User Guide

8
-
10 Tutorial: Fittin
g
Model Data (Bipolar Transistor)
Fittin
g
the Data
The PSpice Optimizer looks for a set of parameters which
minimizes the total squared error between the measured and
simulated curves. When using external data, the optimizer uses
the relative error at each measured point. This means that all
points have equal weight, regardless of the absolute value. This
is essential for fitting semiconductor curves, where the currents
may range over many orders of magnitude. Automatic
normalization is discussed in detail in Target Value Scaling
on
page 4-12.
To fit the data to the model parameters
1
From the Tune menu, select Auto and click Start.
The following table compares the model parameter values
estimated by the PSpice Optimizer to the data values
measured by PSpice for the 2N3055 transistor.
The estimated parameters compare well with the original
values.
The results of the optimization are shown in Figure 8-3.
Parameter
Estimated by
the PSpice
Optimizer
Measured from
PSpice
Is 9.744e-13 9.744e-13
Ikf 4.071 4.029
Br 5 2.949
Bf 98.465 99.49
Ise 9.920e-10 9.025e-10
Ne 1.960 1.941