1426 THE BELL SYSTEM TECHNICAL JOURNAL, NOVEMBER 1953 



nated, since the bridge is balanced. Any alternating component applied 

 will be amplified and sent back through the meter in the opposite polarity 

 to that coming directly from the hne. If then the gain in the amplifier is 

 correct and its phase shift nearly zero, the alternating component through 

 the meter will be greatly reduced. 



The bridge provides the constant resistance feature at the input, since 

 the output of the amplifier can have no effect on the input impedance. 

 It does however necessitate a loss from both line and amplifier to the 

 measuring instrument. 



Practically, the difficulty in design is in securing an amplifier with 

 enough peak output capacity and negligible phase shift at frequencies 

 from one cycle up. This has been accomplished by several methods. Most 

 of the phase shift is introduced by the plate condenser in the last stage. 

 The effect of this is reduced by the constant current power supply. 



The gain of the amplifier is controlled by feedback secured by the 

 potentiometer in the diagonal arm of the bridge. The gain is sufficient 

 so that a substantial amount of feedback may be used with a consequent 

 further reduction in phase shift. 



The final phase compensation is secured by the interstage potentiome- 

 ter. The effect of this is illustrated in Fig. 7. The lower curve is the net 

 phase shift of the amplifier without the interstage circuit. The upper 

 curve is the phase shift which may be introduced at the grid of the sec- 

 ond stage. By proper adjustment, these may be made to compensate 

 each other down to a frequency of one or two cycles. The circuit constants 

 are such that the final adjustment of low frequency phase may be made 

 with a negligible effect on high-frequency gain. It may be of interest 

 that if the bridge is unbalanced by shorting the lower 50-ohm resistance, 

 an error of 20 per cent or more is introduced in the dc reading due to the 

 extremely large effective reactance of the feedback amplifier. If the 

 bridge is unbalanced by shorting the instrument the amplifier is quite 

 likely to "motor-boat" and blow the fuse used for protection. It is for 

 this reason that the instrument key may be set to replace the instrument 

 by a 50-ohm resistor. When in this position the fuse is also shorted to 

 avoid needlessly blowing it due to the high surges which frequently 

 occur when the amplifier is turned on and the condensers start to charge. 



The filter as described, omitting the reactance elements from the bridge, 

 would be entirely satisfactory within the power capacity of the amplifier. 

 With certain measurements, notably those of velocity, the peaks of the 

 wave as applied to the filter, which in this case would be proportional to 

 acceleration, are so high that no reasonable amplifier would be able to 

 handle them. These knifelike peaks however are so sharp that they may 



