AUTOMATIC FEATURE RECOGNITION FOR DATA INTEROPERABILITY ISSUES IN HIGHSPEED ELECTRONICS SYSTEM DESIGN

摘要

Technology trends toward higher speed and density devices have pushed high performance electronic system design to its limits. With fine miniaturization of very-large-scale integrated (VLSI) circuits and rapid increase in the working frequency of system-on-a-chip (SoC), the signal integrity has become a major concern. As the operating frequencies enter the gigahertz range, signal integrity issues such as cross talk, power-ground-plane voltage bounce, and substrate losses can no longer be neglected. In order to design high-performance electronic systems with fast time-to-market, it is often needed to analyze whole or part of the system at one fundamentally deeper level of physics. It has begun to be recognized that electromagnetic (EM) field analysis needs to be rigorously included as an addition to traditional circuit simulation. A common problem in this practice is the lack of efficient tools that enable engineers to easily transfer circuit board design data into EM solvers. To partially solve this problem, ACIS SAT has been introduced as a standard data exchange format and been adopted by many software vendors for data import and export. However, efficient data transfer remains a problem as the geometry created in the design package becomes static and no longer feature-based once imported into the simulation package. In this paper,automatic feature recognition algorithms are implemented to help extract features and parameters from the imported static model in SAT format. Case studies will be provided for some representative high speed electronics designs. This work is supported by Research & Technology Development Grant Program of Washington Technology Center with a goal to achieve improved design process for high-speed electronic systems. The developed tool has a potential to speed up the current design process by eliminating laborious manual preparation of design data for EM simulation and allow what-if analysis to be automated to highlight likely signal integrity issues.