John Reyland, PhD
Instructor of ATI’s “Signal Processing for Software Defined Radio”
ATI offers two scheduled courses on Software Defined Radios (SDR). Additional courses are offered on-site at your facility worldwide to groups of 8 or more.
Practical Software Defined Radio Development
Software Defined Radio Development- Practical Applications
The schedule for these SDR courses is provided at the link below. Check frequently as additional courses are added based on the demand. If you are interested in courses for the SDR area, please post a comment on the ATI blog and send an email to firstname.lastname@example.org. Visit:
First we should define what is a Software Defined Radio (SDR). The author of the reviewed article neglects to do this so I will provide a definition. In my view an SDR is a radio system with control and reconfiguration defined by a highly organized SDR frameworks. This frame work reads configuration files and reconfigures radio hardware and software to bring about the desired radio operation. An example of a standardized set of SDR configuration files is the Software Communication Architecture (SCA) domain profile. SCA domain profile is a set of Extensible Markup Language (XML) files that describe waveform components and interconnections. One example of an SDR that meets this definition is at: http://redhawksdr.github.io/Documentation/
This is a useful article on examples of Software Defined Radios and the source of inspiration for this tutorial blog post.
None of the radios in the reviewed article meet this strict definition. So here is a simpler definition that will cover most of them. An SDR is a hardware device that tunes to a range of radio frequency carriers. The selected carrier is down converted to an intermediate frequency, filtered and converted to a stream of digital samples. Ironically, this definition does not imply any user changeable software at all! Indeed, there is very little user changeable software in most of these 12 devices. The author neglects to mention that the “software defined” part is written in GNU radio, Simulink or some pre-written third party software (for example, see http://airspy.com ) running on the computer the SDR is connected to. These products are more properly called SDR front end tuners. To be fair, I will admit that common usage seems to prefer calling them SDRs.
A disappointment with these products is that most of them do not have a sensitivity spec. This would be, for a digital signal, the minimum bit error rate for a certain low receive level (for example -100 dBm on the antenna input). For a voice signal sensitivity could be the signal to noise and distortion (SINAD) for a similar low level signal. In my opinion, the lack of serious specs like sensitivity puts these products in the hobby or educational category. A professional application would need a link budget and other parameters like probability of outage. Lack of a sensitivity spec makes these difficult.
Another observation is that most of these SDRs do not tune down into the 3 – 30 MHz HF band, where most of the ham radio signals are. This may be because the tuner chips they use were designed for applications, such as TV and commercial wireless, that did not need these low frequencies. There are a few that offer adapters to reach these low frequencies.
This article provides sparse comments about 12 SDR products. Prices, applications, frequencies and ease of use are given as the reason these 12 were chosen. In what follows I am going to reorder the author’s choices, starting with the products that seem to offer the most usefulness to the average SDR experimenter.
This is a low cost (about $25) 24 MHz to 1.8GHz tuner with a USB interface. You can put together a complete SDR receiver by controlling the RTL-SDR with pre-written SDR software (for example SDR#). The well written book “The Hobbyist’s Guide to the RTL-SDR” by Carl Laufer is an absolute must for getting the most out of this receiver.
This is made by the same company (NooElec) and is very similar to the RTL-SDR. The RTL-SDR started out as a TV tuner for use in countries other than the US. The NESDR is specifically designed for SDR use. A primary difference is the TCXO (Temperature Controlled Crystal Oscillator) in the NESDR has much better specs then the cheap crystal in the RTL-SDR.
This is a higher performance wide range tuning version of the RTL-SDR. With 1-6GHz tuning range and both transmit and receive capability, this may be the best value of any of these SDRs. Price is only $299.00. However, for those who want to program receiver DSP into an on-board FPGA, the USRP may be a better value. The HackRF is RF only, no FPGA.
With a $199 price tag, this is another upgrade from the low cost RTL-SDR. However, unlike the HackRF, the 24MHz – 1.8 GHz tuning range is not much better than the RTL-SDR. Also, unlike the HackRF, the AirSpy does not seem to include a transmitter.
This is a USB dongle based SDR with better specs than the RTL-SDR. Tuning range is restricted to the 2.4 GHz ISM bands, not as versatile as the HackRF.
National Instruments USRP
This is similar to Ettus USRP. These devices are both based the Analog Devices AD9361 transceiver chip. The Ettus USRP is commonly used with GNU radio and the NI USRP is probably designed specifically for Labview, a block diagram oriented signal processing tool from National Instruments.
Other devices called SDR
The reset of this list has various circuit cards that could possibly be used as SDRs, however they either have high prices or they simply do not include any radio tuning or they simply were not designed to be SDRs.
Surprise, surprise, this is not an SDR even with our simpler definition. There is a 125MHz analog to digital converter (ADC) and a 125 MHz digital to analog converter (DAC). However, there is no antenna matching, AGC, tuning or IF filtering. Seems like the Red Pitaya is more suited to some kind of electrical mechanical control. Now that I have bashed it as an SDR I will point out a very positive feature. The Red Pitaya has a Xilinx ZYNQ FPGA. Some of the other SDRs have a smaller FPGA that, with some effort, the user can implement radio digital signal processing. The Red Pitaya has a much larger FPGA that could possibly perform all the signal processing need to completely decode the received message – if it was a radio.
Except for a much smaller FPGA, Quadrus SDR is similar to the Red Pataya. There are four channels of phase coherent sampling but no front end RF components. Most radio front end circuits provide frequency down converting, filtering and up to about 100 dB of gain to match the receive signal to the ADC dynamic range. Here, we have up to 30 dB gain going directly into the ADC input. Coupled with properly designed radio front end hardware this device could be a good platform for multiple input multiple output (MIMO) signal processing.
This is a 6U Compact-PCI form factor hardware platform that can be used for SDR. This hardware is quite large compared with the other products discussed here. The fact that a card this large only has a Virtex 5 FPGA and no radio hardware tells me it is probably not a good choice for an SDR. The Rad Pitaya will perform better for less money.
With a tuning range up to 27 GHz, this SDR is unique. The pricing starts at $3500, a bit outside the hobby category and more like professional test equipment. Comes with spectrum analysis software.
This has radio front end components however is it is designed for use as a spectrum analyzer and is not an SDR.