Like almost everything else in electronics, radios are becoming processors with software that communicate via a small amount of RF I/O circuitry. Surely, then, the rise of software-defined radio (SDR) should come as no surprise.
Digital signal processing (DSP) lies at the heart of SDR. Add to that the arrival of faster analog-to-digital and digital-to-analog converters (ADCs and DACs) and processors, and SDR becomes more viable for a wider range of applications. Simply put, software continues to push hardware aside, assuming more and more processing functions.
SDR first showed up in military equipment, but it’s now used in most cell phones. It also is ideal for future public-safety communications by providing a way to deal with the myriad air interfaces and frequency spectra used by first responders in disaster situations. SDR techniques have even joined the mainstream, with services like ham radio adopting them as prices permit.
According to the SDR Forum and the IEEE, “A software-defined radio is any radio, transmitter or receiver, in which some or all of the physical layer functions are software defined.” That means the core hardware is a processor running software that can emulate hardware functions. As a result, the signals must be digital.
The receiver must first digitize the radio signals in an ADC. In most cases, a downconverter is needed to translate the very high radio frequencies, often in the microwave region, down to an intermediate frequency (IF) that’s within the range of a decent ADC.
Today, many SDR receivers convert directly to baseband. Once the ADC converts the signals into digital form, the processor and software can take over. DSP software routinely implements receiver functions like filtering, noise suppression, and demodulation.
The digital signal processor develops the signals to be transmitted, along with any modulation. A fast DAC then converts these signals into analog form. Next, an upconverter stage translates the signal to its final higher operating frequency before it’s applied to a power amplifier and the antenna. The processor uses DSP to perform the modulation, filtering, and other functions previously implemented with analog circuits.
The most common reason for using SDR is flexibility, or the ability to change or adapt to varying radio situations. With SDR, you can accommodate virtually any modulation scheme in the same radio without adding any hardware. All you have to do is download a new software module, and you have a new radio.
Multiple modulation subroutines can exist within the code and allow an operator to change on the fly. A flexible air interface and a wide frequency range make the radio applicable for many different jobs. That’s why the military is so enamored with SDR. One radio can communicate with many different sources and terminals, lowering costs and reducing the number of radios needed.
Furthermore, SDR improves radio performance. For instance, DSP filters can make selectivity many times better than what’s achievable with inductor-capacitor (LC) or crystal filters. Brick-wall filters become a reality. Intermodulation problems can be significantly reduced. Features like automatic gain control (AGC) and noise suppression can also be improved many times over the performance produced by analog circuits.
In amateur radio, hams want the best performance with the most flexibility, and SDR provides it. Hams use multiple communications modes in multiple bands. Continuous wave or CW (Morse code), AM, FM, single side-band (SSB), double side-band (DSB), radioteletype (RTTY), and packet data are just a few of the schemes used in bands from 1800 kHz to 10 GHz.
Since the ham bands aren’t channelized, any station can operate on any frequency, making interference and closely spaced stations a challenge to overcome. Superior DSP filtering is a real blessing in most ham communications. While SDR usually costs more, at least today, it’s chosen for one or more of these benefits.
It’s important to distinguish between SDR and a softwarecontrolled radio, though. A software-controlled radio may not involve common SDR methods, except perhaps for limited use of DSP IF filtering. Instead, it’s typically a computercontrolled receiver.
With a number of models on the market, software-controlled radios use a PC to control all or most receiver functions via a graphical user interface (GUI) that emulates the receiver front panel with its tuning dial, S-meter, knobs, and switches. By pointing and clicking with a mouse, users can change frequency, select band and mode, increase volume, and perform other operations usually actuated with a frontpanel button or knob.
With PCs so common today, it’s an easy transition from a conventional knob, switch, and LCD readout front panel to the virtual front panel with its point-and-click approach. Ten-Tec’s RX-320D and AOR’s SR2200 are just two of the software-controlled radios now available.
SDR AMATEUR RADIO EQUIPMENTM
Hams have dabbled with SDR for more than a decade. These inveterate experimenters have been home-brewing hardware of all kinds since the early 20th century, and that includes some SDR in the late 1990s. SDR is very complex and expensive, though, so it’s been limited to those hams equipped with the knowledge and money.
Continue on page 2
Until recently, only a handful of commercial SDR products emerged. Introduced in 1998, the Kachina 505DSP was the first product to use a virtual front panel on a PC. While not fully SDR, it did use DSP for second IF filtering. The 505DSP didn’t last long, and it wasn’t a commercial success. But it certainly showed what could be done, and it paved the way for other efforts.
A big breakthrough in ham SDR came in 2002. Gerald Youngblood (K5SDR) wrote the first of a series of articles on SDR basics describing the construction of a complete SDR transceiver called the SDR-1000 (“Software-Defined Radio for the Masses, Part 1,” QEX, July/August 2002) for the American Radio Relay League (ARRL). This ultimately led him to found FlexRadio Systems, producing the SDR-1000 as a commercial product.
