Modeling the "Early effect" in bipolar transistors using an empirical but effective parameter

Previous attempts at modeling the "Early effect" in transistors were not valid over wide ranges of operating conditions. A model parameter is defined here which accounts for the desired effects and has the important attributes of being independent of current, temperature, or other parameters for a given device type. The expressions proposed recently by Logan and by Lindholm and Hamilton for including the Early effect in transistor current gain are shown to be equivalent for all temperatures and currents. A new method is proposed for measuring the empirical parameter VN.