56 



NOISE MEASURE 



[Ch.6 



that produced the canonical form, we obtain a 4w-terminal-pair network 

 which has the same terminal behavior as the original dissipative im- 

 bedding network. The resulting network is shown in Fig. 6.8. It has 

 2n available terminal pairs. The 4w resistances can be comprised in a 

 4w-terminal-pair network with a characteristic-noise matrix, the eigen- 



ES 



L 



Passive dissipative 4re-terminal-pair 

 imbedding network 



Fig. 6.8. Imbedding of n two-terminal-pair amplifiers in lossy 4«-terminal-pair network. 



values of which are all negative (Sec. ?>.?)). The n amplifiers can be 

 grouped correspondingly into a 2w-terminal-pair network with a charac- 

 teristic-noise matrix that has (in general, positive and negative) eigen- 

 values equal to those of the characteristic-noise matrices of the original 

 amplifiers. The network operation in Fig. 6.8 is obtained from that 

 corresponding to Fig. 4.3 by a subsequent reduction from 6n to 2« output 

 terminal pairs. Again, the least positive eigenvalue of the characteristic- 

 noise matrix is not less than that of the best amplifier. Thus, we may 

 state the following theorem: 



Consider a general lossless or passive dissipative interconnection of an 

 arbitrary number of different and independently noisy two-terminal-pair 

 amplifiers that results in a two-terminal-pair network. When the resulting 

 two-terminal-pair network is driven from a source that has an internal 



