MEASUREMENT OF PHASE DISTORTION 



541 



measured. If the detector were balanced with a zero delay system and, 

 then, rebalanced with the system under question inserted, the differ- 

 ence in these readings as given by the phase shifter would indicate 

 the delay of the system. An integral multiple of tt might not be taken 

 care of in this measurement, but this is of little consequence. 



For the modulating frequency of 25 cycles, a phase shift of nine 

 degrees in the envelope of the 25-cycle modulation corresponds to a 

 delay of .001 second. For convenience, therefore, the phase shifter 

 used in this set is arranged in steps so that each step corresponds to a 

 phase shift of nine degrees, or a delay of .001 second. In order to 

 read intermediate values of delay, an auxiliary phase shifter, which 



Fig. 8 — Arrangement for direct measurement of envelope delay at low frequencies. 



consists of a variable condenser bridged across the output circuit of 

 the detector tube, is used and calibrated directly in steps of .0001 

 second. 



This particular delay measuring set has been found quite useful in 

 the frequency range of 300 to about 10,000 cycles per second. The 

 absolute value of delay, of course, is not that which is measured, 

 but this can usually be determined from the measured value by 

 adding this measured value to the integral multiple of .04 second, 

 which is suitable for the case in hand. 



(2) For measurements below 300 cycles, the circuit arrangement 

 shown in Fig. 8 can be advantageously used. This is based on princi- 

 ples exactly the same as those just described but differs considerably 

 in the application of these principles. 



Here a relatively low frequency must be used for modulation. One 

 and a quarter cycles per second has been chosen because it is satisfac- 

 tory for measuring at frequencies as low as 10 cycles and is easily ob- 



