What we’re talking about today

• What’s GNU Radio?
• A very basic introduction to SDR
• A word on hardware
• Getting GNU Radio
• GNU Radio Companion
• Using Python or C++ to write a module

What’s GNU Radio?

• Free & open source software
• Provides signal-processing blocks to build software-defined radios
• Works with readily available hardware
• Runs on Linux (best), Windows and macOS
• Runs on both Intel & ARM (including Raspberry Pi!)
• Used in hobby and commercial applications

A very basic introduction to SDR

• Uses digital representations of signals
• Uses lots of math to implement things like modulation & filters

Sampling a signal

• Analog signals sampled periodically
• Sample rates should be at least twice maximum frequency (Nyquist theorem)

Sampling a signal

source: Wikipedia; https://wiki.gnuradio.org/index.php/File:Cont_to_digital.png

Sampling a signal - things to look out for

• Tradeoff between sample rate & computing requirements
• Tradeoff between cost, sample rate, and sample resolution
• Aliasing

A word on sampling

• More samples mean more expensive A/D converters
• More samples means more computing power to process the samples
• Faster sampling means higher cost

A word on resolution

• The number of discrete steps measurable in the range of a signal
• Higher resolution generally means higher cost A/D converters
• Sometimes a lower noise lower-resolution A/D is better than a higher resolution A/D!

Aliasing in sampling

source: Wikipedia; https://en.wikipedia.org/wiki/Nyquist–Shannon_sampling_theorem#/media/File:CPT-sound-nyquist-thereom-1.5percycle.svg

Aliasing in the frequency domain

source: Wikpedia; https://en.wikipedia.org/wiki/Nyquist–Shannon_sampling_theorem

Digital Filtering

• Math!
• Uses signal processing on the digitally sampled signal
• Requires fast processsing for additions & multiplications
• Digital filters can often create far steeper cutoffs for the same component cost

Modulation & demodulation

• More math!
• AM can be accomplished multiplying the carrier by the audio signal
• BPSK can be accomplished by changing the phase of the carrier
• Other modulations possible, too!

Block diagram of a typical SDR

source: Wikipedia; https://commons.wikimedia.org/wiki/File:SDR_et_WF.svg

A word on hardware

• You need a moderately fast PC or Raspberry Pi running Linux, Windows, or macOS
• For RF experiments, you need some sort of SDR hardware
• RTL-SDR is a cheap receiver (check the raffle table!)
• hackRF is a popular transceiver
• LimeSDR is another newer transceiver
• Lots of other choices; just check for GNU Radio compatibility first

What does GNU Radio give me?

• Over 400 DSP blocks, including filters, modulators, demodulators, visualization tools
• Flow-based construction of radios & subsystems
• The ability to add new blocks with code in Python or C++

Getting GNU Radio

• On Linux, use your package manager
• On Windows, use an installer
• On macOS, use MacPorts
• You can also build from the source code
• See the GNU Radio Wiki for more details

GNU Radio Companion

• Drag and drop construction of SDRs using modules
• Uses a data-flow paradigm for connecting modules
• Provides controls and visualization tools

Modeling SDRs using data flow diagrams

• Radio components connected by signal paths
• Each signal is a stream of numbers
• Modules transform the number streams in some way
• Sources provide data, sinks acccept it

Modules and variables

• Most elements of your radio are functional blocks that do something
• Variables are named containers that hold numbers (frequencies, sample rates, etc)
• Modules get connections, variables just sit there.

Graphical User Interface components

• Some sinks display data (e.g., waterfall)
• Some GUI controls to set variables (e.g., knobs)

Basic use

• Drag and drop modules
• Click to create data flow connections
• Control-F is your friend (finds modules)

GNU Radio Companion UI

source: by the author

Toolbar

source: by the author

Modules

source: by the author

Console & variable inspector

source: by the author

A simple NBFM receiver flow graph

source: by the author

Running the flow graph

• First GRC creates a Python file of your flow graph
• The Python file uses code from GNU Radio
• Then GRC invokes GNU Radio with your Python file

A simple NBFM receiver UI

source: by the author

Creating modules & radios

• Uses Python or C++.
• Define modules using XML
• Implement them using Python or C++
• Uses the same data-flow paradigm, just in code.

Modules have

• An initialization block
• A message handler
• A work block that does the math

Connecting modules in code

• Create the flowgraph in Python or C++
• Connect modules using the connect method

Defining the module

• Modules require a definition
• Definitions are in YAML or XML
• This enables GNU Radio to “see” a module

Tips

• Work through the tutorial
• Look at error messages closely
• Windows and macOS ports are flakier than Linux
• Get your SDR working with the software that comes with it first.

Resources - Hardware

• RTL-SDR receiver, frequencies vary
• hackRF transceiver, 1 MHz-6 GHz
• LimeSDR transceiver, 100 kHz-3.8 GHz

Resources - Software

• GNU Radio
• GNU Radio Live SDR Environment

Resources - Reading

• Recent editions of the ARRL Handbook
• “Digital Signal Processing and GNU Radio Companion”, QEX July/August 2014
• GNU Radio Wiki