SDR Presentation - January Meeting
The below slides were shown at the January general meeting in 2016 of the Northern Alberta Radio Club as presented by Stephen Olesen, VE6SLP. Folowup questions are welcome.
SOFTWARE DEFINED RADIO
NARC PRESENTATION – JANUARY 2016
STEPHEN OLESEN – VE6SLP
WHAT IS SDR?
• SDR – Software Defined Radio
• Instead of using discrete components or dedicated ICs for tuning, reception and demodulation,
software may take on many of these roles.
• Hardware component tends to only provide RF to IF conversion (including a zero-IF).
• Uses software to demodulate received signals.
• Uses software to modulate transmitted signals.
WHY USE SDR?
• SDR provides flexibility by removing restrictions imposed by hardware in the signal chain.
• A single RF frontend can be used by multiple signal processors as the incoming RF is digitized early on in
• General purpose computers can be used with high performance to analyze and work with complex
signals or multiple (independent or dependent) signals at once.
• Digitization of the RF signal in both directions allows for very accurate reproduction and analysis of
signals without introducing further noise in analog components.
• The hardware involves:
• An RF frontend, generally can be quite wideband and often found as a monolithic IC.
• Analog to digital conversion, either from an IF or from baseband.
• In a transmitter, a digital to analog converter is also used.
• May send baseband to an IF mixer or direct to RF.
• Components are simpler as less amplification needs to take place before the ADC, filtering and
processing are done digitally in software.
• The software only needs to work with a digital stream of data which is the RF data converted from the
• Software can receive real samples or complex samples (I/Q – in-phase/quadrature).
• Can run as a service, as a user application, in dedicated hardware (such as FPGAs, DSPs or even
• Data can be transmitted digitally over networks without loss of signal fidelity.
• Examples of software include:
• SDR# (SDR Sharp, Windows)
• SDR-Console (Windows)
• HDSDR (Windows)
• GNU Radio (Linux)
• Linrad (Linux)
• SDR-Shell (Linux)
• Gqrx (Linux, Mac)
• WebSDR (Linux, provides web interface to multiple users)
• The RF is received on an antenna.
• RF frontend usually provides normal frontend filtering (bandpass, low pass, etc.)
• RF is either directly mixed to baseband or to an IF
• IF is either directly sampled or mixed to baseband
• Samples are sent to the computer as either real or complex samples
• Real can be converted to complex, and vice versa
• Baseband can easily be sampled with a computer sound card
• IF generally requires dedicated hardware for the conversion of the high frequency IF
• Receive bandwidth limited by the analog to digital converter (or sound card)
DEALING WITH THE DIGITAL SIGNAL
• The real samples generally are converted to I/Q samples, providing a 90 degree out of phase sample of
the RF data
• The complex sample is able to produce amplitude and phase across the entire received spectrum while
eliminating images due to mixing
• Real samples when converted have no direct phase information and will have an image signal present
• This can be removed using digital filters
• Digital signal can be sent through extremely complex or many-poled filters providing sharp, narrow
filters on the raw RF data
• Original signal data can be used in multiple filters/demodulators simultaneously across the entire
SAMPLING RATE, NUMBER OF BITS, BANDWIDTH
• Since the RF signal is digitized, the analog to digital conversion will have a fixed sample rate.
• Depending on the hardware (ADC and DAC), this sampling rate may be at sound card speeds (48 or
96kHz, or 96kSPS/192kSPS) or higher (such as the BladeRF at 40MSPS)
• Usable bandwidth is half the sampling rate (192kSPS provides 96kHz bandwidth, 40MSPS provides
• Bit length provides the total signal amplitude range and relates strongly to the achievable SNR, dynamic
range, accuracy and overload characteristics. 8 bits is common, providing 256 levels of signal strength.
• Simple USB devices known as RTL-SDR (RTL2832) are easy to find and a low cost way to experiment with reception only SDR.
• Available in Canada for as low as $15, but generally in the range of $25-$30.
• Available with a variety of tuner chipsets, most common these days is the R820T.
• Connects via USB to a computer.
• Able to tune from around 50MHz to 1.5GHz and above.
• Provides 8 bit samples at up to 3.2 MSPS (usually only usable up to 2.4MSPS or so)
• The Softrock series by Five Dash (KB9YIG) are a kit version of HF SDRs which use a computer sound card for
the digitization portion.
• Low cost and a simple design allow a good understanding of the hardware side of SDR.
• Range in price from $21 USD to around $100 USD.
