NASWA Journal Columns · Equipment Reviews, May 2008

Alan Johnson, N4LUS • 2490 Sharon Way • Reno, NV 89509 alanjohnson◊

Equipment Reviews, May 2008

Microtelecom Perseus Software Defined Receiver

A sea change has occurred with receiver technology. Over the past three decades, digital technology has gradually supplemented and then replaced analog circuitry in modern radios. At first, microprocessors were put into radios for tuning and function switching and external computers were used to control receivers. The next transition was the replacement of hardware by software (digital signal processing). Now, we have reached the point where the computer (with the addition of a couple of outboard intergrated circuits) is the radio. Not only has the internal hardware of the receiver been eliminated, but the control hardware of knobs and switches has been replaced by a computer mouse.

A software defined radio (SDR) is just that – a radio in which the desired radio frequency signal is digitized and then all further receiver functions such as demodulation, filtering and noise reduction are handled by software. This eliminates the need for mixers, transformers and filters and provides almost unlimited flexibility. Digital Signal Processing (DSP) has been in use for quite some time now, but generally has been used with a traditional superheterodyne front end, with the attendant problems of overload and spurious signals from the mixer circuits which are the basis of the design. In a SDR, the desired RF signal is directly sampled and then digitized for demodulation and filtering. SDR’s can have phenomenal close-in strong signal handling performance as well as having the capabilities of DSP for excellent selectivity and noise and interference reduction. An additional advantage of SDR’s is that the receiver’s performance and feature set can be continuously upgraded through software updates.

The basis of the SDR’s magic is the derivation of an in-phase and a 90 degree out-of-phase samples (I and Q respectively) of the incoming RF signal. There are several techniques to do this, such as the Quadrature Sampling Detector or by direct digital downconversion. By subsequent mathematical manipulation of the I and Q signals in a microprocessor, any type of demodulation can be accomplished. SDR’s represent a completely new receiver topology. For more technical information on SDR’s, go to and click on “FAQs” from the drop-down “Products” menu as well as “Publications” from the “News” tab. The QEX series of articles, “A Software Defined Radio for the Masses” are particularly good. Also see (scroll down to the link to “Introduction the the QS1R”).

This review is a follow-on to Giampiero Bernardini’s review of the Perseus in the March, 2008 Journal. The production version of the receiver software is now available and it adds several new features to the beta software used for Giampiero’s review. The Perseus receiver is designed and produced by the Italian firm, Microtelecom, led by Nico Palermo. Perseus is an acronym for “Pretty Execellent Receiver for Software-Eager Unperceivable Signals”. The name may be somewhat lame, but the receiver isn’t! It is designed for high performance over the range of 10 kHz to 30 MHz and has an intutive, but flexible, user interface, which incorporates all the features one would expect in a top-notch communications receiver. It can also be used with third-party control software for specialized applications, such as spectrum analysis. One the radio’s most exciting features is the ability to record up to 800 kHz of spectrum for later playback and analysis.

The box

The Perseus is truly a black box measuring 1.5 x 4.25 x6.5 inches (HxWxD). The front panel has small rectangular LED’s to indicate power, wideband mode and 10 and 20 dB attenuators (both light for the 30 dB attenuator selection). There is a red LED to indicate overload. The back panel has a USB 2.0 port, a coaxial jack for power (5 volts at 1 amp) and a BNC jack for RF input. The end caps of the enclosure are set into yellow gaskets which provide a nice visual contrast and more importantly keep the unit from sliding. The unit comes with a switching wall-wart power supply. Mine had a round pin European power plug, so I had to get an adapter. Unlike some switching type power supplies, this one is overall free from RF noise. Inside the box are the main components – a bank of half-octave bandpass filters, a 14 bit Analog to Digital converter (Linear Technology LTC2206-14) and the FPGA Digital Down Converter utilizing a Xilinx Spartan IIIE XC3S250E.

The other vital component of the Perseus system is a computer. The current computer system recommendations to take full advantage of the receiver’s capabilities are for at least a Dual Core Pentium with a 2.5 Ghz clock speed and 500 MB of RAM, running Windows XP or Vista. A USB 2.0 port, a standard 16 bit sound card, mouse with center wheel and a minimum graphics resolution of 1024 x 768 are also required. Mine is running fine with a 2.33 GHz Core2Duo with 2 GB of RAM running the XP Pro operating system. If you do plan to take advantage of the recording feature, a large hard drive is a necessity, since 10 minutes of recording with an 800 kHz span width requires 3.6 GB of storage. A sound card is required for audio playback, but its quality is not important, as (unlike some Winradios) no signal processing is done by the sound card.

