FLUCTUATIONS IN MICROPHONES AND OTHER RESISTANCES 199 



which are presented in this paper indicate that the noise observed in 

 each case is of the same nature and is traceable to the existence of 

 contacts between granules or perhaps granular boundaries. 



Apparatus 



The experimental arrangement used in the measurements to be 

 described here is given in schematic form in Fig. 1. The system in- 

 cludes the input circuit, a high gain amplifier, appropriate filters, 

 attenuator and output measuring device. 



The input circuit consists of the resistance under test, a battery for 

 supplying the direct current, a potentiometer for measuring resistance 

 and voltage, a standard signal oscillator for calibration purposes and 

 appropriate resistances and condensers for coupling to the amplifier. 

 In some cases an input transformer having a high-turns ratio was also 

 required in order to raise the signal level above the amplifier noise level. 

 The granular resistance element was shielded from acoustical, me- 

 chanical and electrical shocks by suspending it with rubber bands 



Fig. 1 — Schematic amplifier circuit for measuring contact noise in granular resistance 



elements. 



inside a tightly sealed iron box which was lined with alternate layers of 

 hair felt and >^-inch sheet lead. The remaining parts of the input 

 circuit were also carefully shielded. 



The high-gain amplifier consists of two separate resistance coupled 

 units, each containing three stages. Each unit is so designed and 

 shielded that the effect of external disturbances is eliminated. The 

 total gain obtainable is about 165 db, with the frequency response uni- 

 form to within 2 db from 10 cycles to 15,000 cycles. In most of the 

 measurements described here, however, a filter which transmitted 

 only those frequencies above 100 cycles was inserted between the first 

 amplifier unit and the attenuating network. 



The gain of the amplifying system could be varied in steps of 20 db 

 by means of interstage potentiometers. In addition, a 600-ohm 

 attenuator having a range of 63 db in steps of 1 db was placed between 

 the filter circuit and the second amplifier unit. The output measuring 

 instrument was a 600-ohm vacuum thermocouple and microammeter. 

 The deflection of the meter was closely proportional to the mean square 

 voltage applied to the couple. Individual noise measurements were 



