Notes on Some Sound Card Radio and Modem Programs

Notes on Some Sound Card Radio and Modem Programs

Brian Pease

7/20/06

I recently had a need for soundcard software to test different digital text modes in audio frequency Thru-the-Earth communications. There are many other programs besides the few mentioned here. I used the software directly on their audio output frequencies, not as audio in/out to an SSB radio.

Alberto, I2PHD, has developed three freeware soundcard programs that I have tried. These are available at his website http://www.weaksignals.com. The first two are receivers.

WINRAD is a Software Defined Radio tailored for weaksignal SSB and CW work. It is a sophisticated audio-frequency receiver with adjustable bandwith, I and Q inputs from a mixer, (or a single input without sideband rejection), multiple spectrum displays, and DSP noise reduction.

SDRADIO is an experimental Software Defined Radio, with AM, ECSS, SSB and FM demodulators, with a highly graphical interface. It also has I and Q inputs. It can tune as high as 24 kHz by itself, with a spectrum analyzer display and point-and-click tuning. Both programs are great fun to play with.

JASON is a slow digital text mode designed for extreme weak signal two-way communications over links with random noise interference and multipath. In its normal mode of 2.5 characters per minute, the total bandwidth is only 4 Hz! There are faster and slower modes for better or poorer conditions. Text is coded as the difference in frequency between two tones, sent sequentially, eliminating the need for precise tuning or for a linear amplifier. Only upper-case characters and common punctuation marks are used. There are 17 tones spaced 0.25 Hz apart., each sent for 11.9 seconds. In receive mode, it searches for any continuous tones within a 6 Hz window. It is vulnerable to continuous tone interference. Once locked on, the frequency of each tone is read out to within .001 Hz! If the receive and transmit frequencies are very close (say within 1 Hz), the performance is remarkable! The 4 Hz BW signal can be centered between power line harmonics.

I recently used this mode successfully during some audio-frequency Thru-the-Earth tests in an active mine with extremely high levels of manmade EMI both above and below ground. Besides the laptop computer, only a power amp, receive preamp (with bandpass filter), and an antenna were needed on each end of the link.

SPECTRUM LAB (SpecLab) is a freeware program by Wolfgang Buescher, DL4YHF. It is available at http://www.qsl.net/dl4yhf. SpecLab is a very sophisticated (and complex) audio signal analyzer program for analyzing real-time or recorded signals. It can produce a real-time audio output like the receiver programs, and in fact can be configured as a VLF receiver up to 24 kHz. It can accept I and Q audio inputs to create an SSB DC receiver. It has a long list of features including a built-in noise and multi-tone signal generator, FFT-based FIR filters that can also shift frequency up or down or change USB to LSB, notch out steady carriers, do DSP noise reduction, etc. It can accept inputs from a pair of crossed loops to do direction finding. It can receive and transmit in several digital modes including slow CW, several PSK, MSK and HELL modes. It even has a block diagram mode that can be used to change the configuration. Fortunately, there is an extensive help file! This is a really fine program if you have lots of time to learn it.

BR-UNIVERSAL FILTER is a small freeware program by Bernd Reiser available at http://www.dsp4swls.de/download/download.html. It is intended for real-time receiver audio filtering up to 4 kHz. It can do IIR or IFR low pass, high pass, band pass, and notch filters, fully adjustable in frequency and order. It can implement comb filters for 50 or 60 Hz hum reduction (or adjustable freq.), DSP noise reduction, frequency shifting, and sideband inverting. It displays the original unfiltered signal next to the filtered output that you actually hear. It is easy to use, but there is only a limited help file in rough English, and some of the on-screen graphics are in German. The various settings, including frequency, are on analog scales only. I created a simple receiver that filtered a 3496 Hz beacon signal with about 20 Hz bandwidth and converted it to a ~400 Hz audio output. There is no AGC. It seemed to work well as a substitute for my DQ radiolocation receiver.

STREAM, is a freeware program by Nino Porcino, IZ8BLY. It offers the MFSK-16 (Multi-Frequency-Shift-Keying) mode, which is a robust typing-speed two-way FSK mode with 16 tones at about 16 baud, FEC, and a bandwidth of about 250 Hz. See http://xoomer.alice.it/aporcino . The software also does PSK-31, which has less than 50 Hz BW. With careful frequency selection for MFSK-16 to avoid landing a tone directly onto a 60 Hz power line harmonic, I found in actual audio-frequency tests close to some large power lines, and later in the lab with simulated 60 Hz EMI, that PSK-31 was better. To use the MFSK technique successfully with power line interference would require custom tone frequencies.

HAMSCOPE, a freeware program by KD5HIO, offers several digital two-way modes including MFSK-16 and PSK-31. See http://www.qsl.net/hamscope/HamScope.html . It is very user-friendly with several unique features including a superb spectrum display with a second averaged trace, and the ability to periodically transmit a short canned message and automatically switch to receive mode to listen for a reply. I ended up using HAMSCOPE’s PSK-31 mode, along with JASON, for the active mine testing, for reasons described next.

Lessons Learned:

If one of these programs is going to be used for weak signal work, especially without a preamp ahead of the sound card, your Windows sound card settings should be checked. There is a 20 dB gain boost available in the microphone mode that may or may not be activated (but should be).
In Windows XP, right-click the speaker icon in the system tray and select "adjust audio properties". Press "Advanced" under the Microphone slider then check "microphone boost" for a +20 dB gain increase.
Another way to do this in W-XP is to double left click the speaker icon to bring up the master playback volume controls. Click options/properties; select the playback radio button; then check "show microphone". Next, in either the playback or recording control panels, press the "Advanced" button under the microphone slider and check "microphone boost".
In Windows 98 and 2000 you have an additional step in options/properties/playback where you must check "advanced controls" to create the Advanced button first! Now when Advanced button is pressed, you check "AGC" to boost gain +20 dB.

One problem is related to the accuracy of the sound card’s sampling rate, which the programs rely on for timing. Many programs allow for compensation of the sample rate, and some help you measure it. In my limited experience, this was a minor problem.

The biggest problem with the programs is frequency accuracy of transmission and reception. For my project I needed transmission and reception to occur on precisely the same frequency to avoid failures due to being on the wrong frequency. The problem seems to be related to the operating system one uses. I had 3 laptops, each with a different operating system! For example, JASON appears to have been written for Windows 98 because in W98 the program transmits and receives on precisely the same frequency, which is also accurate. In Windows 2000 and Windows XP, JASON received on the correct frequency but transmitted with as much as 19 Hz error, well out of the 6 Hz receive window. All of the programs I tried that had transmit capability behaved in a similar way. Like JASON, HAMSCOPE received accurately, but on transmit it had frequency errors with all 3 operating systems sufficient to be out of PSK-31’s AFC range.

The reason I chose to use JASON and HAMSCOPE is that they both allow independent setting of transmit and receive frequencies! I was eventually able to "net" all 3 laptops to transmit and receive on the same frequency, eliminating one variable. I posted a calibration chart on each laptop with the transmit and receive frequencies to enter into each program.

The lesson learned is that when you are going to work with signals that may be buried in noise, and free software, bring all the gear together in one place beforehand and don’t assume anything!