THE TRANSISTOR AT HIGH FREQUENCIES At low frequencies it is assumed that the transistor responds instantly to changes of input voltage or current. Actually, such is not the ease, because the mechanism of the transport of charge carriers from emitter to collector is essentially one of diffusion. Hence, to find out how the transistor behaves at high frequencies, it is necessary to examine this diffusion mechanism in more detail. Such an analysis! is complicated, and the resulting equations are suggestive of those encountered in con- nection with a lossy transmission line. A model based upon the trans- mission-line equations would be quite accurate, but unfortunately, the resulting equivalent circuit is too complicated to be of practical use. Hence it is necessary to make approximations. Of course, the eruder the approximation, the simpler the circuit becomes. The hybrid-pi model developed in this chapter gives a reasonable compromise between accuracy and simplicity. Using this model, a detailed analy- sis of a single-stage CE transistor amplifier is made. 11-1 THE HYBRID-P! (IT) COMMON-EMITTER TRANSISTOR MODEL? In Chap. 8 it is emphasized that the common-emitter circuit is the most important practical configuration. Hence we now seek a CE model which will be valid at high frequencies. Such a circuit, called the hybrid-tI, or Giacoletto, model, is indicated in Fig. 11-1. Analyses of circuits using this model are not too difficult and give results which are in excellent agreement with experiment at all frequencies for which the transistor gives reasonable amplification. Furthermore, the resistive components in this circuit can be obtained from the low-frequency h parameters. All parameters (resistances and capacitances) in the 348Sec, 11-1 THE TRANSISTOR AT HIGH FREQUENCIES / 349 Fig. 11-1 The hybrid-IT model for a transistor in the CE configuration. E E model are assumed to be independent of frequency. ‘They may vary with the quiescent operating point, but under given bias conditions are reasonably constant for small-signal swings. Discussion of Circuit Components The internal node B’ is not physically accessible. The ohmic base-spreading resistance ry, is represented as a lumped parameter between the external base terminal and B’. For small changes in the voltage Vs. across the emitter junction, the excess-minority-carrier concentration injected into the base is proportional to Vee, and therefore the resulting small-signal collector current, with the collector shorted to the emitter, is proportional to Vi. This effect accounts for the current generator gn Vy, in Fig. 11-1. The increase in minority carriers in the base results in increased recom- bination base current, and this effect is taken into account by inserting a conductance gm. between B’ and E. The excess-minority-carrier storage in the base is accounted for by the diffusion capacitance C, connected between Band E. The Early effect (Sec. 5-5) indicates that the varying voltage across the collector-to-emitter junction results in base-width modulation. A change in the effective base width causes the emitter (and hence collector) current to change because the slope of the minority-carrier distribution in the base changes. This feedback effect between output and input is taken into account by connecting ge, between B’ and C. The conductance between C and E is Gea Finally, the collector-junction barrier capacitance is included in C.. Some- times it is necessary to split the collector-barrier capacitance in two parts and connect one capacitance between C and B’ and another between C and B. The last component is known as the overlap-diode capacitance. Hybrid-II Parameter Values Typical magnitudes for the elements of the hybrid-II model at room temperature and for Ic = 1.3 mA are gm = 50 mA/V Tr = 100 2 tw. = 1K Tre = 4M Te = 80K Cc. = 3 pF C. = 100 pF That these values are taken as reasonable is justified in the following section.350 / INTEGRATED ELECTRONICS Sec. 11-2 11-2 HYBRID-IT CONDUCTANCES We now demonstrate that all the resistive components in the hybrid-II model can be obtained from the A parameters in the CE configuration. Transistor Transconductance g» Figure 11-2 shows a p-n-p transistor in the CE configuration with the collector shorted to the emitter for time-vary- ing signals. In the active region the collector current is given by Eq. (5-3), repeated here for convenience, with ay = a: Ic = Ico — ale Since the short-cirouit current in Fig, 11-1 is gmVo., the transconductance Jn is defined by _ dlc} _. Oe _| ale 9m = FVas lvoe °° OVun OVE (et) In the above we have assumed that ay is independent of Vx. For a p-n-p transistor, Vz = —Vz-e, as shown in Fig. 11-2. If the emitter diode resistance ist, then r. = 9V2/@Iz, and hence On = 2 (11-2) The dynamic resistance of a forward-biased diode is given in Eq. (8-14) as Vr/Ix,t where Vr = &T'/q, and hence — ele _ Ico = Ie Paresh Ve (11-3) For a p-n-p transistor Ic is negative. For an n-p-n transistor Ic is positive, but the foregoing analysis (with Ve = +Varz) leads to gm = (Ie ~ Ico)/ Vr. Hence, for either type of transistor, gm is positive. Since |Z¢| > |Ico|, then gm is given by ~ Hel Im ~ (1-4) Vee Fig. 11-2 Pertaining to the derivation of gn. Since the recombination current in the emitter space-charge region does not reach the collector, the factor » in Eq. (3-14) is taken as unity in the calculation of gu.