Visual modeling and design of microelectromechanical system transducers

Microelectromechanical systems (MEMS) integrate miniaturized mechanical str uctures with electronics to extend the benefits of planar integrated circui t technology to a broader class of systems, involving sensors, actuators, f ilters, resonators, switches, and wave guides. The mechanical structures, s uch as beams, plates, groves, and diaphragms, implement transduction betwee n energy domains, passive implementations of discrete electrical devices, a nd conduction paths for electromagnetic radiation [F. Frank, J. Staller, Th e merging of micromachining and microelectronics, Third International Forum on ASIC and Transducer Technology, Alberta, Canada, May 1990, pp. 53-60, R . Howe, Silicon micromechanics: sensors and actuators in a chip, IEEE Spect rum, July (1990) 29-35]. To realize the potential and growth of microelectromechanical systems (MEMS ) technology, many new design and manufacturing challenges must be addresse d. The close proximity of the integration of mechanical and electrical doma ins within the small dimensions associated with very large scale integratio n (VLSI) presents new energy-coupling issues. The behavior of the overall s ystem is not the simple concatenation of separate mechanical and electrical behaviors, but the simultaneous combination of the mechanical and electric al behaviors. New modeling, analysis, and design techniques are required to address both mechanics and electronics. In this paper, we address initial design capture and system conceptualization of MEMS transducers based on vi sual modeling and design. Visual modeling frames the task of generating hardware description language (analog and digital) component models in a manner similar to the task of g enerating software programming language applications. A domain is created u sing relevant artifacts and the artifacts are rendered to highlight key des ign aspects. The artifacts may be directly manipulated in controlled ways t o alter design aspects-a process we refer to as design-by-direct-manipulati on. To facilitate the application of visual modeling and design for microelectr omechanical transducers, artifacts, renderings, and associated design aspec ts need to be largely predefined. This requirement leads to a structured to pological design strategy wherein microelectromechanical foundry cell libra ries are utilized. Microelectromechanical system transducer design becomes a process of exploring candidate cells (topologies), varying key aspects of the transduction for each topology, and determining which topology best sa tisfies design requirements. Design renderings and aspects emphasize a circuit level of abstraction. Cou pled-energy MEMS characterizations are presented based on branch constituti ve relations and an overall system of simultaneous differential and algebra ic equations (DAE). The resulting design methodology is called Visual Integ rated-Microelectromechanical VHDL-AMS Interactive Design (VIVID). (C) 2001 Published by Elsevier Science Ltd.