Frequency response of theoretical models of junction transistors

SMALL-SIGNAL parameters are very useful for determining the electric-circuit performance of a transistor. However, for a junction transistor these parameters are not independent of frequency as is now fairly well known. One type of frequency variation in the parameters occurs because of the feedback mechanism and the effect of collector capacity as described by J. M. Early.1 A second type of frequency variation occurs because of the diffusion process by which junction transistors operate. Frequency variations of parameters due to the diffusion process for the ideal one-dimensional model introduced by Shockley have been calculated independently by the writer,2 by J. M. Early3 and by H. O. Johnson.4 Addition of a constant base-spreading resistance to the base lead of the one-dimensional model completes the over-all theoretical model commonly used. In practice, this model applies fairly well to alloy, or fused junction, transistors.6 On the other hand, for grown-junction transistors, the distributed nature of the transistor parameters in the transverse direction of the base region must be taken into account. Analysis of such a two-dimensional model6 has shown that under simplifying conditions such a transistor may be represented by the usual one-dimensional ideal model plus a complex frequency-dependent base impedance in series with the base lead. Hence, the grown-junction transistor can be represented by the same type of model as that used for the fused-junction transistor, but with the constant base-spreading resistance of the latter model replaced by a more general complex base impedance. In practice, this two-dimensional model applies fairly well to grown-junction transistors having a base connection that approximates a line contact. For grown-junction transistors in which the base connection is not a line type of contact, the base impedance may be partly resistive and partly complex.