Among amateur radio enthusiasts (“hams”) the concept of software-defined radios, or SDRs,  has really caught fire over the past several years.  An SDR consists of hardware to digitize a radio signal (or many signals at once) and software to separate the signals from each other and demodulate one or more of them.  Most of the SDRs developed by hams use a PC’s sound card to do the digitization, with the processing software running on the PC.  The most popular examples of SDR software include Rocky, PowerSDR, and Winrad. 

The soundcard is an audio-in, audio-out device which samples the input signal at a maximum rate of 48,000 samples per second (low-end soundcards), 96,000 samples per second (becoming pretty common), or 192,000 samples per second.  The Nyquist sampling theorem says that the input signal must not contain significant energy above one-half of the sampling rate.  A 48,000 sample per second (or 48 KSPS) soundcard can theoretically process input signals at frequencies between 0 Hz and 24 kHz.  A 96 KSPS soundcard’s input bandwidth is 0 – 48 kHz.  This is well below the frequencies of any radio signals, with the possible exception of the very-low-frequency systems for communication with submerged submarines.

A soundcard which has two input channels (stereo input) can actually digitize an input band equal to its sample rate, by using both channels.  This means a 96 KSPS soundcard can digitize signals from 0 to 96 kHz.  That still isn’t enough to directly digitize most radio signals.

Most ham SDRs get around this limitation by using external hardware to “downconvert” a slice of the RF (radio-frequency) spectrum equal in bandwidth to the signal bandwidth of the soundcard.  The first SDR I built was a surface-mount version of a very popular design called the “Softrock 40″.  It converted a slice of the spectrum centered on 14.060 spectrum to the soundcard’s input band, so with a 96 KSPS soundcard it would demodulate everything from 14.012 MHz to 14.108 MHz.  Hams will recognize this as most of the 20-meter CW band, including the very popular PSK-31 activity at 14.070 MHz.  This SDR was very simple, and inexpensive to build.  You can see a photo of the original Softrock 40 at http://www.amqrp.org/kits/softrock40/

The Softrock 40 used a crystal local oscillator, meaning that whatever crystal you used set the center frequency (14.060 MHz in my case) more or less permanently.  There was also a single, fixed bandpass filter between the antenna connector and the downconverter, so the center frequency had to be well within the filter’s passband.  The receiver performed very, very well (at least as well as my FT-870), but was not “frequency agile” outside of a narrow band. 

I decided to develop a new SDR, similar to the Softrock but using a pair of Analog Devices Direct Digital Synthesizer (DDS) chips instead of the crystal local oscillator (LO).  The DDS LO (which sets the center frequency) can be tuned to any frequency between 100 kHz or so and 38 MHz.  The LO frequency is set by a microcontroller which connects to a PC host via a USB cable, and the PC host software is used to select the LO frequency. 

At the same time, I replaced the single bandpass filter with a bank of five bandpass filters, also selected by the host software.  When the user sets the LO frequency, the appropriate filter is chosen.  The end result is an inexpensive but high-performance receiver which covers 2.5 MHz to 30+ MHz – the entire HF spectrum.

I will soon be offering LD-1 SDRs for an initial price of $150 (plus shipping), completely assembled and tested.  For those who wish to build their own, I will be offering kits, semi-kits, and bare PCBs at reduced prices.  The LD-1 is a very nice little receiver for what I think is an attractive price, and would make a good project for students.

I also plan to offer an add-on transmitter module in the near future, to turn the LD-1 into a low-cost transciever.

If you are interested in buying an LD-1, please send me a comment.

Share on Facebook