Simultaneous Parameter and Tolerance Design for an Electronic Circuit Via Computer Simulation and Statistical Optimization

摘要

Conventionally, parameter design is carried out prior to tolerance design for economic reasons. However, a combined parameter and tolerance design may be completed in one stage for further cost reduction and quality improvement. In this study, a simultaneous optimization of component parameter and component tolerance for an electronic circuit design is realized via computer simulation and Response Surface Methodology (RSM). The approach first generates a set of outputs (experimental data) through computer simulation based on various combinations of parameter and tolerance levels as inputs. Then these outputs are converted into a total cost as a response value before applying RSM for statistical analysis and optimization. The response value(total cost) includes quality loss, tolerance cost, and failure cost, which reflect the combined effect of the assigned parameter and tolerance values being assigned. In this study, computer simulation can be carried out by two methods. The first method uses Monte Carlo Simulation to generate the outputs with the condition that the transfer function is known. The second method uses a PSpice simulation program to generate outputs without necessarily knowing the transfer function. The results provide designers with the optimal component parameter and tolerance values, and the critical components, and the response function-all of which are very important during the early design activities as they enable designers to have a repeated application, accurate feedback and appropriate suggestions, particularly under uncertain design conditions.