Behavioral modeling of D/A converter via nonlinear auto-companding wavelet approximation method

In this paper, a nonlinear auto-companding wavelet approximation method is proposed for the behavioral modeling of analog circuits and applied to the behavioral modeling of a switched-current (SI) digital-to-analog converter. The companding function is automatically constructed according to the initial error distribution when approximating the input-output function of the circuit blocks by conventional wavelet collocation method. The proposed method works more efficiently than the published modeling approaches in regulating the error distribution and reducing the modeling errors as well.