534 BELL SYSTEM TECHNICAL JOURNAL 



should be bridged across the measuring trunk to measure the delay of 

 the network under these conditions is obvious from the above discus- 

 sion. Sufficient accuracy can often be obtained by considering only 

 the impedance curve obtained for either the open or the short termina- 

 tion and using the axis of this curve for determining the length of the 

 impedance cycle; e.g., in Fig. 5 (which gives curves corresponding to 

 the open termination) using the lines +150 ohms and zero as the axes 

 of the resistance and reactance curves, respectively. 



b. Determination of Envelope Delay from Phase Shift Measurements 

 Methods will now be briefly described for determining the envelope 

 delay of a system from special measurements of the steady-state phase 

 shift; that is, the difference in phase shift for a definite frequency dif- 

 ference or the difference in frequency for a certain difference in phase 

 shift will be measured and the delay computed therefrom. Three 

 measuring methods will be considered in which the fundamental 

 principles involved are much the same. Practical circuits for measur- 

 ing delay by these methods may be rather complicated from an ap- 

 paratus standpoint in order to facilitate the measurements as much as 

 possible. The details of these circuits are not given here, but they are 

 disclosed in various patents. ^■-' '^' ''* 



The following gives the essential principles on which these methods 

 of measurement are based: The current of the measuring frequency 

 from an oscillator traverses two separate paths and is then combined 

 at the receiving end of the measuring circuit. In the first path no phase 

 distortion is introduced, while in the second the frequency is transmit- 

 ted through the system of which the delay is to be measured. Both 

 paths contain resistance attenuators, so that the magnitudes of the 

 currents received from the two paths may be adjusted as desired. If 

 now a frequency is chosen such that the phase shifts through the two 

 paths cause the two received frequencies to be exactly out of phase, an 

 adjustment can be made so that an observer listening with a receiver 

 to the combined received currents will hear no tone when the two 

 received currents are equal in magnitude. If the frequency is now 

 changed continuously, the observer will hear the tone increase and then 

 decrease again to zero when the two received currents are again exactly 

 out of phase, care being taken that the two received currents are kept 

 equal in magnitude. This means then, that for this difference in 

 frequency, the phase shift in the system under measurement has changed 

 a complete cycle. From what has gone before it is simple to calculate 



^^ U. S. Patent 1 596 941. 



" h'. Nyqiiist and H. A. Etheridge, Jr., U. S. Patent 1,596,942. 



"S. B. Wright and K. W. Pfleger, U. S. Patent 1,596,916. 



