Softwave- A Very Early SDR

VHF Front Panel

Ahead of its time, a DSP-based HF/VHF receiver that ran on Microsoft windows. With its “front panel” displayed on a user’s computer screen, this was truly a software-defined-radio (SDR).

When Softwave appeared in the marketplace in 1994, it was one of the very first software-defined radios. Technically speaking, it was not a true SDR because it had an RF/analog front-end. It utilized a 455 kHz IF which was bandpass sampled and subsequently processed using an Analog Devices 2105 DSP. The receiver had no hardware front panel; rather, all of the controls were hosted on a personal computer screen.

Softwave provided the user with a number of different radio personalities that had different capabilities and limitations.

AM Radio Personality

The AM radio mode offered both nonsynchronous as well synchronous reception modes. The bass fidelity was exceptional for an AM radio at that time.

Full-Featured Shortwave Radio Personality

The communications radio was the most sophisticated personality in Softwave. It had all of the features offered in a high-end amateur radio of the era, but also had a real-time spectrum analyzer built in to the display. This radio was far ahead of its time in this regard alone. Traditional amplitude as well as synchronous AM detection were provided as well as LSB and USB modes. For FM, a cross-product mode was provided (mathematically the same as a noncoherent frequency discriminator) as well as a PLL-based demodulator.

The Comm & VHF radio modes both had extensive database capabilities for logging station information. In the case of the Comm mode, the database was tied to a graphical representation of the entire world with accurate lattitude and longitude capabilities also included. This was to be used in a later version of the product for minimum useable frequency / maximum useable frequency (LUF / MUF) propagation calculations.

Perhaps one of the most impressive aspects of this receiver was its ability to handle Morse Code (CW). A wide range of bandwidths from about 10 kHz down to a about 50 Hz were offered (48 different bandwidths). The CW-TT mode was ahead of its time also in that the signal processing used optimal methods to determine whether a CW carrier was present or not, and the receiver would output the pure tone of a gated digital oscillator whenever a CW carrier was present. This mode was referred to as “tone tagging.” This mode was exceptionally sensitive, and it gave the listener the feel of noise-free reception (no feeling like you were in a long tunnel at the very low bandwidth settings). This mode was a necessary predecessor for fully automatic Morse Code decoding.

This personality included two different types of squelch / impulse noise rejection that worked extremely well.

The user could also engage an adjustable notch filter that appeared as an overlay in the spectrum analyzer display. Notch width, depth, and center frequency were all adjustable.

Automatic Gain Control (AGC) was fully configurable by the user with adjustable attack, hold, and delay time constants.

As true in several of the personality modes, a single mouse click would frequency-calibrate the receive frequency using WWV or any other known frequency (e.g., CHU-Canada) to within one Hertz or so for unprecedented frequency accuracy.


The VHF radio was purposely tailored to VHF reception. Completely arbitrary frequency scanning groups could be set up by the user as desired. The frequency synthesizer within Softwave was a fractional-N synthesizer, making it exceptionally fast compared to other HF/VHF radios of the day. As such, it could simultaneously monitor many more channels than any other comparable products available at that time.

One of the most interesting modes in Softwave was its built-in real-time spectrum analyzer capability. The high-speed synthesizer combined with the digital signal processing made it possible to monitor large portions of any frequency band for activity. If a signal of interest were to pop up on the display, a simple mouse-click on the display would take the user directly to the VHF or Comm radio personality at precisely the right frequency.

A number of other modes were also available in the receiver as described in the user’s manual reference that is listed at the bottom of this page. Many of the features were novel for the day as alluded to earlier, but never patented.

Technical Details

The Softwave product consisted of five primary ingredients:

  1. A digital DSP board that plugged into the PC’s motherboard
  2. An external “black box” radio unit that had an RF input connector on one end, and a sole parallel cable port on the output end that connected to the DSP board.
  3. DSP code image that was downloaded to the DSP board from the PC making use of the dual-port RAM designed into the DSP board.
  4. A DLL written in C++ that provided the software interface between the DSP/radio and the PC.
  5. A user-level interface program written in Visual Basic that ran under windows.

The RF radio portion (item 2) was an HF/VHF receiver. For HF operation, it used a bank of half-band third-order lowpass filters in order to deliver outstanding mixer spurious performance. In VHF mode, a bank of three filters were used to cover from 108 MHz to 175 MHz. The first IF was 45 MHz in both cases whereas the second IF was 455 kHz. A crystal filter was used for the first IF filter whereas a ceramic filter was used for the second IF filter. A fractional-N (third-order MASH with 8-bit fraction) synthesizer was used for the main local oscillator. Smaller frequency steps were implemented using complex I,Q signal rotation in the baseband digital signal processing.

