Introductory Tutorial for SDR

and GNU Radio Beginners

GNU Radio Conference 2023
Murat Sever
Outline
● About me
● About tutorial
● GNU Radio
● Lab: Digital Signal Processing (DSP)
● Lab: Software Defined Radio (SDR)
● Lab: Wide Band Frequency Modulation (WBFM)
About me
Part-Time Lecturer at TOBB ETU, Ankara,
Turkey
● ELE361L Course / Telecom Laboratory
○ Summer 2021
○ Fall 2022
○ Fall 2023

Project Owner & Manager

● SDR-Powered Education

3
About tutorial
● Introduces fundamental DSP concepts and GNU Radio to new users
● Consists of the following lab modules based on Jupyter Notebooks
○ Lab DSP
○ Lab SDR
○ Lab WBFM
● We will use GNU Radio for
○ Exploring signals in simulation mode
○ Sound processing
○ Spectrum watching with RTL-SDR
○ Broadcast FM demodulation
Get the labs (if you haven’t already)
● Labs available @ GitHub
○ https://github.com/ARDC-TOBB-ETU/GRCon23Tutorial
Download or clone
Use README to install
● Linux/Mac
○ Run ./install.sh
● Windows
○ Install miniforge
○ Create a new environment
■ conda config --append channels conda-forge
■ conda create --name GRCon23 --file requirements.txt
Opening Jupyter Notebooks/GNU Radio
● Linux/Mac
○ Run source "${HOME}/conda/etc/profile.d/conda.sh"
○ Activate the environment conda activate GRCon23
○ Run jupyter-lab
○ Run gnuradio-companion
● Windows
○ Open a miniforge prompt
○ Activate the environment conda activate GRCon23
○ Run jupyter-lab
○ Run gnuradio-companion
Jupyter Notebook
Outline
● About me
● About tutorial
● GNU Radio (slides from a previous presentation)
● Lab: Digital Signal Processing (DSP)
● Lab: Software Defined Radio (SDR)
● Lab: Wide Band Frequency Modulation (WBFM)
GNU Radio is…

▪ A signal processing library

▪ Designed for real-time
▪ The software part of an SDR
▪ Not a radio application
▪ The tool to build your own transceivers
▪ FOSS: Free and Open Source Software
GNU Radio
▪ Open-source framework for SDR and signal processing
▪ Founded by Eric Blossom in 2001
▪ Block-based dataflow architecture
▪ Each block runs in its own thread
▪ Data flows through a graph called a Flowgraph
▪ Blocks are nodes in a Flowgraph, and perform
operations and signal processing
▪ Signals normalized between -1.0 and +1.0
▪ Similar in concept to MathWorks SimulinkTM
▪ Running C++ and Python under-the-hood
▪ Can write code directly, or use the GNU Radio
Companion (GRC) graphical tool
Basic Concept: Flow Graph
▪ Transceivers are implemented as flow graphs
▪ Similar to Simulink / schematics
▪ Define structure and parameters of blocks
Basic Concept: Block
▪ Written in C++ or Python

▪ Implement one logical step

▪ Each block run in separate thread

▪ Can be encapsulated into hierarchical blocks

Data Streams
▪ Samples are buffered

▪ Data types are color-coded

Complex Float Byte

Color Types
Click on menu item Help->Types
GNU Radio Companion
Block Library
Edit Flowgraph

Console Variables
GNU Radio Workshop // SDR Academy 17
Search Blocks
GUI Output and Instrumentation
GQRX - a GNU Radio Application
Out Of Tree Modules
▪ GNU Radio can be extended with OOTs
▪ OOTs cover more specific functionality
▪ There is a large number available
▪ CGRAN is our central database

GNU Radio Workshop // SDR Academy

GNU Radio is used by
Commercial
Defense R&D Infosec
Research Research

Academic Hardware
Research Vendors

Professional
Students Service
Hobbyists Providers
GNU Radio is an Ecosystem
▪ Active Open Source community since 2001

