HOLE l.WfKCIlOX l\ (il.KMAMI \l 



.U)5 



substantially all of the emitter current is carried by iioles. The theoretical 

 curve shown on the ['"if^ure is 



10/1 Vc I 



0.2 + O.S X 2.5 X |F,,/1()|(1 - e'"'""-') 



This corres])()n(ls to 



r, ^ 10 ^ lJ:{Rc + Rb) 



T \V \ T a R\V \ ' 



(6.11) 



(6.12) 



from which it was conckulcd that for the particular bridge studied t,, was 

 0.2 microseconds. 



>- 1.5 



Ol (u 



M M 



II 1.0 



<3 



0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 00 



'/Vc IN VOLTS-' 



Fig. 10 — a versus l/d'ci showing agreement with the theory for the \alue of /^. 



If Ti is much shorter than r^ , then the holes penetrate the whole filament 

 and /3 becomes 



^ = 



1 — exp( — iajT<) 



— lUT ( I'l 



sin (cor</2) 



?COT< 



(cor,/2) 



(6.LS) 



I''or small values of uti , li approaches unity since (sin x), .v approaches unity 

 as X approaches zero. For wr(/2 = x, the response is zero. This is the condi- 

 tion that Tt = 27r cj = 1 /. For this case the filament is just so long that 

 the modulation is averaged over the time of one cycle of the input signal 

 and since this average includes all phases, the modulation vanishes. 



Preliminary experiments with filamentary transistors, made in accord- 

 ance with the principles discussed above, appear to confirm the general 

 aspects of the theory. Power gains of 15 db have been obtained and fre- 

 (juency responses showing a drop of 3 db in a at 10'' cycles sec. have been 

 observed. Noise measurements indicate an improvement of 10 to 15 db over 

 tlie average type-A transistor for comparable conditions of preparation. 



