604 BELL SYSTEM TECHNICAL JOURNAL 



We next observe that the network of Fig. 8 is not unique ; in other words 

 it is not the only possible network. To see this it need merely be observed 

 that from the group of four independent /? parameters in (15) there are 

 several ways in which a subgroup may be selected containing only three of 

 them. For example the ''passive part" of the network in Fig. 8 was con- 

 structed from the three parameters jSn , iSi2 and 1822 . But this is evidently not 

 the only choice. Moreover, the impressed force was taken to be voltage- 

 controlled but this again does not represent the only possibility. 



In general it follows that the "passive part" of the network will reflect 

 at least three properties of the complete network. In Fig. 8, for example, 

 the "passive part" reproduces faithfully the two short-circuit driving-point 

 admittances and the feedback admittance of the complete network. 



Finally it is well to observe that only one driving force is needed in the 

 general network formulation. 



With this background of the general ideas involved let us now further 

 explore some of the possibilities suggested as to other "passive network" 

 constituents. Let us, for example, use jSn , ^22 and ^21 for the "passive part". 

 We then write (14) as 



h = /3iiFi - /321F2 + (/3i2 + /32i)F2l 



(17) 

 h = ^2iFi + /322F2 J 



and from (17) the network representation of Fig. 10 follows. In addition 

 this network differs from that of Fig. 8, in that the impressed current appears 

 on the input side and that it is controlled by the output rather than by the 

 input voltage. As an illustration consider again the triode operated with 

 positive grid. The equivalent network is now as shown in Fig. 11. 



Now let it be supposed that the impressed force be current rather than 

 voltage-controlled. We then first transform (14) into 



V — ^^ ^12 jr 



Pii 





(18) 



and then rewrite (18) as 



' /3ii /3ii ' 



Pll \ Pll / Pll 



(19) 



