36 BELL SYSTEM TECHNICAL JOURNAL 



Po, and with these repeat the second step. In this connection it 

 should be noted that, in the first step, it is not necessary to choose 

 the P/Po-curve and (M— &)/P -curve which most closely coincide; 

 on the contrary it suffices to choose two curves that are closely parallel 

 (that is, have closely equal slopes at each/). For, corresponding to 

 the nearly constant distance between such two curves, it will only be 

 necessary to supplement the excess-simulator with a series resistance 

 element Pn — which will thus in the complete network be also in series 

 relation to the basic resistance Pi and hence can be merged therewith 

 (even when the requisite R n is negative, provided it is less than Pi 

 in absolute value). 



After the parameters /, Po, and d = T/f have been evaluated, the 

 values for the elements of the excess-simulators in Figs. 7a and 7b 

 can be readily obtained from the two sets of equations (3-C), (4-C), 

 (5-C) and (6-C), (7-C), (8-C), respectively; it thus being found that 



C 2 = <//27r(l+/)Po, 

 C 3 = d/2ir(l+t)tP , 

 P 3 = Po(l+/) 2 , 



C i = d/2irtP , 

 C b = d/2irP , 

 Rb = Po- 



The requisite value for the supplementary series resistance element 

 Pn is evidently 



which will be approximately independent of / if the curves of P/Po 

 and (M — k)/Po chosen in the first step are approximately parallel. 

 If the requisite value of Pa is negative, the basic resistance Pi will 

 merely be decreased by that amount. 



For the limiting form of excess-simulator in Fig. 10b the design- 

 procedure is considerably simpler, because the parameter / is fixed 

 (/ = 0). The two remaining parameters d and P can be evaluated 

 by inspection of two sheets of curves plotted as functions of/: One 

 sheet containing a set of curves of P/Po with d as parameter, and 

 curves of (M—k)/Po with P as parameter; and the other sheet, 

 curves of Q/Po with d as parameter, and curves of N/P with P as 

 parameter. 



