Publications

Abstract not provided.

Abstract not provided.

VOLMAX is a three-dimensional transient volumetric Maxwell equation solver that operates on standard rectilinear finite-difference time-domain (FDTD) grids, non-orthogonal unstructured grids, or a combination of both types (hybrid grids). The algorithm is fully explicit. Open geometries are typically solved by embedding multiple unstructured regions into a simple rectilinear FDTD mesh. The grid types are fully connected at the mesh interfaces without the need for complex spatial interpolation. The approach permits detailed modeling of complex geometry while mitigating the large cell count typical of non-orthogonal cells such as tetrahedral elements. To further improve efficiency, the unstructured region carries a separate time step that sub-cycles relative to the time-step used in the FDTD mesh.

The finite-volume hybrid-grid (FVHG) technique uses both structured and unstructured grid regions in obtaining a solution to the time-domain Maxwell`s equations. The method is based on explicit time differencing and utilizes rectilinear finite-difference time-domain (FDTD) and nonorthogonal finite-volume time-domain (FVTD). The technique directly couples structured FDTD grids with unstructured FVTD grids without the need for spatial interpolation across grid interfaces. In this paper, the FVHG method is applied to simple planar microelectronic devices. Local tetrahedron grids are used to model portions of the device under study, with the remainder of the problem space being modeled with cubical hexahedral cells. The accuracy of propagating microstrip-guided waves from a low-density hexahedron region through a high-density tetrahedron grid is investigated.

A technique to integrate a dense, locally non-uniform mesh into finite-difference time-domain (FDTD) codes is presented. The method is designed for the full-wave analysis of multi-material layers that are physically thin, but perhaps electrically thick. Such layers are often used for the purpose of suppressing electromagnetic reflections from conducting surfaces. Throughout the non-uniform local mesh, average values for the conductivity and permittivity are used, where as variations in permeability are accommodated by splitting H-field line integrals and enforcing continuity of the normal B field. A unique interpolation scheme provides accuracy and late-time stability for mesh discontinuities as large as 1000 to 1. Application is made to resistive sheets, the absorbing Salisbury screen, crosstalk on printed circuit boards, and apertures that are narrow both in width and depth with regard to a uniform cell. Where appropriate, comparisons are made with the MoM code CARLOS and transmission-line theory. The hybrid mesh formulation has been highly optimized for both vector and parallel-processing on Cray YMP architectures.

The Hybrid Thin-Slot Algorithm (HTSA) integrates a transient integral-equation solution for an aperture in an infinite plane into a finite-difference time-domain (FDTD) code. The technique was introduced for linear apertures and was extended to include wall loss and lossy internal gaskets. A general implementation for arbitrary thin slots is briefly described here. The 3-D FDTD-code TSAR was selected for the implementation. The HTSA does not provide universal solutions to the narrow slot problem, but has merits appropriate for particular applications. The HTSA is restricted to planar slots, but can solve the important case that both the width and depth of the slot are narrow compared to the FDTD spatial cell. IN addition, the HTSA is not bound to the FDTD discrete spatial and time increments, and therefore, high-resolution solutions for the slot physics are possible. The implementation of the HTSA into TSAR is based upon a ``slot data file`` that includes the cell indices where the desired slots are exist within the FDTD mesh. For an HTSA-defined slot, the wall region local to the slot is shorted, and therefore, to change the slot`s topology simply requires altering the file to include the desired cells. 7 refs.

Two methods for modeling arbitrary narrow apertures in finite- difference time-domain (FDTD) codes are presented in this paper. The first technique is based on the hybrid thin-slot algorithm (HTSA) which models the aperture physics using an integral equation approach. This method can model slots that are narrow both in width and depth with regard to the FDTD spatial cell, but is restricted to planar apertures. The second method is based on a contour technique that directly modifies the FDTD equations local to the aperture. The contour method is geometrically more flexible than the HTSA, but the depth of the aperture is restricted to the actual FDTD mesh. A technique to incorporate both narrow-aperture algorithms into the FDTD code, TSAR, based on a slot data file'' is presented in this paper. Results for a variety of complex aperture contours are provided, and limitations of the algorithms are discussed.

Sub-gridding techniques enable finite-difference time-domain (FDTD) electromagnetic codes to model apertures that are much narrower than the spatial resolution of the FDTD mesh. Previous thin-slot methods have assumed that the slot walls are perfectly conducting. As the slot depth-to-width ratio becomes large, interior wall losses for realistic materials can significantly affect the coupling through the slot, and therefore these loss effects should not be neglected. This paper presents two methods for incorporating loss for walls with good, but not perfect conductivity, into the FDTD calculations. The first method modifies an FDTD equation internal to the slot to include a surface-impedance contribution. This method is appropriate for the usual FDTD thin-slot formalisms. The second method includes the losses into a half-space'' integral equation that can be used by the recently introduced Hybrid Thin-Slot Algorithm. Results based on the two methods are compared for a variety of slot parameters and wall conductivities.

Finite-difference time-domain (FDTD) codes can, in principle, be used to determine the electromagnetic response of complex scatterers. However, the extent to which structural details can be accommodated is limited by computer resources and one's ability to specify necessary parameters. By embedding into the FDTD code alternative numerical methods that solve the aspects of the problem which are not practical, or possible, for the FDTD code to handle, power and flexibility can be added. This report investigates three such hybrid schemes. Topics include: (1) embedding a transient multiconductor/circuit-analysis code so that coupling down to the component level can be directly computed; (2) the effectiveness of using a multiconductor transmission-line code to analyze shielded multiwire cables in FDTD calculations; and (3) the effectiveness of using two-- and three-- dimensional aperture transfer functions to model narrow apertures in FDTD formulations. These topics were selected because of their immediate need in system assessments. Experimental measurements and/or alternative solution methods are used to verify the hybrid approaches. 56 figs.