The analog signal output from the receiver was at 455 kHz. This signal was subsequently bandpass sampled using a single A-to-D converter that was resident on the DSP board. The sampling rate (44.39 kHz) was fixed in the DSP code and Hilbert transform techniques were used to derive precision in-phase and quadrature-phase signal pairs. A high-level block diagram of the digital signal processing is provided here and a block diagram of the DSP hardware is provided here.

All of the C++, Visual Basic, and DSP code is provided in a single ZIP file here. Without the Softwave hardware and suitable (old) software platforms, this code is of limited value, but is provided for the most curious readers nonetheless.

x

x

x

x

x

x

x

The external radio was connected with a standard parallel cable to the Softwave DSP board resident in the user's PC. Power was provided to the radio over the cable as well as all other signaling necessary to support its operation. Note carefully the plastic RF shield connector on the front panel. The metalic box served as a Faraday shield for the receiver, making the plastic bushing necessary rather than a metalic one.

DSP Board, Hosting Analog Interface to Radio Including ADC, Dual-Port Memory Interface to PC, and Analog Devices AD2105 DSP.

x

x

x

x

x

x

x

xxxxxxxxxxxx

x

x

x

x

Close-up of the HF filter bank of half-band third-order LC filters. These were necessary to deliver excellent mixer spurious performance over the multiple octaves represented by the HF band.

xxxxxxxxxxxxxxxxxxxxx

x

Close-up of the tuned-IF stages at 45 MHz. The MSI-packages in the center hosted the digital word control necessary to adjust the receiver's AGC settings.


x

x

x

x

x

x

x

x

x

x

x

Power supply anti-hashing filters at the connector input of the RF receiver. Great care was taken to insure that the (terrible) noise from the PC's voltage supply did not degrade the performance of the radio.

The power supply from the PC to the radio was of special concern since the PC was full of low frequency contaminations. The CRT-based monitors, disk-drive and other hardware meant that there would be unwanted frequency content from near DC to well into the VHF range present. The DC power was filtered with heavy “hash coils” as soon as it arrived on the PWB, along with additional LC filtering to quell higher-frequency contaminants. It was then passed through an active voltage regulator that finished the task. Signal grounding was carefully thought through so that the power supply contaminations did not get on to the exterior of the chassis or cable. Although difficult to see in the photo, there are large hash inductors for both the +12V supply as well as the ground return.x

x

x

x

Close-up of the RF connector end of the Softwave Receiver. The nearest shield covers the high-power first-mix area. The large can in the center rear is the fractional-N frequency synthesizer. The short-shield in the center housed the 10.1 MHz TCXO.

A better view of the entire external RF radio for Softwave is shown here. The small blue cable brought the local oscillator from the fractional-N synthesizer (top) down to the first-mix area. The bank of HF and VHF filters is shown lower-center. Notice that the DB-connector that included DC power is positioned almost as far away as possible from the RF connector.

The bushing of the RF connector was purposely made of plastic rather than of metal. In order to perfect the shielding methodology as much as possible, the radio chassis was used as a Faraday shield thereby avoiding any direct connection between the metal & shielding involved with the DB-connector and the sensitive LNA input of the radio.

Other Files:

Softwave sales slicky (4 MB)

Softwave User’s Manual

Additional Planned Modes for Softwave

Militarized Softwave Radio Concept

Discussion of Patentable Softwave Radio Concepts- 1994

Zip-File Containing All C++, Visual Basic, and DSP Source Code

Mathcad Work

AM Demodulator Algorithm

Audio Bass & Treble Algorithms

Notch Filter Coefficients

Hilbert I,Q Sample Extraction

Hilbert I,Q Sample Extraction after Mitchell

Softwave Morse Code Algorithm Development

More Modes for Softwave

Chebyshev Bandpass Filter Design

Rate Decimation (for I & Q Channels)

Softwave Channel Filter 2

Softwave Channel Filter 3

Softwave Channel Filter 4

Softwave Channel Filters

Radio Interface Control Registers

Although Softwave was a technical success, it fell prey to insufficient market traction due to a variety of issues. The product would have been better aimed at the military communications market, and although proposals of this nature were made through U.S. Air Force SBIR offerings, it was too little too late because of the long procurement cycles involved with government funding. One graphic from one of the proposals is shown below where the Softwave concept was extended to include cellular and SATCOM UHF bands.

Comments are closed.