▪ PyBombs, OOTs

▪ GRCon since 2011

▪ GNU Radio Foundation

▪ FOSDEM SDR DevRoom

▪ GSoC, SoCIS, R&S Competition, SDR Academy

▪ GNU Radio Europe

Outline
● About me
● About tutorial
● GNU Radio
● Lab: Digital Signal Processing (DSP)
● Lab: Software Defined Radio (SDR)
● Lab: Wide Band Frequency Modulation (WBFM)
Exploration of Signals in Frequency Domain
Sampling
● Communication signals are continuous-time
● We (ADCs) take samples at regular times
● Ts is sampling period
● Fs is sampling frequency
Baseband & Bandpass
● Baseband: Information signal
● Bandpass: Communication signal
Nyquist Sampling Theorem
● The Nyquist Sampling Theorem states that a baseband, bandlimited signal
must be sampled at greater than twice the bandwidth present in the signal,
i.e.
○ fs > 2 * fmax
○ fs > 2 * (f_high - f_low)
Aliasing
● Sampling produces aliases (spectral replicas)
● To prevent aliasing Fs must satisfy Fs > 2 * BW
Nyquist Zones
● Partitions of bandwidth 0.5f s in the frequency domain
● Any signal components present in higher Nyquist Zones are ‘folded’ down into
the 1st Nyquist Zone as a result of aliasing
Folded Spectrum View
Examples of aliasing with reference to Nyquist Zones
Sampling and Aliasing
Digital Filters
● A filter modifies the frequency contents of an input signal
● Types
○ LPF
○ HPF
○ BPF
○ Notch
Filters Using GNU Radio
Multirate Signal Processing
● Multirate operations are required to change the sampling rate in a DSP
system to optimise computational efficiency
● Some example scenarios
○ To match the sampling rates of two signal paths that will be combined
○ To adjust the sampling rate closer to Nyquist when the signal bandwidth changes
○ To match the sampling rate of an external interface, such as a DAC
○ To ease analogue anti-alias or image-rejection filter requirements
Decimation
● Reducing the sample rate by an integer factor
● Retain every 𝑃𝑡ℎ sample and discard the remaining samples
● The new slower sample rate is 1/𝑃 of the original faster sample rate
Decimation
● Decimation involves two processes:
○ anti-alias low pass filtering, followed by
○ downsampling
Interpolation
● Increasing the sample rate by an integer factor
● Insert 𝑃 − 1 zeros between the original input samples and interpolate
● The new faster sample rate is 𝑃 times the original slower sample rate
Interpolation
● An interpolator is composed of
○ an upsampling operation, followed by
○ a low pass image rejection filter
Other Multirate Operations
● There are other types of operation to be aware of, beyond simple decimation
and interpolation by integer factors
● Resampling a signal by a rational fraction
○ If the sampling rate is to be changed by the ratio of two integers, e.g. a rate change from 100
MHz to 150 MHz could be expressed as R = 3 / 2 . Rational fractional rate changes can be
achieved using a cascade of an interpolator and decimator, e.g. L = 3 and M = 2 in this
example. The resulting structure can be optimised using polyphase methods.
● Resampling a signal by an irrational fraction, or by a factor that changes
over time
○ Where there is no convenient integer-based expression for the resampling ratio, or where it is
dynamic, a different type of approach is required. Popular methods include highly
oversampled polyphase filters, and Farrow structures.
Frequency Xlating FIR Filter
Suppose this is the
stations in FM radio
● Frequency Xlating FIR Filter is a
example!
block that:
○ performs frequency translation on the
signal, Our aim is to select
○ downsamples the signal by running a only one channel
decimating FIR filter on it.
● It can be used as a channelizer:
○ it can select a narrow bandwidth
channel from the wideband receiver
input.
Frequency Xlating FIR Filter
● If you have Real taps, then your FIR filter ● If you have Complex taps, then your FIR
will be symmetric in the frequency domain. filter will not have to be symmetric in the
frequency domain.
firdes.low_pass(1,samp_rate,samp_rate/(2*deci
mation), transition_bw) firdes.complex_band_pass(1, samp_rate,
-samp_rate/(2*decimation),
samp_rate/(2*decimation), transition_bw)
Frequency Xlating FIR Filter
● Decimation: the integer ratio between the ● Center frequency: the frequency
input and the output signal’s sampling translation offset frequency.
rate. ● In practice, it is the frequency offset of the
● Example: signal if interest to be selected from the
○ Input sample rate = 240000 input.
○ Decimation factor = 5
○ Output sample rate = 240000 ÷ 5 = 48000
Frequency Xlating FIR Filter
LabDSP.ipynb
Outline
● About me
● About tutorial
● GNU Radio
● Lab: Digital Signal Processing (DSP)
● Lab: Software Defined Radio (SDR)
● Lab: Wide Band Frequency Modulation (WBFM)
What is Software Defined Radio (SDR)?
“A radio in which aspects of functionality are implemented in, or controlled by,
software.”

