624 BELL SYSTEM TECHNICAL JOURNAL 



two sets differ only in the indicators used. One requires approximately 

 2 seconds to reach its maximum deflection with a steady impressed 

 signal, and the other requires about 0.2 second. The differences in 

 level corresponding to different types of tubes do not vary greatly, 

 but are nevertheless appreciable. They are, to some extent, a measure 

 of merit of the tube, for a larger difference indicates higher damping of 

 the microphonic disturbance, and high damping is of course desirable. 



A Microphonic Noise Measuring Set 



The type of test set which has been built in the course of this study 

 for use in the laboratory, comprises an arbitrary standard of agitation, 

 a calibrated amplifier, and an indicating instrument. The agitator 

 consists of a heavy, rectangular slate base at one end of which are 

 mounted sockets for several types of tubes. At the other end is an 

 electrically driven vibrating armature carrying a hammer which 

 strikes about 9 blows per second against a steel block bolted firmly near 

 the center of the slate base. This unit is set on a thick felt pad in a 

 felt lined copper box which provides electrical shielding and some de- 

 gree of sound-proofing. The sockets used (except those for the bay- 

 onet-pin bases) are of the type in which contact springs push each base 

 prong firmly to one side, against the body of the socket. This type 

 has been found to stand up well under repeated insertions and with- 

 drawals of tubes and gives as good correlations between repeated micro- 

 phonic measurements as any type which has been tried. 



The amplifier is basically a simple resistance-choke coupled unit 

 having a frequency-response characteristic which is essentially flat 

 (within 3 db) between 80 and 30,000 c.p.s. The tube under test, 

 whose plate voltage is supplied through an 80-henry choke, works 

 directly into a 100,000-ohm potentiometer, variable in 10 db steps, 

 whose output is connected to the input of the amplifier. The indicator 

 is a sensitive thermocouple galvanometer whose scale is marked off 

 in db and half db divisions so that the noise level may be read directly 

 from the setting of the input potentiometer and the position of the 

 indicator. It has been found convenient to think of the noise level in 

 terms of the root-mean-square voltage developed by the tube across 

 the 100,000'Ohm load resistance and to use 1 volt as the reference level. 

 Accordingly, unless otherwise noted, the noise levels given herein are 

 expressed as db below 1 volt across a 100,000-ohm load resistance. 



In order to correct for time shifts in tube characteristics and battery 

 voltages, provision is made for checking the amplifier calibration at 

 any time by throwing a switch which transfers its input circuit from 

 the tube under test to a local oscillator. This oscillator delivers a 



