Distributed high frequency effects in bipolar transistors

The origin of distributed high-frequency effects in bipolar transistors is reviewed. The occurrence of these distributed effects as a function of frequency is related to the approach of characteristic frequencies, f/sub T/ and f/sub ac/ for vertical and lateral effects, respectively. It was shown that a hybrid- pi AC model provided with a transcapacitance between base and collector correctly predicts the phase of the transadmittance up to f/sub T/. The transcapacitance value naturally follows from a physical large-signal charge model and incorporates neutral base and base-collector depletion region delay. AC crowding in the internal base can be effectively modeled by adding a capacitor in parallel to the intrinsic base resistance. The capacitance value is related to the total input capacitance at the base. A critical emitter width which makes it possible to determine when AC crowding dominates vertical NQS (non-quasi-static) effects is defined. It is shown that when the model for the intrinsic transistor is properly extended with parasitic elements, as has been done in the MEXTRAM model, HF small-signal transistor behavior can be correctly simulated up to the highest frequencies of practical interest.<<ETX>>