The use of linear programming for the design of SAW filters and filterbanks

A linear programming algorithm is proposed for designing surface acoustic wave (SAW) filters or filterbanks with arbitrary amplitude and phase responses. A modified sampling theorem representation is employed for the transducer frequency responses which allows the number of independent variables to be minimized without degrading the filter characteristics. The method can also be used as part of an iterative procedure to generate optimal corrections for second order effects such as diffraction and circuit loading. A simplified algorithm for this procedure is given, and the method is illustrated with theoretical and experimental data from a three channel contiguous SAW filterbank design. Although the method given in this paper is formulated primarily for SAW filters, it is equally applicable to FIR digital filter design.<>