608 BELL SYSTEM TECHNICAL JOURNAL 



appropriate, since the choice is not quite a matter of indifference. Broadly 

 speaking, the choice will depend upon the relative advantages of nodal and 

 mesh analysis and in most practical situations the former has proved 

 to be the more convenient of the two. One cannot, however, be too dog- 

 matic in this regard. Consider the networks of Figs. 8 and 13. Suppose, 

 for example, that parasitic elements appearing as a passive 11 network had 

 to be superimposed. It is then more conventient to use Fig. 8, not only on 

 account of the ease with which this may be done, but also on account of 

 the fact that the effective transadmittance /3i2 + 1821 is invariant with respect 

 to such a superposition. If, on the other hand, parasitic elements appear 

 as series elements (lead inductances for example) the network of Fig. 13 

 might be more convenient since the effective transimpedahce Z12 + ^21 

 now remains invariant. 



It is also desirable to choose an equivalent network whose elements are 

 capable of being determined by simple measurements; from this considera- 

 tion the network on Fig. 8 is of distinct advantage. 



Application to Triodes 



The preceding section was primarily directed towards the development 

 of possible forms of network representations of the general four-pole equa- 

 tions. In this section one of these forms, namely that given in Fig. 8, will 

 be used to represent the three modes of triode operation. Depending upon 

 which electrode is at a-c ground pot^^ntial, we may distinguish between the 

 following methods of operation: 



1. Grounded cathode operation. 



2. Grounded grid operation. 



3. Grounded plate operation. 



The schematic diagrams, together with assumed voltage and current 

 directions for these modes, are shown on Figs. 16, 17 and 18 respectively. 



With a given set of available terminals the first step in obtaining the 

 networks consists in calculating the four-pole parameters with respect to 

 these terminals. It will be assumed that the coupling circuits have been 

 designed with such efficiency that lead effects can be disregarded, so that the 

 available terminals actually coincide with anode, grid and cathode. This 

 set of available terminals brings us as close to the electron stream as it is 

 physically possible to attain and it represents the ideal towards which design 

 tends. 



It is beyond the scope of this paper to consider the details involved in the 

 calculations of the four-pole parameters. The basic tools needed are the 

 result of a study of the dynamics of the electron stream, which started from 



