382 BELL SYSTEM TECHNICAL JOURNAL 



shielding is complete insofar as it limits the electrostatic coupling to 

 specific points in the bridge, and eliminates coupling between the 

 bridge and the input and output circuits. However, in addition to 

 capacitance across the diagonal corners and across arm AB, capaci- 

 tances are introduced across ri, across R, and across arm AD. The 

 capacitance across AD may be made small enough to neglect since it 

 consists of the capacitance of one condenser lead to the shield. Capaci- 

 tances across ri and across R do not enter as first degree errors in the 

 value of L but do directly affect the measurement of Re. However, 

 where the bridge is used primarily for the accurate measurement of 

 inductance this compromise is justified. Even for the measurement 

 of effective resistance, although the corrections may be larger due to 

 the presence of the shielding, the bridge will give more consistent 

 results and the corrections may be fairly well estimated. 



The method of shielding shown requires one transformer having 

 two shields between the windings and one transformer with a single 

 shield between windings. It is essential that these shields be as 

 perfect as possible. The other shielding shown is comparatively 

 simple, no equipment requiring more than a single shield. The 

 ground is shown at the point B simply because grounding at this 

 point results in the simplest shielding. It would be desirable to have 

 the ground at C in order that one terminal of the coil under test would 

 be grounded, but at the time of balance the points B and D are at 

 the same potential, and provided that ^2 is only a small fraction of 

 the total impedance of the coil under test we may consider that one 

 terminal of the coil is practically at ground potential. However, it 

 should be noted that for a coil having a considerable capacitance 

 from intermediate points in its winding to ground, a ground at B 

 cannot be considered exactly equivalent to a ground at D. This 

 difficulty is only appreciable in the case of very large inductances of 

 large physical size when measured at high frequencies, and in such 

 cases the effective inductance will be dependent on external conditions, 

 whatever bridge circuit it is measured in. In the case of shielded 

 coils, the ground should in all cases be connected to D rather than 

 to B. In spite of the slight disadvantages noted, this method of 

 shielding appears to be the most satisfactory, and a bridge has been 

 constructed in accordance with it. 



Construction of the Bridge 



From the equation giving the value of L, it is seen that we may 

 obtain an additional range for the inductance by having either ri, 

 C3, or both, variable in steps. In the present bridge we have 



