Accelerated numerical modeling of RF circuits using network characteristic mode analysis

A fast numerical modeling approach based on network characteristic mode analysis (CMA) is presented and investigated for analyzing electrical layouts in miniature RF filters, such as surface acoustic wave filters. In this approach, a generalized eigenvalue decomposition is performed on the Z-parameters of an electrical layout at one or two sampling frequencies that can be computed and extracted with any numerical full-wave method. The obtained eigenvalues are used to extract modal resistance, inductance, and capacitance matrices for each eigenmode. The frequency dependence of the modal resistance matrix can be assumed a priori or determined automatically, and the modal inductance and capacitance matrices are assumed frequency independent. These modal matrices are then used to approximate the Z-parameters at any other frequencies to provide the response of the electrical layout, which can then be combined with the frequency responses of other components, such as resonators, to yield the electrical response of an entire RF filter. Compared with the previously developed analytic extension of eigenvalues, this fast CMA-based method is less affected by the frequency variation of eigenmodes since the frequency dependences of the eigenmodes are implicitly canceled out in its formulation. The accuracy of this approach is evaluated by comparing with results from full-wave analyses. For RF circuits whose electrical sizes are small and whose frequency range of interest is relatively small, the proposed CMA-based fast frequency sweep approach is found to be sufficiently accurate and highly practical for engineering applications.