Modified processes capability assessment with dynamic mean shift

Chia Huang Wu*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Rapid advancements in manufacturing technology have seen electronic products becoming lighter and thinner. This has led to production yield requirements of sawing processes in the semiconductor industry becoming increasingly strict. For sawing processes requiring a very low fraction of defective, the process capability index Cpk is widely applied for production yield measurement under the assumption of process stability. However, in practice, even when the typical control charts are implemented, dynamic and unobserved natures of the actual process mean shift may occur. Since the trial control limits are established with reference to random samples, the associated performances are also with uncertainty because of sampling error. In this paper, we investigate the effects of parameter estimation error on the development of the phase I control chart and further compare three methods with various estimators for the process standard deviation. Then, we tabulate the modified adjustment magnitudes for the dynamic Cpk under various subgroup sizes and sample sizes. On the basis of the simulation results, the proposed method can avoid overestimating the true process capability and make more trustworthy decisions. Finally, for demonstration purposes, an application of process capability measures for a sawing process is presented.

Original languageEnglish
JournalQuality and Reliability Engineering International
DOIs
StateE-pub ahead of print - 14 Apr 2020

Keywords

  • control chart
  • dynamic processes
  • mean shift
  • production yield

Fingerprint Dive into the research topics of 'Modified processes capability assessment with dynamic mean shift'. Together they form a unique fingerprint.

Cite this