Load Carrying Capacity of Amplifiers 



By F. C. WILLIS and L. E. MELHUISH 



Synopsis: This paper describes the adaptation of the cathode ray oscillo- 

 graph to the determination of the overload point of vacuum tube amplifiers. 

 Using the input voltage to produce a horizontal deflection, and the output 

 voltage or current to produce a vertical deflection, the amplifier performance 

 is readily determined by noting the resulting figure on the fluorescent 

 screen. So long as the figure is virtually a straight line or an obviously 

 undistorted ellipse, it was found that the amplifier output is free from 

 harmonics. As soon as overloading begins, the oscillogram shows either 

 a sharp bend at either or both extremities of the line or apparent distortion 

 of the ellipse. The method has the advantage of being cjuick. 



IN an\' device used for the amplification of a complex electrical wave 

 such as that necessary for the transmission of speech or music, 

 distortion may arise in tw^o ways, (a) The amplification may not be 

 the same for all frequencies in the band to be transmitted, (b) The 

 relationship between input voltage and output current may not be 

 such as can be described by a straight line when r.m.s. values are 

 plotted against each other. 



Considering the distortion due to cause (b) alone it is true that for 

 most practical devices the relationship between input and output can 

 be described by a cur^•e which is approximately straight for a portion 

 of its length but as the amplitude of the wave to be transmitted in- 

 creases, operation is over a longer portion of the curve and it ceases 

 to be possible to regard the characteristic curve as straight. It there- 

 fore becomes necessary to determine for any design the maximum 

 energy that the system can carry without noticeable distortion from 

 this cause. For a system intended to transmit speech or music the 

 final decision as to how much distortion is permissible must depend 

 upon the judgment of the expert listener but analytical methods are of 

 service in establishing reference points by measurements which can be 

 duplicated without reference to any particular person. The purpose of 

 this paper is to describe some work undertaken for this purpose. 



Departure from the ideal straight line relationship between input 

 and output results in the production of harmonic overtones of all the 

 frequencies present in the input and in the production ot beat notes 

 between frequencies if there is more than a single frequency in the 

 input. It ft)llows from this that if a complex wave is passed through 

 a dexice ha\ing a characteristic of this natin-e the output will differ 

 from the input in the proportion of the different frequencies and may 

 contain frequencies that were not present in the input at all. A 

 change in the proportion of the different frequencies may be partly 



573 



