Broadbanding:Gain equalization directly from data

摘要: This paper begins by presenting a powerful method which is easy to apply many broad-band circuit design problems. Anyone with problem, in the narrow-band case amounts finding point inside of certain circle (say on Smith chart), might find here very useful (Section I). Gain equalization problems fall into this category and main subject conceptually appealing, highly practical, flexible theory gain equalization. The clever matching developed Fano Voula principle handles passive one-ports well except for some difficulty computing gain-bandwidth limitations. It converts problem solutions system nonlinear equations are practice so formidable that typical text book treatments [8], [9] never address issue solving them systematically. Also classical requires load gains be specified as rational functions. Our does good job limitations, reduces all ones eigenvalues eigenvectors given matrix, only data frequency band. effective multiports settles old impedance multiport circuits. concrete results we present are: (1) Two numerically efficient ways determine theoretical bandwidth limitations n-ports; (2) For quick way compute response function optimal coupling directly from answer obtained (1). recent advance microwave technology has produced need more general theories type called based measured avoids functions spectral factorizations until late process. One typically specifies desired profile G(j\omega) then wants largest multiple \kappa G it realizable. procedure described herein suited these needs since determination automatic. A different broadbanding fills was Carlin [10]. approach quite few compromises. possible use lengthier rigorous would check accuracy Carlin's method. In addition quantitative (much easily learned) qualitative properties every (passive or active) designed optimize possesses. They considerable practical any designer can learn instantly thereby obtain (small) amount orientation cheaply. section results, Section Ill, read independently rest (except Fig. 1.1 environs) best designers who have little taste theory. describe viewpoint itself. From perspective an elegant mathematical fits solidly long line mathematics. mathematics underlies computation "prediction error" Wiener's prediction contention natural analog error problem. formulation broaden appeal remember intriguing systems theorists schooled linear IV).