18 BELL SYSTEM TECHNICAL JOURNAL 



A supplementary semi-graphical method as an aid to finally arriving 

 at the best proportioning of the networks will be found outlined 

 in Appendix C. 



The Basic Resistance and the Excess Simulator 



The first approximation to a network for simulating the character- 

 istic impedance K of a smooth line is evidently a mere resistance Ri 



(a) o — wwvm — o o — j — 0(b) 



(C) o — WWMV 



R. 



j 



Fig. 5 — Synthesis of the General Form of Complete Network, (a). Basic Resist- 

 ance Element R t for Simulating Nominal Impedance, (b). Excess-Simulator J 

 (Abstractly Symbolized) for Simulating Excess Impedance, (c). Complete Net- 

 work for Simulating Line Impedance 



(Fig. 5a) approximately equal to the nominal impedance k of the 

 line, that is, 



R X = ^L/C, (30) 



and this is a very close approximation, for instance, in the case of 

 open-wire lines at the frequencies of carrier current transmission. 



Over the voice frequency range, however, a mere resistance does 

 not suffice; since there the excess characteristic impedance K — k is 

 not negligible, particularly at the lower frequencies. But the re- 

 sistance Ri equal to the nominal impedance may be retained as the 

 natural basis of a network if it is supplemented by an element or 

 elements such as to approximately simulate the excess characteristic 

 impedance. Such a supplementary network is here termed an " ex- 

 cess-simulator " 7 , and is symbolized abstractly by Fig. 5b; while 

 Fig. 5c represents the corresponding complete network consisting of 

 the basic resistance Ri in series with the excess-simulator, whose 

 impedance is denoted by J. The requisite excess-simulator is obvi- 

 ously less simple in structure and proportioning than the mere basic 

 resistance; whence most of the remainder of this paper will be con- 

 cerned with various specific types of excess-simulators. 



7 But in practice the term "low frequency corrector" has become rather firmly 

 established. It was suggested by the fact that the excess impedance to be simulated 

 is largest at relatively low frequencies. 



