LOAD RATING THEORY 



635 



voltage to the rms voltage for a given number of active channels at 

 constant volume. The ratio of the overload voltage for n active 

 channels to the rms voltage of one active channel is given directly by 

 the broken curve of Fig. 3, and the rms voltage for n channels is 

 simply Vw times that for one channel. A simple computation then 

 gives the multi-channel peak factor. This is plotted in Fig. 4 as a 

 function of the number of active channels n. The reduction in multi- 

 channel peak factor as the number of active channels increases reflects 

 the transition from the single-channel distribution curve to the normal 

 curve, as depicted in Fig. 2. 



I 5 10 50 100 500 1000 



NUMBER OF ACTIVE CHANNELS (n) 



Fig. 4 — Multi-channel peak factor for n active channels. 



The Distribution of Equivalent Volume 



The multi-channel peak factor deals only with the effects of changes 

 in the instantaneous voltages of the channels, all other variables 

 being fixed. It is next necessary to extend the treatment to include 

 the effects of the other load variations that occur in practice — those 

 in number of active channels and in channel volumes. It is im- 

 portant, first of all, to notice that the instantaneous-voltage variations 

 occur very rapidly, while changes in the other two quantities are, in 

 comparison, very slow. In the experiments described above, the loads 

 were so fixed that the equivalent volumes could be changed only by 

 changing the operating transmission level of the amplifier; in practical 

 cases the amplifier transmission level is kept fixed, but the equivalent 

 volume is constantly changing because of changes in number of active 

 channels and in channel volumes. 



The amplifier is thus loaded with a constantly changing equivalent 

 volume but because of the great difference in the time-scales of the 

 two classes of variations the load may be regarded as a succession of 

 equivalent volumes, each constant for a small interval of time that 

 nevertheless is long enough to include a representative sample of the 

 resultant instantaneous voltage distribution. If the distribution 

 function for equivalent volume is computed, and then corrected by the 



