An intelligent data acquisition system for the characterization of power NMOS transistors has been developed. The system uses a pulsed method with adaptive duty cycle. The temperature is empirically determined by using the transistor under test as a thermometer. An accurate database can be obtained for modeling or parameter-extraction purposes. Rules for the rejection of anomalous data due to thermal or electrical transitions have been included in the software. A criterion based on the first and second derivatives has shown that the acquired database quality is better when these rules are used. This criterion is independent of the model considered and the parameter-extraction algorithm used
Published in:
Instrumentation and Measurement, IEEE Transactions on
(Volume:37
,
Issue:
3
)
Date of Publication: Sep 1988