454 BeIl system TECHNICAL JOURNAL 



range of a half a wavelength in waveguide. As for the last case it is neces- 

 sary that the reflection coefficients of the couplings A, B and C be small if 

 good accuracy is desired. 



When measuring the coupling FG the procedure and evaluation are the 

 same as for case II. 



Part of a laboratory setup as used at about 4 kilomegacycles is shown in 

 Fig. 8. It includes a hybrid junction, a variable attenuator, a movable 

 shorting piston, a straight section of waveguide and a movable termination 

 which consists of a cylinder of phenol resin and carbon with a tapered section 

 at one end. It is mounted in a phenolic block so that it may be moved 

 axially in the wave guide. 



Fig. 7 — Part of a laboratory setup as used at 4 kilomegacycles. 



In cases III and IV if the hybrid junction has "poor balance" so that 

 voltage appears at C when input is applied to arm A even though B and D 

 are perfectly terminated, the adjusting procedure will cancel this voltage 

 as well. Measuring accuracy will not be impaired provided the other 

 assumptions are fulfilled. 



Measuring TI'— A Fitting Which Does Not Admit of Measuring 

 Each End Separately 



A piece with a configuration unsuited to the preceding technique may be 

 measured by connecting it between two straight pieces of waveguide such 

 as between flanges F and G in Fig. 5. The IT due to the vector sum of 

 the reflection coefficients of the coupling at one end, any irregularities and 

 the coupling at the other end, is measured. Due to the distance between 

 the mismatches, the vector sum will vary over the frequency band of 

 interest. 



