Calculating Power Output

Radio Record, Volume III, Issue 29, 31 January 1930, Page 38

Calculating Power Output

| Single Triode Valve-Concluded

By

Cathode

‘(Continued from last week.) HB reader has already learnt how ¢o insert a "load-line" or dynamic curve (corresponding to a given resistance in the plate circuit) across a family of plate-voltage plate-current curves. He is now faced with the converse proposition; a suitable load-line having been established by use of the distortion-rule or otherwise, it becomes necessary to find the load or resistance to which it corresponds, This can readily be done by reversing the process previously described. At least equally convenient, however, is the following formula :- E max, — B min. Load resistance = I max. — I min. I max. and I min. have been ‘previously defined; E max. is the instantaneous voltage at the plate when bias is twice its steady value (in Fig. 3, reading down from the termination of the heavy load line gives this as 628 volts), and H min. is the instantaneous voltage at the plate when bias is zero (176 volts In the instance quoted). It will be-noted that, as previously explained, the maximum instantaneous voltage is substantially greater than that of the source. Load ‘resistance in the present instance is, then,.- a 628 — 176 ~ 0.058 — 90.003 az «=«--« 060 Ss ohms. This is the value of combined resistance and reactance (added vectorially, as will be explained shortly) which the speaker should possess if the maximum possible output is to be secured from the valve. ‘Naturally, the condition can only be satisfied at certain fre-

quencies, since the speaker reactance varies with frequency. ‘ The calculation of the actual power output presents: no additional difficulty. .The formula is:Power output = 4 (BH. max. — B. min.) (I, max. — I min.). _ Applying. this to the instance previously given :- Power output + (628-176) (0.059 — 0.003) = 8.16 watts. Where the plate voltage does not approach the maximum for which the particular valve in use is rated (so that one is not limited by considerations of plate dissipation) it may be helpful to remember that the optimum load in these circumstances will be twice the a.c. plate resistance at the point of maximum plate current. This condition must be emphasised, as it is one that is often overlooked, the statement being loosely made that the optimum load is twice the a.c. ‘plate resistance of the valve; naturally the reader concludes that the plate resistance referred to is that taken under operating conditions, and is thus misled. Some valve manufacturers ex: pressly state that the listed a.c.' resistance is taken at zero grid bias, and this approximates to the point of maximum plate current. " Most manufac- turers, however, give the a.c. plate resistance under operating conditions. With a valve of sound design, the necessary condition is fulfilled when the load is approximately 1.6 times the a.c. plate resistance of the valve under operating conditions. It must be distinctly understood that it is permissible to make an assumption as to the optimum load only when the conditions are

such that the plate dissipation is nota determining factor. The reader -should now: be in a position .to ascertain with accuracy the maximum output of any power valve concerning which he has, or is prepared to obtain, sufficient . information, This is a very valuable accomplishment. At the same time, it is not always necessary or convenient to determine the output with such great accuracy; for example, it is sometimes desired to make a rough comparison of the output of two valves, and to this end some simplified, method of calculation is called for. It seems to the writer that by far the best approximation of, this kind lies in a simplified version of the foregoing process. Remembering that the optimum load for a power valve is twice its a.c. resistance at the point of maximum plate current, it ‘can be

shown that, under these conditions, the output equals :- } Ra | Output = (1 max..- 1 min.)2 x -- 4 where Ra is the a.c. plate resistance of the valve, and the other terms have the same interpretation as heretofore. 1 maa. and 1 min. may be estimated by setting down on a separate piece of squared paper (asin Fig. 5) a skeleton anode-current anode-voltage curve derived from the curves published. In the figure, A and B are "wo points taken for Hg (grid-bias) = 0 from published curves of a power valve at plate -voltages of 150. and 100. 1 max..is: given by the point where the line CD, whose slope is 2Ra, cuts the curve AB. The-position of CD can be fixed because an anode current, or a grid bias, for maximum plate voltage, is always recommended by the manufacturer. In this example the maximum. plate | voltage .was 200 and the recommended

eae on bias such that the plate current is cal milliamps. (In practice, as previously explained, it is sometimes convenient to draw first any. line C’D’ of the required slope and move CD along ‘parallel to. C’D’ until: it. cuts the vertical corresponding to maximum anode volts at the required point). The permissible value of 1 min. will usually. be betweer 0.5 milliamps and 5 milliamps. In any case this figure will be very small’ compared. with 1 magz., and for approximation will, not affect the result. , Thus; if we assume in ‘the éxanipie of Fig. 5 that 1 min, is 1, 5 milliamps, we have Output = (0.0495 — 0.0015)2 x 1750 TT = 1 watt ‘approximately. It is, of course, necessary to’ satisfy oneself that the second-harmonic dis‘tortion does not exceed the’ permissible ' five per cent., but the method of doing this has already been explained. There are certain -formulae available ‘for calculating power output’ without reference to the characterictic curves . o£ the valve involved. . Any such ~method; however, ‘suffers from ‘the disadvantage that one has ‘no: means: of . ascertaining the distortion: under the particular conditions for which the out- _ put is calculated... The most familiar of these formulae is > — m2 Hg2 Rn Output = . 8. (Ru +.Ra)2 Where m = amplification factor, Ra the a.c. resistance of the valve, Rt -the , impedance of the load at a given fgequency, and Eg the peak, grid swing. It is, in some instances, permissible to assume that the optimum value of Ru is twice Ra. The familiar statement that Eg may Je ascertained by doubling the grid-bias voltage is, however, very open to quéstion, since, with most power valves rsed according to the maker’s instructions, a peak gridswing as great as this would result in far too great a second-harmonic distortion. There would, of course, be no danger of grid-c urrent distortion with this grid-swing, but the reader who has properly grasped the foregoing ex‘planation will appreciate the necessity of avoiding harmonic distortion as well. American power valves are usually

This paper has already reached such a length as to render it inadvisable to treat push-pull, parallel, and pen valves in the one series.’ These Subjects, then, will be covered in a further paper, which will’ appear shortly, and which will contain, in addition, information regarding’ the:load' which different types’ of speaker may be expected to comprise.

stated to have'a certain "maximum undistorted output." The same information is. not generally given with valves of English manufacture, so the following table for power valves of the Marconi and Osram series may. ) useful :- VALVES OPERATED AT MAXIMUM PLATE VOLTAGE. Type of Plate Optimum Output valve volts . load

impedance (ohms) D.E.P.240 150 7,000 0.386 watts P.425 150 5,000 0.26 45 P.625A 180 ‘2,600° 090 ,, P.625 250 4,800 0.96 5, LS5A 400 9,150 253 4 LS6A 400 .4,600 . 514. ,