The SDR-1000 was a true SDR product with direct conversion and DSP for filtering, modulation/demodulation, and software control. It used a PC sound card for the ADC/DAC and featured a complete virtual front panel. The PC executed all of the DSP functions. A commercial success, the SDR-1000 was available up until recently. It’s still supported, but more importantly, it led to the creation of FlexRadio’s latest product, the FLEX-5000A.
Typical of what you’ll see from other manufacturers in the future, the FLEX-5000 is a full-blown SDR transceiver (Fig. 1). It follows and improves upon the SDR-1000’s design. Specifically, it eliminates the need for the PC sound card and adds its own ADC and DAC. The PC still handles the DSP, receiver control, and display functions and communicates with the radio by an IEEE 1394 FireWire interface.
The FLEX-5000A is stealthy and probably the most powerful ham transceiver available. It covers all ham bands from 1800 kHz to 54 MHz (160, 80, 40, 30, 20, 17, 15, 10, and 6 meters). Also, it can operate in virtually all common ham modes, including CW, AM, SSB, DSB, FM, and RTTY, along with a variety of digital packet modes like PSK31. Its directconversion receiver has a unique front end (Fig. 2).
The transmitter puts out 100 W of RF power in all bands. The specifications for both transmitter and receiver in terms of harmonic and sideband suppression, third-order intermodulation distortion, sensitivity, and image rejection are world class. An impressive and flexible switching matrix allows any of three antennas to be connected to any of the two possible receivers or the transmitter.
The input to the antenna is applied to a low-noise amplifier (LNA) and bank of input bandpass filters (BPFs) for each major ham band. After the LNA, a low-pass filter (LPF) serves as an anti-alias filter. The signal is then applied to the quadrature switching detector (QSD). While the first stage in most receivers is a conventional mixer, the FLEX-5000A uses a unique switching circuit called a Tayloe detector, named after its inventor, Dan Tayloe (N7VE) (Fig. 3).
The circuit is essentially a one- to four-MOSFET switching demultiplexer with a capacitor on each of the four outputs. The demux switch commutates at a rate four times the desired signal detection frequency. The output resistance of the driving source and the switched capacitors form a selective bandpass filter. The switching rate sets the filter’s center frequency, and its bandwidth is a function of the resistance and capacitance values.
But more importantly, the outputs are shifted 90° from one another as they sample the input signal. Combining the output signals in low-noise operational amplifiers produces the familiar in-phase (I) and quadrature (Q) signals at baseband, which makes the Tayloe detector a direct-conversion circuit. Given the I/Q signals, the demodulation of any kind of signal is possible.
The I and Q signals are then sent to a 24-bit sigma-delta (S-?) ADC. Sampling rate is selectable between 192, 96, or 48 ksamples/s. The ADC outputs go to the PC via the FireWire interface. An audio codec takes the demodulated, filtered, and otherwise processed signal from the PC and sends it to the audio amplifier, speaker, or headphones.
The Tayloe detector’s local-oscillator signal is derived from a direct digital synthesizer (DDS1), whose input is derived from a 10-MHz temperature-compensated crystal oscillator (TCXO) driving a phase-locked loop (PLL) with a 500-MHz voltage-controlled crystal oscillator (VCXO) output. The DDS1 oscillator can set the receiver to any frequency with an increment as small as 1 Hz.
A second receiver that’s identical to, and independent of, the main receiver can be installed as an accessory. A separate local oscillator (DDS2) is used as well. For the transmitter, the audio-frequency (AF) input signal from the microphone is digitized in the codec and the PC for modulation. The signal generated in the codec is then sent to a DAC, where it’s converted into the I and Q signals. A quadrature switching exciter (QSE) serves as the modulator.
Continue on Page 3
After amplification and low-pass filtering, the signal is sent to the final power amplifier (PA), which is a push-pull MOSFET class AB amplifier with an output of 100 W. The signal is low-pass filtered one more time to clean out the harmonics before making its way to the switching matrix and the antenna. DDS1 sets the transmitter frequency.
While the FLEX-5000A’s hardware is impressive, its PowerSDR software makes it all work. It performs all of the DSP functions and implements the front panel and all other radio controls (Fig. 4). Written in C#, this full open-source software is available for free download at www.flex-radio.com. It’s maintained and updated by a bunch of highly interested and knowledgeable SDR experts. Additionally, the software is designed to run under Windows XP or Vista. A generic version is in the works so customers can run it on a Mac or Linux machine.
The receiver local-oscillator frequency displays are at the top of the video display screen on the PC, showing the receiver and transmitter frequency and the second receiver frequency if installed. The really interesting display is in the middle of the PC video screen. This is the so-called panadapter display, which shows a segment of the band being received. The width of this band is half the sampling speed of the receiver ADC.
