THE SOUNDS or SI'EECH 



599 



toppod wave, \vi(l» its vory lar^i' fuiulanuMilal coinponcnt, marie this 

 distortion appear iiuicli worse than would an ordinary sjieerh wave. 

 Fig. 8 ilhistrates the apparatus used to produce the aroustic s(|uare- 

 topp«.-d wa\e. An elertHnle resembling the hark plate of the condenser 

 transmitter was mounted in front of the transmitter diaiihragm. Be- 

 tween this electrtxle and the diaphragm was apjilied a high potential 

 which was made altirnatcK- positive and negative by a commutator. 



Exciter^ 



CJSS 





.001 inch from back folate. Condenfer Transmitter 



Kig. J(-^Con<lfnser transmitter coupled with sf|uarc-toppecl-\vave exciter 



Exciter Parts 



a. Steel Electrode 0.006 inch from Diaphragm, h. Micarta. Insulation. 



c. Supporting Ring. d. Electrode Terminal. 



By this arrangement the desired positive and negative pressures were 

 produced on the diaphragm. The distance between the au.xiliary 

 electrode and the transmitter diaphragm was about .006 inch. This 

 electrostatic coupling was foimd to be sufficiently close to give a 

 suitable deflection of the transmitter diaphragm, while the stiffness and 

 damping of the air film did not alter the (Kiiamical characteristics of 

 the transmitter. 



Fig. 9 is an oscillogram showing the wave form recorded by the 

 apparatus when acoustic square-topped waves of frequencies 84, 153 

 and 306 cycles per second are impressed on the transinitter. Timing 

 waves of frequencies 75, 150 and 300 are also shown. Analysing the 

 original wave by the Fourier mcthcKJ, and allowing for the distortion 

 in amplitude and pha.se of each cotnponent frequency, a computation 

 has been made of the wave form in the output in the case of the square- 

 topped waves of 84 and 153 frequency. The results are shown in Fig. 10. 



