Matching networks, as used between the consecutive stages of an amplifier to manage the transfer of power between the complex source impedance, the active devices, and the load, continue to be of importance, particularly in microwave systems. Synthesis techniques for these networks have evolved from the classical theory to novel numerical approaches that yield significant practical advantages. We review here several well-known design methods, including the classical derivation, sloped-approximation synthesis, and the real-frequency technique. Also considered are enhancements and variations of these methods. Furthermore, we describe the iterated analysis approach, which appears to provide the greatest design flexibility, eficiency, and accuracy to date. Examples are given which illustrate the relative merits of some of the more modern computer-assisted techniques. Finally, we discuss ideas for further research in the area of matching network synthesis.
MATCHING NETWORKS IN MI C R 0 WAVE C I R C U ITSIn the microwave regime. most amplifiers employ matching networks between their consecutive stages to manage the transfer of power between the active devices. Typically, the job of a matching network is to achieve maximum transfer of power between complex impedances over a specified continuous frequency band. subject to potentially conflicting constraints. such as amplitude equalization or noise-figure minimization. We often seek the best compromise among the aforementioned design constraints. Classical design cannot readily handle such simultaneous requirements. and nowadays the most effective synthesis techniques are necessarily numerical adaptations of the classical methods or other computer-based approaches. With this in mind, the design task is most conveniently rephrased as a numerical problem which can be summarized in virtually all cases as follows.
The Numerical Broadband Matching Network Design Problem0 Given n frequency points that define a passband: . . .A useful rule of thumb for choosing n, the number of frequency points, is that n should be at least twice the degree of the network to be designed. Regarding terminology, we note that the "doublematching" problem refers to the case where complex impedances are given for both the source and load; whereas "single-matching'' refers to the case where either the source or the load is constrained to be a constant resistance. Furthermore, we observe that "equalizer" has been commonly used as a synonym for "matching network" (e.g., refs. 24,26-31), and we shall adopt this usage here. Strictly speaking, the classical interpretation of matching network design has been to achieve the highest value offlat gain possible in the frequency band; however, microwave amplifiers often require amplitude equalization to compensate for the gain roll-off of the active devices. Consequently, most modern matching network design algorithms allow for the optional inclusion of such equalization.Depending upon the application, either lossless lumped-element or transmission-line ...