In radio communication parlance, “noise” doesn’t just mean random and meaningless audible noise such as that made by a tube train or a passing aeroplane, but it also includes random and meaningless electrical signals. These may eventually well cause audible noise from a loudspeaker, or the familiar “snow storm effect” on a visual display as is sometimes seen on an analogue TV receiver when tuned in to a rather weak station. Electrical noise may be generated externally to a radio receiver quite naturally by lightening or other sources of atmospherics, or by radio frequency radiation from the sun or from other cosmic sources. It may also be man-made by a neighbour using an unsuppressed electric drill or some other similar cause. Or it may be generated internally in the receiver. Here, it is usually generated by less than perfect components, (which includes most of them), within the receiver. This internal noise is caused by the random movement of electrical charge carriers, (usually electrons), in the conductors of every circuit.
An electric current is the flow of electrons through a conductor, rather like the flow of fluid through a pipe. However, electricity is not a perfect fluid but consists of discrete particles, (electrons), which “bump their way through the conductor” hitting some atoms and avoiding others. Although the average position of the atoms of a solid like copper are fixed at constant distances from each other, they vibrate about their fixed positions with amplitudes which increase with the temperature of the solid conductor. This vibrational energy is transmitted to the electrons, adding further to the randomness of their passage. The net result is that the current through the conductor or other component, (such as a resistor or an amplifier), varies from moment to moment about the average current. This variation accounts for the noise current and the randomness in the current generates a noise voltage, (or potential difference), across the conductor. Even when there is no average current, the thermal vibrations of the atoms jostling the electrons produces a noise voltage.
The “non streamlined flow” of electrons caused by their colliding with the atoms of the conductor, accounts for the electrical resistance of the conductor. Both the resistance and the noise contribution of a conductor can be reduced by lowering its temperature, but it can only be eliminated by lowering the temperature to absolute zero, -273C. Most forms of electrical noise, whether internally or externally generated can be thought of as a mixture of alternating currents of variable amplitudes and all possible frequencies.
Because of the randomness of the noise generating processes, the noise frequencies cover virtually an infinite range, metaphorically speaking, “from DC to light”, (actually even beyond light frequencies). But we don’t have to be bothered by all of them! We can restrict the range of frequencies that we allow through to the loudspeaker or visual display and thus restrict the amount of noise allowed through. After all, if the frequency range we are interested in only extends from 300Hz to 3000Hz, (as is the case for the human voice), there is no point in letting through frequencies from 20Hz to 10kHz which may contain noise in addition to the wanted voice frequencies. A range of frequencies such as this is called “the bandwidth”. If the noise frequencies are evenly distributed throughout the range of frequencies which we are interested in, the noise is known as “White Noise”, by analogy with white light which contains all the colours which the human eye can see in roughly equal measure.
If a wide bandwidth of noise voltages are fed to a loudspeaker, the sound generated is that in “sssssss”, as pronounced in “hiss”. If the higher frequencies of this range are curtailed, (for example by a “Tone Control”), the sound is more like “shshsh”, the “sh” pronounced as in “Shop”. In summary, restricting the bandwidth restricts the amount of noise at the output of the system whether it be loudspeaker or visual display. In general, to maximise the “signal to noise ratio”, the system bandwidth should be no larger than that of the signal, what ever it is.
In a future article we will look at how extremely narrow bandwidths at the receiver, and of the transmitted signal, can allow signals, (of a sort), to be sent and received over great distances even when propagation conditions are extremely poor.