With 192 ksamples/s, the displayed bandwidth is 96 kHz. Therefore, users can see all of the signals in that range via the spectrum analyzer. With its point-and-click tuning, users simply move the cursor to a point on the frequency display and click. This tunes the receiver there and sets the transmitter to that frequency, which becomes the center frequency on the display.
Its waterfall display option appears above the frequency spectrum display. Sometimes called a spectrogram, the waterfall display shows frequency on the horizontal (x) axis, time on the vertical (y) axis, and signal amplitude on the (z) axis in color, indicating signal strength. The display moves from top to bottom and looks like a waterfall. It’s great for picking weak signals out of a noisy background.
Options for the FLEX-5000 include a built-in automatic antenna tuner and a transverter that will convert the transceiver to the 2-m and 70-cm bands (144 and 440 MHz). With its 28-MHz IF, the unit can operate on these two popular ham bands, both of which are also used for amateur satellite communications.
AN IMPORTANT HAM SDR EFFORT
The High Performance SDR (HPSDR) project is conducting an aggressive SDR development effort. This volunteer, non-profit organization comprises more than 600 hams who design modularized hardware and software for hams who want to tinker with or learn more about SDR. Once all of the proposed products are available, hams will be able to build a world-class SDR radio for the amateur bands or for general shortwave listening (SWL)—at a cost of less than about $1000.
HPSDR offers the Atlas, a six-slot backplane that accepts other modules. Multiple modules are under development as well. For example, the Janus ADC/DAC board uses the AKM 24-bit sigma-delta ADC with a 192-kbit/s sampling rate and some fast DACs. It also uses an Altera complex programmable logic device (CPLD) and several interfaces to connect to the other modules over the Atlas bus (Fig. 5).
The Ozy interface board features an Altera Cyclone II FPGA with USB and other I/O ports. The Mercury receiver board also uses a Cyclone II and includes Linear Technology’s 2208 130-Msample/s ADC. Designed from scratch, the Penelope 0.5-W transmitter module employs digital upconversion. At least a dozen other boards are being developed, including receiver front ends, filter sets, and other pieces of the SDR puzzle.
HPSDR sells the boards pre-wired, most for less than $200, through Tucson Amateur Packet Radio (TAPR). A non-profit ham group that started out in the late 1970s and early 1980s, TAPR created the first packet node controller for transmitting digital data over the ham bands.
As for software, the HPSDR boards work with the PowerSDR software available free from FlexRadio. Check out the HPSDR Web site at www.hpsdr.org and peruse the group’s work and extensive projects and offerings.
Continue on Page 4
OTHER SDR PRODUCTS
Most SDR products are receivers. The WiNRadio model WR-G303e generalcoverage shortwave receiver covers 9 kHz to 30 MHz (Fig. 6). This traditional dual-conversion superheterodyne offers IFs of 45 MHz and 12 kHz. DSP filters provide selectivity as narrow as 500 Hz for CW, 2500 Hz for voice, and 6 kHz for shortwave broadcast AM. Demodulation for AM, SSB (LSB, USB), CW, and FM is also DSP. The radio connects to the PC via a USB port. A software front panel provides the displays and controls.
Perseus offers an unusual SDR receiver that isn’t superheterodyne or direct-conversion. Instead, it’s a direct digital sampling receiver. After some frequency pre-selection at the antenna with LC bandpass filters, the signal goes to an LNA and then directly to a 14-bit, 80-Msample/s ADC that converts the entire band to digital. DSP implemented in a FPGA then does all of the filtering, demodulation, and tuning. It connects to the PC by a USB port. A software front panel implements the displays.
RF Space makes two SDR receivers. The SDR-IQ covers the 100-kHz to 30-MHz range. It features a 190-kHz bandwidth and a wide range of software demodulation methods. The SDR-14 has similar features but can also digitize spectrum from another receiver up to 130 MHz. Both of these receivers feature full-blown frequency-spectrum and waterfall display capability. They also use RF Space’s SpectraVue software, which runs on the required PC.
Designed by Tony Parks (KB9YIG) and Bill Tracey (KD5TFD), the SoftRock SDR receiver can be found on a small printed-circuit board (PCB) with a USB connector that plugs into a PC. It produces I and Q outputs that must go to a PC sound card for analog-to-digital and digital-to-analog conversion.
Similarly, Phil Covington (N8VB) developed the Quick Silver SDR receiver. This direct digital conversion device digitizes the antenna input with a Linear Technology LTC2208 ADC. The overall frequency coverage is 15 kHz to 55 kHz. An Altera Cyclone III FPGA handles the DSP and related functions. The GNU Radio project and its software also are worthy of investigation. Check their Web sites for details.
If you want to learn SDR, the ham radio community is a great place to start, whether or not you’re a ham. The available products make it possible to get up to speed quickly without the hundreds of hours typically needed to design hardware, write code, and run endless tests.