358 THE BELL SYSTEM TECHNICAL JOURNAL, MARCH 1954 



of the measuring electrodes do not affeft the fundamental balance con- 

 dition of the bridge network. 



By the arrangement described not only are the residual effective 

 capacitances between the measuring electrodes and groimd reduced to 

 a desirable minimum (actually below one /x/xF, including calibrating 

 capacitor and balancing networks), but also any adverse capacitance 

 effects of the cables connecting the bridge to the measuring probes are 

 practically eliminated, even though these cables are several feet long. 



The calibrated grounded direct capacitance range of the bridge ex- 

 tends over 0.32 /x/iF in either direction off balance center position. Any 

 unbalances within the ± 0.25 /x/xF range can be read in increments of 

 0.005 fiixF per division on a recorder. Since covering such a limited 

 capacitance range directly by an adjustable capacitor could present 

 various practical difficulties, a network, dividing electrically the range 

 of a 100 /X)LtF differential capacitor by the ratio of 150 (approximately), 

 has been applied. Using such a network facilitates calibration and ad- 

 justability and greatly reduces effects of the mechanical instability of 

 the variable capacitor. (Similar networks are applied for capacitance 

 and conductance residual balance controls.) 



Stationary unbalances of the bridge network can be measured directly 

 in a conventional manner by rebalancing the circuit with the calibrated 

 capacitor. For unbalances rapidly varying in time, however, this null 

 method could not be applied simply. Therefore, a proportional off-bal- 

 ance deflection method had to be used and various means to ascertain 

 overall linearity between incremental capacitance unbalances and indi- 

 cator deflections were provided, so that eventually variations in linearity 

 no larger than 0.4 db over periods of several days and 0.2 db over several 

 hours have been observed in the actual operating conditions. 



Measurements with the l)ridge depend essentially on the calibrated 

 capacitor. To avoid necessity for freciuent and quite elaborate calibra- 

 tion checking (within a few one-thousandths of a mmF) of this capacitor 

 in a laboratory, a set of supplementary, high stability auxiliary standards 

 has been provided in the test set assembly. The capacitance values (1.05 

 jixjuF; 1.20 hijlF; 1.35 nnF) of these capacitors are so chosen that differences 

 between any pair of them can be compared directly with the calibrated 

 capacitor in the bridge circuit. Reliability of this system is based on a 

 reasonably high probaliility that change in the calibrated ^'alue of any 

 single capacitor will be revealed in the process of mutually comparing all 

 four capacitors. It was felt that this method of ascertaining calibration 

 accuracy at the operating position was particularly recommended in the 

 case of this circuit as its sensitivity to incremental capacitance unbalances 



