the first thing you do after powering up the USRP than the signal should only be available on the I connector To see the effect of the complex phase shift set the carrier frequency to 1 Hz and rerun the program Next set the carrier frequency to 1 MHz and explain the results 4 4 Spectrum Analyzer using USRP Now connect the I and Q transmitter connectors TX A and TX B of the LFTX daugherboard to the I and Q receiver connectors RX A and RX B of the LFRX daugherboard via the two channels of the oscilloscope The amplitude should then drop to half since the output and input impedances of the daughterboards are both 50 Q The USRP and GNU Radio can be used as a spectrum analyzer A general analyzer built on the old wxgui GUI library is included in the GNU Radio distribution with the name usrp_fft py Set the transmitter carrier frequency to 100 kHz and run usrp_fft py d 64 f 0 to see the spectrum of the received signal Explain the results The frequency axis can be changed by instead running usrp_fft py d 64 f 100000 4 5 Building a Receiver In this
6. e flowgraph shown in Fig 1 The simple flowgraph is composed of the following blocks e a signal source gr sig_source_f producing a stream of floats e a multiplier gr multiply_const_ff accepting float inputs and producing float outputs e a signal sink audio sink accepting a stream of floats and sending them to the speakers Study the code and then try to run it which should produce a tone in the computer s speakers You run it from the terminal like so base py Note Even though the gr sig_source_f block is created with a sampling frequency of 48 kHz this does not ensure that samples are produced at this rate The block has no notion of time and the sampling_freq argument is only used to determine the successive phase shifts between the samples The audio sink block is the only block affecting the processing rate in the system by having a fixed rate at which samples can be fed to the soundcard gr sig_source_f gr multiply_const_ff audio sink src scale sink Figure 1 Simple flowgraph generating a sine signal and playing it on the computer speakers There is also the file base_gui py which does the same thing as base py but also has a GUI with two graphs The upper graph shows the signal in the time domain the same way as an oscilloscope would The lower graph shows a power spectrum of the signal in dB scale based on an FFT Open that file to see how that is accomplished The differences compared to base py are clearly i
7. efore be using the block blks2 rational_ resampler which combines interpolation and decimation to perform rational resampling The following relation has to hold fir_interp usrp_rate usrp_interp fs_src fir_decim All interpolation and decimation rates have to be integers 4 3 Building a Transmitter Begin by saving the file base py with a new name usrp_tx py and do the changes in this file Alternatively if you want a GUI do this based on the file base_gui py Then perform the following changes to the code To interface to the USRP in the Python code and to be able to use blks2 rational_resampler first import the relevant modules from gnuradio import usrp blks2 Next remove the audio sink instance and the self connect command and replace it by the following code Lines two to five below should have been done as preparation to this lab These should all result in integers As a result of the calculations some of them may be represented as floats though and need to be transformed into integers which is done using the python internal function int usrp_rate int 128e6 usrp_interp lt choose gt fir_interp lt compute choose gt fir_decim lt compute choose gt carrier 0 self fir blks2 rational_resampler_fff fir_interp fir_decim self uscale gr multiply_const_ff 2 15 self f2c gr float_to_complex self u usrp sink_c interp_rate usrp_interp nchan 1 subdev_spec 0 0
8. ndicated in the code using comments 4 1 Connecting the USRP Make sure that an LFTX LFRX daughterboard pair is connected to the A side of a USRP unit Then connect a power cable to the USRP and connect the USRP to the PC using a USB cable You should see a LED blinking with approximately 3 Hz frequency on the USRP motherboard right next to the power connector This indicates that the USRP is ready to be initialized Ensure that the computer can find the USRP by running usr local gnuradio 3 3 0 bin lsusrp The command should print the USRP its serial number and the connected daughterboards After this the LED blinks with approximately 1 Hz frequency This indicates that the USRP has been initialized 4 2 Interfacing to the USRP In the example files base py and base_gui py the signal source is generating samples at a rate of 48kHz which is much lower than the USRP can support This is the value of the variable fs_src used below The D A conversion rate is fixed at 128 MHz and the USRP supports any interpolation rate f 4n where n 1 128 In this particular lab it is not possible to use the smallest allowed values of this interpolation rate We need to choose a suitable interpolation factor for the USRP and in addition to that we also need to upsample the signal in the host computer before sending it to the USRP In our case the interpolation rate needs to be rational since 48 103 does not divide 128 10 4 We will ther
9. task you will build a receiver for the signal generated with usrp_tx py The USRP ADC conversion rate is fixed at 64 MHz and you need to bring this down to the 48 kHz that the audio card can sink The FPGA firmware supports any even integer decimation factor in the range 4 256 Choose a suitable FPGA decimation factor and then compute the additional decimation needed at the host Call the receiver program usrp_rx py You will have a use for some of the following blocks usrp source_c gr complex_to_real gr complex_to_imag gr multiply_const_ff blks2 rational_resampler_fff audio sink Change the carrier frequency back to 0 then build and run the flowgraph This should produce the 440 Hz tone in the speakers Note There is no phase synchronization between the transmitter and receiver and hence the received signal will not necessarily be real However we know that the source signal was real The goal in this task is not to produce a perfect receiver Therefore you can allow yourself to just use complex_to_real and hope for the best Note If the produced audio signal has the wrong frequency combined with that you notice a stutter or buzzing of the audible signal then most likely you need to check your interpolation and decimation rates used by the rational resampler You would then also notice print outs like repeated Uo in the terminal The reason for this is that the flowgraph contains two blocks with fixed sampling r
10. version is needed In wav_rx py insert three FIR filters as shown in Fig 2 one each for bass mid and treble You may use the following helper functions to design the filters i e to determine the filter taps See the manual for further information gr firdes low_pass gr firdes band_pass gr firdes high_pass They return tap values as a list which is then used as the argument taps when instantiating the filter with gr fir_filter_fff decim_rate taps where decim_rate should be 1 in this case Suitable bandwidths and transition bands are given in the following table Filter Bandwidth Transition width Bass 0 200 20 Mid 200 2000 100 Treble 2000 24000 400 Run your sender and receiver to verify that the code works as intended 5 Resources Online USRP FAQ http gnuradio org redmine wiki gnuradio UsrpFAQIntro Firas Abbas Hamza The USRP under 1 5X Magnifying Lens http gnuradio org redmine attachments 129 USRP_Documentation pdf Third party documentation of USRP e GNU Radio C block documentation file usr local gnuradio 3 3 0 share doc gnuradio 3 3 0 html index html Firas Abbas Simple Gnuradio User Manual http rapidshare com files 72169239 Simple Gnuradio User Manual v1 0 pdf Third party documentation of GNU Radio This one is a bit out of date but still useful A printed copy is available at the lab desk
Download Pdf Manuals
Related Search