CODE30 Professional DSP Data Analyzer and Decoder


DSP of CODE30

Please click on the picture to maximize


But how does the DSP part of the CODE30 software work?

After the initial A to D conversion, the centre frequency of the system being monitored (which maybe anywhere within the normal SSB or CW passband of your receiver), is shifted down to DC. The signal will now consist of two, in-phase and quadrature, components.

This complex signal is then passed through a software 'roofing filter'. Because the filter is performed by the software rather than hardware, the specifications are far better than any analog filter could ever be. How about a transition from Passband to Stopband of less than 200 Hz!

The bandwidth of this filter is automatically calculated and optimized dependant on the set shift and baud speed. Change the baud speed or the shift by even small increments and the roofing filter will always compensate automatically for the best bandwidth! The FIR filter is configured using two possible sources of information. One can either manually select the shift and baud speed settings or better still, use the measured parameters of the signal. After this 'roofing filter', the signal is then ready to be demodulated by one of many demodulation methods, some of which are user selectable as follows.

Frequency discriminator demodulators must always have a low pass filter to remove any high frequency keying components. Therefore a special and optimized low pass filter is constructed in software which, just like the roofing filter, is dependant on the set baud speed and shift. Because the roofing filter and the post detection low pass filter are both set up from parameters derived by measuring the signal being received, this frequency discriminator will give a much better performance than the frequency discriminators you may be used to working with up to now. Another benefit of software filtering is that this discriminator will accurately demodulate signals with a shift from 2000 Hz down to an amazing 10Hz with almost identical performance throughout its range!

All frequency discriminators are intolerant of some degree of mistuning, CODE30's FSK discriminator allows for mistuning of up to about one third of the shift without any degradation of performance so is very useful for when one is simply 'tuning around'.

The Frequency Exchange Keyed (FEK) Demodulator uses matched filters. Using matched filters requires a little more care. With most hardware units, the filter widths and shifts are preset, and any signal decoded which does not match these pre-set values will not decode at the hardware's maximum performance. With CODE30's matched filters they are infinitely variable! After using the software tools within CODE30 to measure the monitored system's shift and baud speed, CODE30 automatically sets up optimized matched filters to those precise measured values. No two systems have identical shifts and baud rates but CODE30 will always be able to ensure that it will be decoded using matched filters that give you the maximum performance.

A simple 2 tone RTTY signal only requires two matched filters and so it is easy to construct dedicated hardware to decode them very well. But what about 4, 6, 8 and 13 tone (or more) systems? No hardware decoder can compete with CODE30 in this respect because on a 386DX it is easy to implement 13 matched filters for a 13 tone system! Can you imagine the complexity of constructing a 13 tone demodulator? No wonder CODE30 is the only commercially available general purpose system able to demodulate these more complex systems. Other demodulators include OOK (On Off Keying) for Morse code, 2DPSK and 4DPSK.

But decoding is not always about taking one receiver's output signal and decoding it. The CODE30 A-D card comes with 2 identical A to D channels. Therefore, although not yet developed, with two audio inputs, HOKA Electronics will be able to offer diversity reception!

But that is not all - there is even a THIRD port on the A to D card! This time it is an output from a D to A which means that one can generate signals. Yes, generate a Piccolo Mk6, a Coquelet or ARQ-E or TDM or ARQ6-90 etc. One is able to select from a growing number of signals which are being written and added to this module all the time. At the moment one can select from approximately 10 different systems AND add in a precise amount of Gaussian 'white' Noise too! Select a certain level of S/N and use the signal as a test signal source for testing other decoders, or for training purposes even!

All that is required is a low pass anti-aliasing filter attached to the output connector and you have perfectly generated 'real' signals! Digitally record the output and play it back through the CODE30 input for testing the performance of our software even!

[Please click on the picture to maximize]
0 - 3800 Hz Audio Spectrum Analyzer showing VFT signal (J7B)


[Please click on the picture to maximize]
Zoom 2 - showing just one of the channels of the VFT signal
Note the total lack of adjacent signals due to the filtering !!!


[Please click on the picture to maximize]
Waterfall in Skew mode 0 - 3800 Hz


[Please click on the picture to maximize]
Phase display module showing PSK on 198.0 kHz Radio 4



[Main Page] [previous page] [next page]

[Home Page] [Send E-Mail]


(c) 1989 - 1999 HOKA Electronics (NL)