An appropriate procedure enabling the evaluation of the sensitivity of a MESFET small-signal model given unavoidable measurement errors is presented in the first section of this article. This technique, which is based on calculation of the condition number, has been applied to an NE 76000 MESFET device. The results obtained enabled us to highlight the difficulties of an accurate MESFET characterization and to predict the equivalent-circuit element errors. In the second part, we introduce a new approach that makes it possible to reduce the sensitivity of the parameters of the model. When applied to the same transistor, this approach enables us to reduce the sensitivity of the model of 30% on average. © 2004 Wiley Periodicals, Inc. Int J RF and Microwave CAE 14, 174–181, 2004.
Read full abstract