104 



BELL SYSTEM TECHNICAL JOURNAL 



small relative to the sound wave-length that its pressure calibration, as 

 obtained say by Method 1, may be taken to coincide with its field 

 calibration. 



The normal field calibration of a No. 394-Type Transmitter is 

 shown in Fig. 10. The contour of the particular instrument used is 

 shown in Fig. \B. It was suspended from a thin rod clamped to the 

 metal band B. The measurements were made with a Rayleigh disc 

 (0.5 cm. diameter, 2.46 second period), using the modulated sound 

 method.^ The transmitter was placed 32 cm. from the sound source, a 

 1-cm. diameter tube attached to a loud-speaking receiver. The data 

 obtained for frequencies below 500 c.p.s., are believed to be not so 

 reliable as the rest because of appreciable reflections from the chamber 

 walls. 



A = THERMOPHONE CALIBRATION 

 B = ELECTROSTATIC CALIBRATION 

 n_e^, -0.05a MILLIVOLTS 

 *^~p '" BAR 



20 



500 1.000 



FREQUENCY IN rvCLES PER SECOND 



6,000 10.000 20,000 



Fig. 9 — Comparison of two pressure calibration methods. 



For purposes of comparison, the pressure calibration (Method 7) of 

 the same instrument is shown. At the lowest frequencies the two 

 calibrations nearly coincide, as might be expected. At high fre- 

 quencies, say from 1,000 c.p.s. upward, the divergence of the two is 

 quite marked. It has been pointed out by several writers that the 

 difference may be regarded as due to two effects. First, i" as X de- 

 creases, the transmitter tends to cause a doubling of the pressure in 

 front of it as would a rigid wall. Second," the recess in front of the 

 diaphragm (Fig. 1) introduces a broad resonance which has its maxi- 

 mum approximately at 3,500 c.p.s. An estimate of this effect is given 

 in Appendix V. 



The observed differences between the field and pressure calibrations, 

 from 500 to 8,000 c.p.s. are in fair agreement with those computed for 



