Satgen 374 Why Fourier Transforms by GM4IHJ 25 May 1996 The reported use of Fast Fourier Transform software in Satgen373 brought in a lot of questions. Here is an attempt to explain what is involved, with the warning that it is not easy to give simple answers, but, the method is easy to use. " Why go to complex digital processing" ? Because we need very narrow filters, if we are to pick out weak signals from noise. All mechanical, RC , and xtal filters give up below 50 Hz and they "ring" dreadfully at times. FFTDSP42 gives us clean 2Hz filters and gives us 640 of them, so we do not miss the signal by poor tuning. Mathematician Baron Fourier pointed out over 160 years ago that any waveform which can exist can be generated/duplicated by adding up sine waves.Conversely we can go the other way breaking down any waveform into a combination of sine waves, and display this data on a frequency spectrum analyser as AF9Ys software does. We describe the waveform mathematically in what are called Fourier Transforms computing the FT digitally by performing a numerical integration to produce a Discrete Fourier Transform. The Fast Fourier Transform used by AF9Y is an algorithm ( mathematical statement) which computes this Discrete Transform using an arithmetical short cut ( hence " Fast ") which is only an approximation, but a good enough one, for our practical purposes. In effect the computer samples the incoming signal in 2Hz filters, over a 640 x 2Hz = 1280 Hz bandwidth of your receiver audio ( 260 to 1540Hz ). Each sample has discrete amplitude and phase data. This date is converted into vector addition and averaged .Then every 0.5secs the average amplitude is displayed under a frequency scale , with colour varying from blue through green to yellow white with increasing amplitude. The program adds a new sampling line across the screen every half second, if integration is set to 1. This is fine for EME, sub horizon sats and meteor scatter, but integration can be increased to show up very weak signal like radio astronomy examples or the Nov 1996 Mars Surveyor 70cm probe. Be aware however that even the shortest integration interval does not allow you to read Morse or telemetry from the screen unless it is very very slow indeed. But this technique does alert you to the presence of a signal, even when there are lots of stray intruding signals and artifaxes in the frequency band . Indeed this software is superb as a tool for identifying all the electronic signal rubbish getting into your shack. So you can do "Electronic spring cleaning " before going on to look for a difficult target like the Mars probe (IHJs old 286 micro will definitely be off when the hunt for the Mars beacon starts ). While 70 cm tests looking for Fo20 435.795 MHz beacon as the sat goes below the station horizon will be used to get experience prior to the Mars probe launch. Equipment at IHJ is presently a 21 element long yagi on 70cms ( to be changed before the Mars probe ), feeding a very quiet Jamsat mode J converter 435 to 29 MHz into an NRD 535 Rx, with audio out into a Sound Blaster 16 value card in a Pentium 90 MHz micro ( you can get by with a 486 micro, but slower micros will probably not handle the demands on the CPU fast enough). Try it if you can ,and have fun.