Self-adjusting diagnostic system for the manufacture of crystal resonators

A self-adjusting diagnostic system for monitoring the frequency trimming process is described. The SAFUDS (self-adjusting fuzzy diagnostic system) is capable of adjusting its knowledge to the changes with two incorporated neural networks serving as a convenient online process monitoring system. The development of the SAFUDS not only fulfills the requirement of automating the diagnostics of the trimming process, but it also demonstrates the advantage of fuzzy logic over conventional logic in implementing the system. The use of fuzzy logic avoids the rigidity of conventional reasoning where it is difficult to process information that is definite or imprecise. One of the major advantages of the fuzzy monitoring strategy lies in the intelligibility and flexibility by which the process condition and control actions can be described directly from the experience and advice of an expert. With fuzzy logic, the system is able to heuristically interpret the condition of the process, like human operators. After the diagnostic system had been implemented for six months, the production yield rate of the frequency trimming process improved from 85% to 95%. During the implementation of the system, the knowledge of the production engineer as well as of the operators regarding the process improved