.SY)i//; ciKcri r asi'ECTS of the traxsisioh 



.VJ.S 



circuit purposes may be obtained by adding two noise generators to thr 

 equivalent circuit of four signal parameters, as shown in Fig. 27. 



These noise representations are on an entirely similar basis to the signal 

 representations. Just as four elements in any independent configuration 

 suffice for signal description, so two noise generators in either series or 

 shunt in any convenient independent locations can be added to account 

 for the noise. All these representations give the same signal and noise be- 

 havior for any external connections. Still, some may be better than others 

 in corresponding to the actual physics of the transistor; presumably the 



EQUIVALENT CIRCUIT 



Equations: hiZg + ?ii) + iSvi = ''■'a © -Vi 

 ii%A + /2(?- + 2,) = © .v. 



Circled ® signs indicate addition with attention to any correlations which may exist 

 between noise generators or mean square additions if no correlation exists. 



Noise Figure F = 1 + 



1 



_ _ /%, + Zo\ 



4 kTBRa 

 Fig. 27 — Synopsis of general four-pole, including noise. 



better representations will show particularly simple behavior, for example, 

 in their dependence upon the d-c operating point of the transistor. The 

 usual choice puts noise voltage generators in series with the emitter and 

 collector leads, as shown. 



If the two noise generators were truly independent physical sources 

 of noise, their outputs would be expected to show no correlation and their 

 noise power contributions would be simply additive. This independence is 

 not usually the case for the Type A transistor. By adding the noise outputs 

 and comparing the power in the sum to that in the separate components, 

 correlation coefficients ranging from — .8 to +.4 have been found. From this 

 the conclusion can be drawn that the physical sources of noise in the network 

 do not act in series with the leads but at least to some extent arise elsewhere 