● Flexible functionality
○ the operation of a radio can be changed without making any physical alterations to the device
● Algorithms from DSP and communications theory running as real-time
software on a CPU, GPU and/or FPGA
● Joe Mitola first coined the term in 1991
Why SDR?
● Traditional radios are hard-wired to specific frequency bands and
communication protocols
○ Fixed-function, Black Box
○ Can’t be easily modified, can’t easily access internal values and states
● SDR provides:
○ Flexibility
○ Upgradability
○ Reconfigurability
○ Lower Cost
Key SDR Parameters (Features)
• Frequency (Tuning) Range
• Instantanous Bandwidth
• Bit resolution
• Interface (USB, Ethernet, PCIe)
• Rx/Tx, half-duplex, full-duplex, MIMO
• Preselectors
• Budget: 50$-...k$
RTL-SDR
• “I smell a very cheap poor man’s SDR here☺”
• Cheap man’s radio since 2012
• Hams, DIY, hackers, makers, students,…
• Demodulator
• Named by RTL2832U chip, DVB-T
• Tuner
• R820T: 24-1766MHz
• E4000: 52-2200MHz
RTL-SDR

• Receive-only
• 8-bit ADC
• 24MHz-1.75GHz (depends on tuner chip)
• 2.4MSPS BW (stable) upto 3.2M
• “HamItUp” upconverter
• HF coverage
RTL-SDR Driver Installation#1 - Windows
● Plug in your dongle
● Right click zadig.exe file and select "Run as administrator".
● In Zadig, go to "Options->List All Devices" and make sure this option is
checked. If you are using Windows 10 or 11, in some cases you may need to
also uncheck "Ignore Hubs or Composite Parents".
RTL-SDR Driver Installation#2 - Windows
● Select "Bulk-In, Interface (Interface 0)" from the drop down list. Make sure it is
Interface 0 (ZERO), and not "1".
RTL-SDR Driver Installation - Linux
● Linux users may blacklist RTL so that default DVB-T driver is not loaded when
dongle is plugged in.
○ # cd /etc/modprobe.d/
○ # sudo gedit blacklist-rtl.conf
○ # append: blacklist dvb_usb_rtl28xxu
○ OR
○ # echo "blacklist dvb_usb_rtl28xxu" >> /etc/modprobe.d/blacklist.conf
LabSDR.ipynb
Outline
● About me
● About tutorial
● GNU Radio
● Lab: Digital Signal Processing (DSP)
● Lab: Software Defined Radio (SDR)
● Lab: Wide Band Frequency Modulation (WBFM)
Frequency Modulation (FM)
FM Radio Multiplex
● It is common practice to multiplex multiple information signals together before
performing modulation, as this allows for multi-channel transmission using
one carrier.
Broadcast FM (WBFM)
LabWBFM.ipynb
 Thanks!
[email protected]