SOMK CIRCUIT ASPECTS OF THE TRANSISTOR W) 



figure of which the unit is capable. This optimum source impedance is 

 best for signal-to-noise performance, not for signal performance alone; 

 hence, as is well known for vacuum tubes, it is usually not a mitch for 

 the unit, and in general both the resistive and reactive components of 

 impedance may be mismatched to the unit. 



For the transistor at low frequencies in the grounded-base connection, 

 reactive effects are negligible and the emitter noise generator may usually 

 be neglected. Under these conditions the optimum noise figure is obtained 

 from a generator of impedance equal to the open-circuit input resistance 

 of the transistor (not the actual working input resistance, which may be 

 quite difTerenl). 



The best operating point for low noise is usually obtained at a moderate 

 collector voltage (20 volts) and a small emitter current (0.5 ma.). 



SUMiL^RY 



A tentative evaluation of the Type A transistor may be made on the 

 basis of presently available information. Before making it, we should say 

 that a comparison with the field of electron tubes is obviously unfair — 

 there are many against one, and a little one at that. Furthermore the little 

 one is a baby not only in size but in length of time under development. 

 It is only natural that the full possibilities are not yet apparent. With 

 these reservations, we can make the following statements about the present 

 Type A transistor: 



Gain: the transistor figure of about 17 db per stage is somewhat low 

 compared to 30 or 40 db obtainable from audio tubes. When the band- 

 width is taken into consideration the gain-band product of the transistor 

 is good but, since the excess bandwidth cannot be exchanged for gain, 

 this number is in this case illusory for narrow-band amplifiers. For video 

 amplifiers the comparison is more favorable. 



Stability considerations differ from the electron tube in such a way 

 as to be likely to give more trouble at low frequencies. At video frequencies 

 this difference is less marked if we play fair by comparing with a triode 

 tube instead of a pentode. The latter is of course better shielded than the 

 transistor. 



Frequency response appears to be practical up to 10 megacycles or more. 



Power output efficiency of around 30%, Class A, seems fully comparable 

 to an electron tube, so that a comparison between the two can be based 

 on input d-c power. 



Noise figure of 60 db at 1000 cycles is much worse than that of a good 

 electron tube, which can come close to db. In view of the frequency de- 

 pendence which brings the transistor noise figure down to 30 db at a mega- 

 cycle, the comparison at video frequencies is less unfavorable, particularly 

 if some developmental improvement can be made. 



