Extraction of BJT model parameters using optimization method

A method for parameter extraction of a SPICE 2G.6 bipolar junction transistor (BJT) model is presented. The proposed approach consists of minimizing a nonlinear objective function to produce a least-squares fit of the model equations to a set of measured device characteristics. All parameters are extracted using a combination of the direct-search optimization method and the variable metric algorithm. This makes it possible to obtain satisfactory results even in the presence of redundant parameters and poor initial values. A special strategy of calculating an objective function during optimization that requires very little CPU time is proposed. An example of parameter extraction of the SPICE 2 G.6 BJT model shows poor model accuracy. Therefore a model modification that gives a very good fit to the measured data is proposed.<
>