452 



BELL SYSTEM TECHNICAL JOURNAL 



Figure 4 shows Fi , F2 and their sum TT'3 — TI'4 plotted versus T. It may 

 be noted that Wz — Tr4 versus T has the same values as SWR versus W 

 in Fig. 2. 



Using equations (16) and Fig. 4, TI'' and W" may be evaluated for the 

 particular values of Ws and Wi in equation (11). In the evaluation, if 

 there is uncertainty as to which reflection coefficient belongs to the wave- 

 guide coupling BE and which belongs to the termination /., a termination 

 with a different magnitude of reflection coefficient should be used and the 

 technique repeated. The reflection coefficient which is the same in the 

 two cases is of course that due to the waveguide coupling BE. 



Fig. 5 — Block schematic for cases III and IV. 



It is assumed in the above solution that multiple reflections between 

 the two impedance mismatches are inconsequential. Appendix A outlines 

 a procedure for evaluating the maximum probable error due to multiple 

 reflections. 



III. Termination Z Perfect, Four Couplings on Hybrid Junction 



In this case the setup might be as shown in Fig. 5. This setup differs 

 from that shown in Fig. 1 in that the hybrid junction has four couplings 

 shown, termination Z' has been replaced by a variable attenuator and a 

 movable shorting piston, and the waveguide coupling FG is to be measured 

 instead of coupling BE. The hybrid junction and the termination Z are 

 assumed to be perfect as defined for case I. 



Since it is the object of the measuring method to measure impedance 

 mismatches in branch B, it is desirable to make the voltage at C depend only 

 on power reflected from branch B. This is accomplished by adjusting 