• Use USB for power and frequency control, regular 3.5mm audio cables for the baseband audio to be digitized.
• Compatible with a wide variety of software as the only requirement is the sound card driver.
• Capable of both transmission (low power, around 1 watt) and reception, 96kHz bandwidth (192kSPS)
• Flex Radio provides all-in-one solutions for high end performance.
• A combination of hardware based digital processing and computer based signal processing.
• Higher power outputs than the SoftRock.
• Range in price from $1000 to $11,000.
• Familiar rig layout and design with many input and output options.
• NARC owns a Flex Radio SDR at the shack.
• Some support dual RF frontends, wide bandwidth (not limited by a computer sound card)
BLADERF AND HACKRF
• Both the BladeRF and HackRF are small USB controlled SDR hardware frontends.
• Provide onboard ADC and DAC.
• BladeRF capable of full duplex communication.
• Very low output powers (10mW or less).
• BladeRF is capable of 40MSPS and 300MHz to 3.8GHz transmit/receive, 12 bits.
• HackRF is capable of 20MSPS and 1Mhz to 6GHz, 8 bits.
• BladeRF runs around $600-$1000 USD, HackRF is around $350-450 USD.
• BladeRF has an onboard FPGA, HackRF uses a CPLD.
• Designed for experimentation, protocol design, advanced decoding.
• They have been used to provide GSM/LTE cellular network services, digital TV (ATSC) transmission, and more.
USING WHAT YOU HAVE
ANY RIG USING A SOUND CARD
• Since the real idea of SDR is that software handles all the signal processing, any RF frontend can
theoretically be used to get the signal into the software.
• Using a normal radio (HF, VHF, UHF, etc.) with an audio connection to a computer can provide low-
bandwidth SDR services.
• Bandwidth limited by the in-radio filters (USB, LSB, FM, etc.)
• This sort of setup is commonly found with PSK, JT65, WSPR.
• The RF frontend is a full radio capable of demodulation of audio, but is unaware of what that audio is.
The computer then processes the input like an SDR to provide these digital modes.
• Fldigi and Ham Radio Deluxe provide convenient interfaces to classic radios via the sound card (and
often a connecting interface to provide PTT and level control).
USB AND SOUND CARD INPUTS
• The simplest SDR RF frontend outputs the baseband signal via regular analog audio signals to a
computer’s sound card.
• Limited bandwidth (audio bandwidth ranges).
• I/Q imbalance is more common and harder to fix.
• USB inputs vary, with devices like the RTL-SDR, Flex, BladeRF and HackRF having integrated ADCs/DACs.
• Some USB SDR devices use built-in soundcards which connect over USB, but are still limited to audio ranges.
The RigBlaster series is an example of an SDR-like interface with integrated sound card.
• Starting with equipment you already have, experiencing the basics of SDR and the integration of
computers with radios, can cost nearly nothing.
• Using an RTL-SDR based USB adapter can get you up and running for under $25.
• Adding on transmit capabilities can start as low as $70 for the SoftRock, up to the Flex Radio series at
many thousand of dollars.
• Because the RF portion is relatively straight forward, designing your own SDR frontend can be done (this
generally involves oversampling your desired frequency to get I/Q samples and mixing with an IF to
convert to baseband, and not much else).
• In general, SDR provides for a clean output signal.
• Sharp and complex filters can be used in software to remove as much extra sideband/noise as possible.
• Extremely complex signals can be sent digitally to a high performance DAC which produces the voltage
output to the RF transmitter without needing complex analog filters or integration to join multiple
• Signal noise is influenced by the bit size and sample rate used to generate the signal.
• Since the RF frontend is straightforward (ADC/DAC, mixer and amplifier), the signal path is short,
preventing extra oscillators and parasitics from compromising the signal.
SOME EXAMPLE USES
• ADS-B (1090MHz) plane transponder tracking
• Multi-channel trunking scanners
• Wideband band listening / waterfall (including multi-user demodulation), ie. WebSDR
• Remote listening / transmitting
• PSK, JT65, WSPR (all at the same time, with one antenna and SDR front end)
• Experimental digital or analog mode design
• Satellite tracking and telemetry
• Discovering the RF world around you without spinning the dial (a wide bandwidth receiver showing
20MHz of RF spectrum at once gives a nice overview of nearly all the HF spectrum at once)
• Or feel free to send me questions at [email protected]c.net
Written on January 20, 2016 by Stephen Olesen, VE6SLP