Installation consists of plugging in the power supply and connecting it to the radio, connecting the radio to a USB 2.0 port on the computer with the supplied cable and then installing the drivers and control software. There is no instruction manual in English, but the software disk that came with my receiver had a Quick Start PDF file in English and there is comprehensive information on installation and operation in English on the web. The current software version is 1.0f.

The interface

This is, of course, what is most exciting about SDR’s, since the receiver itself is connect and forget. The Perseus and the Flex5000A software were my first exposure to using a real-time spectrum display and mouse tuning. I had used the static spectrum displays on the TenTec RX-350 and ‘320, but not much. With the Persus, the spectrum display is real-time and tuning is essentially instantaneous. A waterfall display is also an option. In one available tuning mode, one uses the mouse to point to a desired signal on the display, double left-click and the receiver tunes to that frequency (reminds me of shooting fish in a barrel!). The Perseus control interface is designed to resemble a standard communications receiver front panel with pushbuttons and sliders and will be intuitive to use for most hobbyists. Refer to and click on the lower screen shot for a reference. The main portion of the panel is the spectrum display which can be set for span widths of 1.6/3.1/6.3/12.5/25/50/100/200/400/800 kHz. I’m tending to use the receiver as a conventional receiver (i.e. not recording spectrum for later analysis), so I find myself using the 200 and 400 kHz spans, which yield a less crowded display. There is also a secondary display window which shows the signal that the receiver is tuned to in more detail. The span width of the secondary window is determined by the filter bandwidth selected. The secondary display gives a visual representation of the signal’s modulation, the filter passband and the notch filter.

All aspects of the receiver are controlled via the mouse, which needs left and right buttons and a scroll wheel. There are no keyboard commands or shortcuts, with the exceptions that the notch function can be accessed by holding the control key instead of clicking the Notch button beneath the secondary display and frequencies can be directly inputted via the keypad. Many people are resistant to adopting computer controlled radios because there are no knobs – well, after a few minutes of adapting to the Perseus interface, I can say that I’m not looking back. I use a trackball, so I can access all receiver functions by moving my thumb and two fingers.


There are a myriad of tuning options available, in addition to the point and double click method. For band-scanning the mouse wheel can be scrolled in steps which are mode dependent (5 kHz for AM and 100 Hz for SSB are the defaults) but can be changed via the Up/Down arrows next to the Wheel Step window. The pointer must be over the main display or the frequency ruler at the bottom of the main display for scroll tuning. The frequency ruler can be shifted by pointing to it, holding a left-click and dragging the ruler. Beneath either end of the ruler are left and right “Click Frequency” arrows that can be clicked to tune in user selectable steps (default is 100 kHz with a range of various steps from 1 kHz to 1 MHz). The tuned frequency is displayed in the Frequency Pane with a resolution to one hertz. The receiver can also be tuned by pointing to any of the digits in the Frequency Pane and rolling the scroll wheel. I find scrolling the MHz digit to be the fastest way to jump between broadcasting bands. By double-clicking in the frequency pane, a direct frequency entry window pops up for direct frequency input via the keyboard. This is one thing I would like to see changed in the software – my preference would be to directly enter a frequency via the keypad without having to open a separate window. One last tuning method is to grab and drag the dial pointer in the main display, but this is only available if the “Center” button is deactivated in the “Tuning” window.

The secondary display

This window (also known as the demodulation window) is the control center for filter selection and manipulation. The signal (both carrier and modulated sidebands) is displayed in the window and the filter passband is represented by a gray column within the window. If a wide enough filter is selected, adjacent channel signals are displayed as well. There are six filter selection buttons next to the secondary display window to choose different bandwidths ranging from 0.8 to 25 kHz – these are just starting points, as each filter can be dynamically varied by scrolling the mouse wheel while pointing to the secondary display window. The bandwidths are not infinitely variable, but the bandwidth steps are dependent on the selected bandwidth button, e.g. if the 12 kHz button has been clicked, the scroll wheel varies the bandwidth in approximately 50 Hz steps, for the 6 kHz button the steps are about 25 Hz and for the 3 kHz button the steps are 12 Hz. The -6 dB width of the chosen filter is shown numerically in the upper left corner of the demodulation window. The filter shape factors are excellent, with a quoted specification of 1:1.08 for the 2.4 kHz filter at the -6/-60 dB response points. The minimum filter width available is 57.2 Hz.

