TRACING DISTORTION

Evening Star, Issue 21138, 25 June 1932, Page 4

TRACING DISTORTION

AVith the object of eliminating distortion in receivers, an Australian writer savs:—

Let ns consider what happens to the average broadcast receiver omploving, sav, five valves. .After the incoming signals have been amplified bv two stages of tuned Li. I-', and rendered audible by a detector valve, they are amplified by a further two stages of audio frequency generally employing transformer coupling. It is here that the distortion commences. By audio frequency one infers the range of speech or tone 'frequencies over a band of 150 to 5,000 eveles per second. Frequencies up to 10,000. or oven more in some eases, may be considered as audible. Every experimenter has at some time or other encountered the familiar how] or whistle while experimenting with variable grid leaks, and will remember that the howl can be made so low in note that the period between the chicks '•an almost he younted. By adjustment ot the leak this note may lie gradually raised over the whole audible range until it reaches a high pitch whistle which apparently disappears into inaudibility. It is hoped that this will

servo ns an explanation of audio frequency.! A new distortion in audio frequency amplifiers of the transformer-coupled type may be usually traced to one or more of the following causes:— (1) The presence of oscillation of high frequency in the amplifier circuit'. (2) Resonance effects in the windings of tho transformer due to natural response of the winding. (3) Valves being worked at saturation. ✓ (4) Operating the valves to near tho bends of their characteristic curves; and (5) Distortion due to the iron core. With the many types of laminations used in transformers cores to-day the experimenter should at least be able to find a satisfactory compromise. The writer was somewhat surprised upon dissecting a well-known transformer to find that the “ burr ” of each lamination touched the “ burr ” of the next, thereby destroying the laminated effect of the core. Let us examine tho first cause of distortion —oscillations of high frequency being present in the amplifier circuit and also local oscillations, set up due to couplings, between the output and input sides of the amplifier, are a frequent source of trouble. Tho wireless enthusiast who has somewhat foolishly tried three or more stages of transformer coupling is already familiar with the high-pitched whistle which makes itself known in most cases when he tries to obtain more amplification than tho valve and the transformer can handle. It is not an uncommon occurrence to find that many two-stage amplifiers howl in this manner, Tin's can usually be prevented by correct by-passing in the amplifier itself. Assuming that the first stage has a fixed condenser of perhaps .0002 mfd. across its primary, any audible howl which may then be present can be eliminated by a series of experiments in the form of connecting a .001 mfd. fixed condenser across (a) tho first transformer secondary, (b) the second transformer secondary’, (c) the primary and secondary of each transformer in turn. Different combinations of connections mav bo tried, there being no definite solution to anv particular elusive bond. Again, a howl may he caused by incorrect grid potential, and sometimes a “C" battery works wonders. Resonance effects in the winding of tho transformer may bo the cause of distortion in cases where the output amplitude is fairly large. With some transformers, when receiving strong signals, one may hear tho programme by placing the ear near the transformer in question. The writer lias a power amplifier which, when the loud speaker is removed, can bo hoard to “ play over a room of fairly large area. The obvious solution is not to overload the amplifier or in extreme cases discard the transformers. Where two or more stages of amplification are used, distortion is often caused by tho valves being worked at saturation. In many cases whore the amplitude of the signals is great, the voltage variations applied to the grid of the following valves are so great and at such a variance that operation takes place over all the .straight portions and right around both bends of the characteristics of the valve, causing the peaks of tho waves to bo flattened out. The solution of this problem is tho use of a power valve in place of the offender. Whore a power valve is not available or desirable the plate voltage of the existing valve may be raised and the filament current also increased. It must be borne in mind that any increase in plate voltage must be accompanied by an increase of the negative bias of the grid, in order to work the valve at the correct portion of its characteristic curve.

The writer realises that the average experimenter does not possess such instruments as a sensitive milliameter, volt-meter, etc., to check the curves of his valves. He can, however, obtain in many cases a much better result by careful experimenting in plate voltage, grid bias. . J filament current to bring bis amplifier valves to the correct working point. Too many experimenters to-day use as much as 130 volts on the plates of their amplification valves without anv negative bias whatsoever. So Jar we have dealt with only the audio side of the receiver. Now jet ns consider another part ot the receiver—• namely, the radio frequency and detector stages. The present-day broadcasting stations are each allowed a frequency band or wave band of 10,000 cycles' wide. This allows the addition of frequencies of from zero to 5,000 cycles both above and below the carrier frequency. In a receiver where the radio frequency circuits or tuning circuits are tuned' too sharply this effect will bo found to cut off from the signal. This effect is caused because the peak of resonance is made so sharp that, in place of extending nearly to 5.000 cycles, the band of frequencies actually amplified may be onlv 1.000 or 2.000 cycles wide. This difficulty may be overcome by reducing the sharpness of the tuning. The usual method employed is to increase the coupling between the coils. In the case where a regenerative receiver is used, or any tuned radio frequency receiver having an adjustable control for oscillation, distortion will be experienced if the oscillation control is brought too close to the point of oscillation. When regeneration has been brought up to this stage, the oscillation causes the reproduction of a local oscillation current, which combines with the incoming signal, and thus produces beats or heterodynes The result of this permits only 'the carrier wave frequency to come through properly, and badly distorts the side bands and high notes. Tliis is because on these side bands the locally generated oscillations do not exactly match the oscillations of the signal) and distortion is consequently set up. From the foregoing description the reader will then sec that it is incorrect to force the regeneration to such a pitch that it will introduce distortion into tiro circuit.

Since the advent ot li class stations many listeners are experiencing distortion from overloading detector valves. When the receiver is in close proximity to a powerful broadcasting station

the incoming signal is so groat that it forces the grid of the detector valve so far negative or so lar positive that the valve is unable to handle it. The remedy is to detune the receiver a little. Another point which is very important in the detector stage is to have the correct value, grid leak. If tho value of the grid leak is 100 high, the detector valve will be found to block when a strong signal is received. Ily “ blocking ” it means that an excessive accumulation of negative charges have | taken place on the. grid of the valve, almost overcoming tho signal ampli • tudos. By tho use of a lower value grid leak this" trouble may he overcome. When operating tho receiver on local stations it will be found that a value up to.two megohms will bo suitable. > In the case when tho receiver is timed to a distant station a value up to five megohms can be used. The next point of consideration in the detector circuit is for the purpose of discovering the escape, of radio frequency current in the audio frequency j side of tho receiver. 1 Two serious difficulties result from this trouble. First, the tone quality is damaged, and, second, the volume is decidedly reduced. The audio stages are designed only to amplify the signals placed on tho grid of the first audio valve; therefore, when radio frequency currents get through into these stages, part of the amplifying power is wasted in building up the radio frequency. It , is wrong and must be stopped. - j If the receiver were designed correctly in the first place it would not bo necessary to take those steps, but | for the unfortunate who has made some slight mistake in tho designing of his receiver the following description will show how to elminate the radio frequency from the audio side of the receiver. Place a condenser having a capacity of .001 mfd. across the plate and filament, insert a radio frequency choke in the plate lead to tho first audio transformer, shunting another small bypass condenser across tho primary of tho audio transformer.