148 



BELL SYSTEM TECHNICAL JOURNAL 



the loud speaker to the input terminals of the attenuator and the 

 attenuator adjusted to give the same meter deflection. The variations 

 in the attenuator settings with frequency showed the variations in the 

 performance of the loud speaker in TU. When the ratio of the open- 

 circuit voltage of the oscillator to the square root of its output imped- 

 ance equalled 1, the setting of the attenuator which gave a voltage 

 between the attenuator output terminals equal to the voltage across 

 the condenser transmitter terminals with a pressure of one bar on the 

 diaphragm, gave the reference zero of one volt, one ohm, and one bar 



30 



3 20 



h 



z 



O 10 



a 



If) 



50 



100 



500 1000 



FREQUENCY 



5000 



10000 



Fig. 4 — Response-frequency characteristics of 115 cycle cut-off exponential horn 

 with moving coil type receiver. Measured outdoors 12 feet from horn mouth at the 

 specified angles with the axis. 



described above. The response of the loud speaker was then read 

 directly from the attenuator setting. 



This measuring method has the commendable feature that the 

 results obtained are independent of any variations from day to day in 

 the sensitivity in the voltage indicating system. The voltage indi- 

 cating system simply serves to compare the voltage produced by the 

 condenser transmitter with the voltage across the output terminals of 

 the attenuator and any variations in battery voltage, tubes, etc., or any 

 variations in the amplification with frequency can in no way affect the 

 accuracy of the measurements. Aside from the condenser transmitter, 

 the only element in the measuring system which must be calibrated and 



