170 



BELL SYSTEM TECHNICAL JOURNAL 



As a further check on the performance of this unit, two inducta.ices, 

 each of about 0.01 henry inductance, were compared at two frequencies, 

 25,000 and 50,000 cycles. Table VI gives the readings obtained in 

 these tests. 



TABLE VI 



At the 25,000-cycle frequency the maximum difference in inductance 

 from the probable correct balance does not exceed ± 5 microhenrys 

 or 0.05 per cent. The resistance balances check to within 0.1 ohm. 

 At 50,000 cycles, the inductance change due to ratio arm reversal is 

 still within ± 0.01 per cent while the resistance change is within 0.1 

 ohm which would be just under one per cent for a coil of this reactance 

 and a reactance to resistance ratio of 300. This is a critical test of 

 the ratio arm phase-angle balance. Hence it may be concluded that 

 over the entire frequency range the ratio coils meet all balance require- 

 ments. The changes in inductance occurring at the higher frequencies 

 when the impedance arms were reversed indicated that the residual 

 capacitance unbalance of these arms was too large. Readjustment of 

 the balancing condenser reduced the changes to less than 0.02 per cent. 

 The difference in resistance balance at the 50,000-cycle frequency 

 indicates that the conductances shunting the impedance arms are not 

 negligible at this frequency. For more accurate results these con- 

 ductances would require balancing. This would be quite practicable 

 by means of a variable high resistance shunt. 



In making each series of tests outlined above, the testing potential 

 applied to the bridge was varied by means of a resistance potentiometer 

 from the lowest value at which a balance could be made to the maxi- 

 mum of the supply oscillator. This was to check the completeness of 

 the shielding and to detect the presence of any coupling with the 

 supply circuit. In no case was there any discernible change in balance 

 produced. 



