A number of numerical techniques are described for the analysls of passive components. Fundamental concepts and typical applications are presented and discussed . A list of references is included for representative works reported in literature.
INTRODUCTI ONAlthough numerical and analytical, if possible, characterizations and modelings of miniaturized passive components have been exercised by a large number of researchers and engineers for some time; the necessity of such activities are increasingly important in recent years. This is in line with increased research and development in the millimeter-wave circuits and the monolithic integrated circuits. It is no longer economical or, in many cases not even feasible, to tune the circuits once they are built. Therefore, extremely accurate characterization methods are needed to model the structures.Since most structures used in today's printed and planar integrated circuits are not amenable to closed form analytical expressions, the numerical methods needed for characterizations are in fact a necessary evil. The circuit designers would like to use CIO packages which in most cases consist of curve-fitting or empirical formulas. However, these formulas must be backed up by accurate characterizations. In addition, any numerical methods for characterizations need to be as efficient as possible both in CPU time and temporary storage requirement, although recent rapid advances in computers impose less severe restrictions on the efficiency of the method. AMother effort in the development of numerical methods has been the versatility of the method. In reality, however, choice of the numerical methods is the product of trade-off between accuracy, speed, storage requirement, versatility, etc. and is quite structure dependent.Since the advent of microwave integrated circuits, a number of methods have been invented in addition to the somewhat more classical methods refined for these modern structures.
SURVEY OF NUMERICAL METHODSWhen a specific structure is analyzed, one has to make a choice of the method best suited for the structure. Obviously, the choice is not unique. Therefore, the user must make a critical assessment for each candidate method. In this paper, we list a number of numerical methods and present geherally accepted appraisals for these methods. It is not possible to make an exhaustive list of all available methods. Rather, most representative ones are reviewed. Although the assessment will be aimed at the characterization of passive components, 3-dimensional structures, many methods are also effective for 2-dimensional problems. Fur-1059 ther, 2-dimensional methods can be used in an integral part of the composite characterization program.(1) Finite difference method [I] This method is well known to be the least analytical. The mathematical preprocessing is minimal and the method can be applied to a wide range of structures. Certain precalutions have to be taken into account when the method is used for an open region problem.(2) Finite element method [2][3][4][5...