DYNAMIC MEASUREMENTS (J.\ I.1J.( TUOMAGNETIC DKVK K- 



I lo5 



ment. The output transformer obeys equation (8), and its characteristics 

 are tabulated on Figure 19. From this equation and a cycle time of 0.1 

 second we find that the change in plate current for a 0.705 henry mutual 

 inductance amounts to 2.83 ma. The dc plate voltage change therefore 

 is 11.32 volts and the corresponding input dc voltage is 1.0. These values 

 are extremely small compared to the voltage swings available, but never- 

 theless the fact that the instantaneous voltage swings are finite imposes 

 a limit to measurement of step discontinuities. 



One direction of transformer current change is where we wish to de- 

 crease the current suddenly. The signal blocks the output tube and the 

 transformer current decays through the equivalent of a 50,000-ohm 

 shunt. The primary inductance is 11.5 henries and the time constant 

 therefore is 0.23 millisecond. The quiescent current is 40 microamperes, 

 so a drop of 2.8 microamperes, assuming the 6L6 blocks completely, 

 would require at least 0.02 millisecond. This means that if there is a 

 measurement discontinuity, the best the circuit can do is to respond as 

 the first part of an exponential over at least this time interval, rather 

 than abruptly. This is shown in Fig. 20, where the indicated time actu- 

 ally approaches 0.1 millisecond. This extension evidently is contributed 

 to by the distributed capacitance of the winding. 



The current rise case is about as favorable. The current rise rate is 

 limited by Lenz's law to a maximiun value of 



di E 2657 ^, , ,,o^ 



5^ = z = ii:55 = ''^/^ (^^^ 



A change of 2.83 microamperes therefore requires at least 0.12 milli- 

 second. This is also shown in Fig. 20. 



The foregoing discussion has been directed toward establishing that 

 the initial delays of Fig. 20 are entirely overload effects and not a fre- 

 quency response effect. Thus if required rates of change during a test 

 do not exceed the available rate, then no error of this type will occur. 



Fig. 21 is a measurement of a fast relay. One curve designated x, is 

 the displacement versus time. The other curve marked i, is the velocity. 

 In neither the displacement curve nor the velocity curve before impact, 

 do the rates approach the overload condition. 



The impacts are shown on the velocity curve to have the extreme rate 

 of change. Such abrupt curves cannot be taken as being literally true. 

 Following such abrupt changes the 10-kc oscillations also are found, but 

 have not been drawn as they have to be discarded. 



