Ph.D Thesis

Ph.D StudentArie Elka
SubjectA Synthesis and Modeling Approach for Micro-
Electromechanical Filters Using Structural
DepartmentDepartment of Mechanical Engineering
Supervisor Full Professors Bucher Izhak
Full Thesis textFull thesis text - English Version


This work deals with the design and topology optimization of micro-electromechanical systems (MEMS) acting as filters by employing structural dynamics concepts.

In the wireless communication field, the integration of filters in the electronic circuit plays a major role in the construction of modern, high performance transceivers. Filters are one of the most important elements in frequency selective circuits. The advance in micromachining manufacturing processes has enabled the integration of micromechanical filters made of silicon in these circuits. MEMS has the potential to replace signal processing circuits that require narrow band-width (high-Q), good signal to noise ratio, and stable temperature and aging characteristics. Recently MEMS filters have been shown to be superior to conventional electrical component in terms of signal quality and energy consumption and size.

It has been shown that the design topology of basic two-port network with lossless mechanical system and modeling approach of electrical models in electrical terms, mechanical models in mechanical terms seems to capture the best of the two 'worlds'. This modeling is in contradiction to the usual modeling in terms of an electrical circuit via the exploitation of mechanical-electrical analogies and conveniently implemented in a circuit simulator.

In the research, novel topologies for mechanical filters, and new conceptual design and synthesis approach are developed. Several configurations have been studied. A multi-degree of freedom system for mechanical filter is tailored to possess superior dynamical behavior that can produce higher order filters with improved characteristic (such as sharp roll off, better stop-band rejection, etc.).  This novel topology can be operated with several different combinations of spatial actuation, sensing and damping. Each of these actuation/sensing combinations can create for the same mechanical filter design a unique spectral shaping.

The research also addresses the optimal shape design of segmented spatial sensors and actuators that excite selected mode shapes in real time and perform modal filtering. It is shown that a precise model-based optimization stage, overcomes the inaccurate implementation of the theoretical solution. An effort has been made to develop a practical method to design an RF-MEMS filter that has the ability to tune the shaped electrodes, which can reduce the modeling errors in the filter computation i.e. fringing field and implementation due to the discretization of the continuous model.

These methods can lead to a general design methodology and provide design guidelines.

The proposed methods are analyzed and simulated for 2D and 3D structures to illustrate their usefulness.