80-Bus News

  

September-October 1983, Volume 2, Issue 5











Page 55 of 67











55

All About RFI

Still, perhaps a more thorough investigation into RFI is called for. For the uninitiate, RFI stands for Radio Frequency Interference, and if you are still wondering what that is I suggest you stand your average portable radio next to the computer whilst it’s running. Now lets have a quick look at where this muck is coming from. Its origin is the system clock, or clocks, but it isn’t the clock which is necessarily causing the trouble. The problem lies in the fact that virtually all the computer uses square waves throughout its logic. In fact most of its logic would refuse to work if fed with anything other than square waves.

It is a law of nature that square waves contain odd harmonics of the fundamental frequency, the better the shape of the square wave the more harmonics there are. In practice the harmonics from the high frequency clock is not likely to contain much in the way of harmonic content due to the relatively poor wave shape at those frequencies, but it is successively divided and shoved through gates to perform the various functions of the computer. All this causes the fundamental frequencies to be reduced and consequent improvement in the wave shapes. The net effect is that there are square waves busily clocking things around the computer at all sorts of fundamental frequencies and at the same time generating many harmonics of the various frequencies involved. All this causes broad band noise with a spectrum from a few thousand Hz to several hundred MHz. All that is required now is for something to radiate this switchng noise and we have considerable RFI.

Of course the pe tracks and wires connected to the computer all act as aerials, and worse, because of the broad band nature of the noise, it is inevitable that some of the tracks will be tuned lengths at the higher frequencies, improving the radiating efficiency of the tracks and leads at those frequencies. Any radio placed in the vicinity of the computer will pick up the noise and if the computer is doing something repetitive, then the sound heard will be the familiar chirps and whistles associated with a busy computer. Even if the computer is doing nothing, it is still performing its own housekeeping, keeping the video and dynamic RAM refreshed, scanning the keyboard for something to happen, etc. The keyboard scanning rate of a Nascom 2 is about 1.5KHz, and guess the frequency of the predominant tone heard when listening to a Nascom waiting for a keypress. This is probably the strongest signal heard, as the radiating element is the keyboard cable which is of substantial length.

And so on to the cure. Well I don’t claim it to be a complete cure, as it has not totally eliminated RFI, what it has done is to reduce RFI to sensible proportions so that it is no longer a factor contributing to the difficulty of working dx on 2 metres. It has also removed the faint background hiss which I had attributed to poor aerials on stereo VHF reception and an AM portable radio can now be used in the same room as the computer with no interference on the stronger signals. As I said, I bunged it all in a box. -Well, that’s not entirely true, more precisely, I had a box made. I thought about 19" racks, but by the time you have acquired a pretty Vero case to go round it, it works out prohibitively expensive. No my box measures about 15" by 9" by 8" and is made out of 16g aluminium sheet. It was made with a base plate with a half inch flange all round, the longer two side plates were made flanged on the ends, and two blank end plates. The metal chopping and folding was done to my rough sketches by H. L. Smith of Edgware Road, W2 who specialise in metal bashing.


This is an OCR’d version of the scanned page and likely contains recognition errors.











Page 55 of 67