Calculating Power Output

Radio Record, Volume III, Issue 28, 24 January 1930, Page 28

Calculating Power Output

Continued from Last Week

By

Cathode

WE have now reached the part of the subject which the redder will probably have most difficulty in grasping; namely, the means of securing and calculating the maximum output available without exceeding the limit of second-harmonic distortion previously mentioned as being permissible, ie., five per cent. Involved in this question is that of the determination of the load resistance necessary for maximum output. | ’ There are two cases to be considered. Firstly, the case in which the voltage approaches the maximum permissible, and the output is limited hy, the plate dissipation. Secondly, the case where the plate voltage.is low and the most suitable load can be used without oyerrunning the valve, : In the first case, the initial operating point can readily be determined in the manner previously described. As a refresher examine the plate voltage plate current family for the UX842 in Fig. 3. Here the permissible plate dissipation is 12 watts and the maximum safe plate voltage 425; with this voltage the plate current must not be greater than 28 milliamps if the rated dissipation is not to be exceeded. To keep the plate current at this value with the plate voltage specified a negative grid bias of 93 yolts is shown to’ be necesgary by the curves. There are two limits which must be observed to avoid distortion: the grid must not become positive and so draw current, and it is necessary to operate on the straight portion of the curves. Fig. 8 seems to indicate that it would be inadvisable to permit the plate current to fall much below 10 milliamps at its maximum, as the "curves" begin to justifiy their name below this value. If the filament of the valve is heated

with direct current, the grid will start to draw cirrent at zero grid voltage; but if the filament is heated with alternating current, it will not be possible to go to zero voltage but the point at which grid current commences to flow will be approximately 4 volts negative. (the UX842 has a 7.5 volt filament). Assuming d,c. on the filament, and utilising the whole of the grid base, it is clear that a bias of .98 volts will permit a swing from 0 grid voltage to twice 98, or 186 volts. No curve for a negative grid voltage of 186 is shown in Fig. 3, but its position may be judged and the lower plate eurrent, point set on this curve at the value of 10 milliamps previously determined as the probable lower limit for distortionless operation. A straight line.is then drawn from this point through the onerating point previously plotted and continued to the 0 grid-voltage curve. This line is shown dotted in Fig. 3. It is now necessary to determine the percentage of second-harmonic distortion, for the figure of 5 per cent. must not be exceeded. Percentage distortion is usually calculated from the equa- tion: 1 max. +1 min. 2 -1mean x 100 1 max. -1 min. Where 1maa.=plate current at least négative grid voltage, 1 min.=plate current at most negative grid voltage, and imean=steady plate current with no signal input (initial operating condition).

In the present case, of course, 1 mean is 28 milliamps, 1 min. 10 milliamps, and 1 mae. is seen from the intersection of the dotted line with the curve corresponding to zero grid voltage to be just short of 48 milliamps. Calculation will show that the second-harmonic under these conditions is less than 1 per cent., so, as the permissible amount is 5 per cent., a lower minimum value of plate current can be chosen and the load characteristic line redrawn. Actually, in this case, a minimum plate current of 8 milliamps may be used without generating too great a secondharmonic, and the solid load line (upper) of Fig, 3 is the result,

The grid voltage at which the curves were made are shown on them. The oblique dotted line is the preliminary load line drawn through the operating point and the 10-milliampere current point at twice the operating grid , voltage. The distortion under these conditions was quite small and a lower limit of minimum plate current was chosen. Thé most suitable line for a plate voltage of 425 is the solid one passing through the same operating point as the dotted line. The other load line 7 ae a plate voltage fs) . Diagram 4, lower. HE process of trial and errér involved in the above method may be considered unduly laborious. The work may be rendered easier by the use of a new device known as a "five-per-cent. distortion: rule." ‘This piece of apparatus is nothing but a short rule which may be conveniently made from a scrap of white cardboard. At a point not far . from the centre a zero is marked, and from this zero divisions are marked off on either side; the size of the divisions is immaterial, so long as they ean be conveniently read, but their relation one to the other is important. The divisions on the right-hand side of the nero are 9-Lliths of those on the left.

That ix to say, if the divisions on | left-hand side are each 1,1 inches, divisions on the right-hand side will each be 0.9 of an inch (in practice the centimetre is a more conyénient measurement than tHe inch). ‘The operation of the rule is simplicity itself. The two plate-voltage plate-current curves corresponding to zero grid voltage and twice the operating grid voltage are first distinctively marked (say with coloured ink). The central zero of the distortion rule is then placed on the predetermined oper--ating point and the rule rotated about this point until the readings on each side of the rule where its edge inter-: cepts the two. distinctively marked curves are the same. The load-line is then drawn jn slong the edge odpthe rule. The ¢cnvenience of this "iitile instrument jis particularly marked where, instead of having the initial operating point determined by the plate dissipation, a plate voltage considerably lower than the maximum permissible is employed; in such casés, a quite éxtended process of trial and error is necessary to determine the best bias and load, and the distortion. rule eliminates most of the labour and calculation involved. (To, be continued next, week.)