MEASUREMENT OF CAPACITANCE 427 



accordingly the constant K by which the calibrated values of the bridge 

 condensers must be multiplied to give their absolute values, is always 

 very near unity. If the absolute value of the capacitance of a single 

 condenser can be determined the factor K can readily be evaluated by 

 measuring this known condenser on the capacitance bridge. If K is 

 known the absolute value of any condenser can then be determined by 

 measurement on the capacitance bridge because of the consistency of 

 the bridge calibration. The practical reason, therefore, for determin- 

 ing the absolute value of a primary standard of capacitance in terms of 

 resistance and frequency is to permit the determination of the error in 

 the bridge calibration, i.e., to evaluate K. Accordingly, the actual 

 measurements are carried out from the viewpoint of determining the 

 value of K for the precision capacitance bridge rather than from the 

 viewpoint of determining the absolute value of the capacitance of a 

 single condenser. Of course, the latter determination is included in 

 the former. 



Special Apparatus 



Aside from the shielded capacitance bridge the following special 

 apparatus employed in the determinations is worthy of mention. The 

 unit standard condensers were dry stack mica condensers potted in an 

 asphalt moisture-proofing compound and shielded by brass cans. 

 They had been kept in the laboratory a number of years so that they 

 were thoroughly aged and their values extremely stable. The phase 

 difference of these condensers was very small, even for high grade mica 

 condensers. 



Unit resistances were made up especially for this series of tests and 

 consisted of bifilar windings in 100 ohm sections connected in series on 

 hard rubber spools ^ in. in diameter. No. 40 B. & S. gauge advance 

 wire was used throughout. All the coils had phase angles less than .1 

 minute at 1,000 cycles. The 6 dial shielded resistance box used in 

 some of the measurements was a laboratory standard variable from .01 

 to 10,000 ohms and calibrated for phase angle. The oscillator em- 

 ployed as a source of current was a specially constructed vacuum tube 

 oscillator designed to maintain an extremely constant frequency, and 

 to deliver a practically pure sine wave. The reference standard of 

 frequency was a 100-cycle tuning fork surrounded by a constant 

 temperature bath,^ the average frequency of which from day to day was 

 constant to .001 per cent. The reference standard of resistance against 

 which the resistances of the units were calibrated was of the well-known 



*J. W. Horton, N. H. Ricker, W. H. Marrison, "Frequency Measurement in 

 Electrical Communication," A. I. E. E. Transactions, June, 1923. 