In addition to changing the bandwidth, the secondary display allows for several other functions. By double-clicking on the signal peak, the signal will be centered in the passband. A left drag will allow shifting the signal peak within the passband for fine tuning. By pointing to either side of the passband column and left dragging, the edges of the passband can be individually shifted. By right dragging, the entire passband can be shifted to effect passband tuning. If the “Notch” button is clicked (or the Ctrl key held down) the Notch filter is activated and the notch can be tuned to a desired frequency by a double click or variably tuned by a left drag. The width of the notch can be varied with the scroll wheel and it is deactivated with a right click. All of these adjustments are graphically displayed in real-time, providing a visual indicator of filter adjustment in addition to the aural effects. This is another aspect of SDR’s that I find fascinating.

The other buttons

For a tour of the other features of the Perseus, I’ll start in the upper left hand corner and work my way around the front panel. There are four stages of attenuation: off, 10, 20 and 30 dB. Below those buttons are buttons to select the preselector (1.7 MHz low pass filter and half-octave bandpass filters from 1.7 to 30 MHz), a pre-amp that yields about 4 dB of gain and Dither, a software function to reduce spurious signals. Next are the Amplitude controls to set the reference level and height of the waveform of the main (in Spectral mode) and secondary display. The level of the waveform can also be set with the Up/Down arrows at the right side of the main display. Continuing down the left side of the front panel are the Frequency pane and windows and control buttons for Span Width, Click and Wheel Step settings. Below this is the Tuning window with buttons for Center, Cal and CalClear. Clicking on the Center button puts the tuned frequency in the center of the main display. The Cal and CalClr buttons permit accurately calibrating the receiver to a frequency standard, such as WWV. This can be done on the fly, without having to open any configuration files. The calibration procedure is outlined on Guy Atkins’ blog (see below). Next are the Sampling Rate selection buttons. Increasing the sample rate increases the frequency span covered by the receiver. Unlike the Flex-5000A, changing the sample rate doesn’t seem to affect the filter performance of the Perseus. Below this are the input select buttons to choose between the receiver and recorded WAV files.

Along the bottom of the front panel are the controls for recording and playback of Wav files – again the Perseus can record up to 800 kHz of spectrum for a time limited only by the size of the computer’s hard drive. To the right of the secondary display window are the sliders for adjusting the response time of the main and secondary displays. Next to the right is the noise blanker control – the threshold level can be adjusted with a slider control. Then there is a column of seven pushbuttons. The top three are for the Marker function, which permits tacking the frequency and amplitude of up to four signals in the main display – the markers are invoked by a right click on the desired signal in the main display. The bottom four buttons are for the AGC – off, slow, medium and fast. The AGC parameters are not user adjustable. The lower right half of the front panel contains a large horizontal bar format S-meter which shows both S units and signal level in dBm. It can be set to respond to either average or peak signal level. Below the S-meter is a window that displays station information for the tuned frequency based on either the HFCC, EiBi or a user-defined database. The databases are in files stored with the program, not downloaded from the Internet, so they will need to be updated over time. To the far right of the lower half of the panel are the buttons and sliders for the noise reduction and volume controls. Between the main display window and the bottom half of the panel are the mode buttons – AM, Synchronous AM, CW, RTTY, USB, LSB, FM, DRM and USER (which feeds zero-IF IQ samples to the Virtual Audio Cable output for use with third-party applications). Synchronous detection is double sideband, but single sideband sync detection can be obtained by using passband tuning, as with the AOR AR-7030 or the early Drake R8. DRM detection requires the use of two separate external programs – Dream for decoding and Virtual Audio Cable to route the digital audio and avoid the need for two sound cards. The DRM Software Radio software is not supported. In the far bottom right is the status indicator for Vcom status. VCom stands for Virtual Com port – this feature allows the Perseus to be controlled by an external program using the CI-V instruction protocol. Finally, along the right-hand side of the main display, are the Up/Down arrows to adjust the reference level of the spectrum display and sliders to adjust the speed, brightness and contrast of the waterfall display.

