1240 THE BELL SYSTEM TECHNICAL JOURNxVL, NOVEMBER 1952 



Computing V^p first, 



The calculated value of — 10.9 \'olts compares quite fa^'orably to the 

 measured — U.O volts. 



Region II is given in this case, approximately by. 



F. 



. ' "^ 7?6 f re + Rc^ 



V Rk 



/?6 + r-c + /?c 



F. ^ (^6.8K + e.8A- + 19i^ + 2.2k ) ^' " '^"^ ^^'^ 



V, ^ (-8.9/0/. - 10.9 (42) 



The first term is of course the slope in Region II and is the magnitude 

 of the negative resistance. The calculated value is —8900 ohms whereas 

 the measured value was approximately —9200 ohms. 



The Region III approximation, derived also from the general rela- 

 tionship is, 



Rb -{-Re Rb -\- Re 



^ (mKK2.2K)) ^ 45(6.8A0 . . 



(Q.SK + 2.2K) ' 6.8/C + 2.2K ^ ^ 



or F.iii = (1785)/. - 34 (45) 



The relation for Region III agrees quite well in slope but not in dc 

 value as may be seen in Fig. 20. Since in this example the Region III 

 behavior is determined essentially by the circuit parameters, it is 

 surmised that the nominal 45-volt battery employed in taking the data 

 was actually 47 volts. 



The Region I check is essentially perfect since the approximation 

 given in Fig. 17 is quite good. 



Note the error at the intersection of the Regions II and III. The 

 broken-line method predicts a sharp transition whereas the actual case 

 is gradual. The deviation is due to the gradual changes in r^ and re as 

 the collector voltage approaches cutoff and is the largest gross error in 

 the approximation. 



It is believed that analysis of this sort will reasonably predict circuit 

 behavior and lead to device requirements. There must be a thorough 

 understanding of the approximations involved and the accuracy will be 

 directly related to the degree to which the original idealized character- 

 istics are approximated. Extended, by means of more than three broken 