The Specs

The Perseus’ specifications (at 14.15 MHz, as published by the manufacturer) are as impressive as its functions and controls. The frequency coverage is 10 kHz to 30 MHz. Selectivity stopband attenuation is > 100 dB and image rejection is 90 dB. The third order intercept is >31 dBm and dynamic range is 100 dB for SSB /2.4 kHz filter and 104 dB for CW/500 Hz filter (signal spacing not specified). The Minimum Detectable Signal level is -131 dBm in CW/500 Hz filter and -124 dBm in SSB/ 2.4 kHz filter, both with preamp on. SSB sensitivity is quoted at 0.39 uV for 10 dB S+N/N (pre-amp off/preselector on). These are very respectable specifications – I look forward to seeing confirmation of the test results from either Sherwood Engineering or the ARRL.


I found with the Flex-5000A that what seemed to matter most for successful operation was not the computer specifications, but how many processes were running in the background. The Perseus seems to be much more tolerant and runs without glitches or hiccups. The average CPU usage on my system is 10 per cent.

Based on my experience with the Perseus at the Winter SWLfest, the receiver needs a good antenna to perform properly. At the ‘Fest, I tried the Perseus with an AOR Window loop and with a 20 feet length of wire – with either antenna, the Perseus was out-performed by an Eton E1 with it’s internal whip. This relative lack of sensitivity is not a problem at my home with either full-size wire antennas or outdoor active antennas. It is possible that some Perseus users with marginal antennas might benefit from a low-noise preamp ahead of the receiver, since the built-in preamp only provides about 4 dB of gain. I’m planning on trying a Kiwa shortwave pre-amp to see what difference that makes.

In operation, the Perseus is a dream. My quick tests for selectivity and resistance to overload are trying to listen to the BBC on 6005 kHz with R. Habana on 6000 kHz, CKZU Vancouver 6160 kHz vs. R. Nederland on 6165 and R. Educacion, Mexico on 6185 kHz vs. R. Habana on 6180. In all cases (propagation premitting) I was able to obtain a listenable signal on the weaker stations. Splatter from the stronger stations was not completely eliminated, but significantly reduced allowing intelligible listening. The filter shape factors are super sharp and the available bandwidth steps are more than enough to find the right bandwidth for the received signal. I really enjoy being able to visualize the filter width vs. the signal’s modulation in the secondary display. The notch filter is very sharp and deep. The notch width is variable and the effect of the notch is visible in the secondary display. The front-end seems to be very overload-resistant, with no breakthrough of mediumwave signals into the lower shortwave frequencies.

The audio output is crisp and clear, so a good set of computer speakers is called for. I’ve found the M-Audio Studiopro 3 to be a good choice for a non-subwoofer setup. My favorite speaker system with a subwoofer is the Klipsch Promedia 2.1. The AGC action is smooth and also contributes to the pleasant sound of the receiver. One feature that is missing is a RF gain control or, as found in the FLEX-5000A, an AGC Threshold control, which can be adjusted to reduce background noise to near inaudibility. Hopefully, this will be addressed in future software upgrades. The volume control appears to perform as a RF Gain control when the AGC is set to off. The noise reduction in the receiver is one of the best I’ve heard – it significantly reduces noise on the signal without altering the signal’s intelligibility or introducing artifacts. The synchronous detector is very good, but will very rarely lose lock with deep signal fades. ECSS reception is very good – to the point that synchronous detection is almost redundant. Although I have not made much use of it, the ability to record a broad swath of spectrum for later analysis is quite exciting.

The bottom line

Software defined radios represent a tremendous leap forward, not just in receiver design, but in the user interface as well. I have hardly turned on any of my other radios since getting the Perseus. It would be nice if the developer would publish a demo version of the software for perspective users to experience the interface and features. Hopefully, there will be a US distributor for the Perseus in the near future once FCC approval has been obtained, but ordering one from Europe is not that difficult – it can be ordered from Woodbox Radio in Italy with Paypal – current pricing for the US (March 2008) is 687.50 Euros, which appears to not include shipping. A US-made competitor to consider is the Quicksilver QS1R VERB which should soon be in production – it features open source software and will be able to sample and record up to 2 MHz of spectrum.

Internet resources

The manufacturer’s site:

Guy Atkins’ blog:

Yahoo discussion group:

A guide in English for operation of the Persesus along with recorded audio:

SDRs in general:

Read more Equipment Reviews columns.

Comments are closed.

Quick Information

Solar Conditions

At 2023 Mar 14, 0300 UTC

  • Solar Flux: 143
  • A-index: 3
  • K-index: 3

Full Report from NOAA