Introduction Statistical process control (SPC) is by now a commonly used quality assurance tool. One of the most important SPC techniques is the use of control charts for maintaining process performance; their effectiveness lies in their ability to alert users of any deterioration of process behaviour so that potential damage can be alleviated or averted in time. As a result, the attention of control chart users, mostly technical personnel, is normally focused on process deterioration. From a management point of view, the ability to detect process improvements in the form of "out of control" states is no less important, since improvements can take place as a result of process stabilization or learning effects. In this article we address this subject with reference to a conventional process control chart and its application in a high yield or low defective manufacturing environment, and offer some practical implementation procedures. Many manufacturing processes today have attained quality levels far higher than those commonly found when Shewhart control charts were first proposed in the 1920s[1]. One of such control charts, namely np-chart for the control of nonconforming items (defectives), faces some implementation problems when the proportion nonconforming p is low and the size of each sample n is not large enough. In particular, based on 3ơ deviation from process mean, the lower control limit (LCL) is often below zero which has no physical meaning. It is often advised in textbooks or journal papers that this LCL be ignored or zero be taken instead as the lower limit. However, such practice contradicts the fundamental idea of SPC by control charts. A process is judged out of control when it, with a certain probability, performs worse than expected; in such a case action should be taken to make the necessary corrections to the process. The process may also be "out-of-control" in the sense that it has become better than expected; in such a situation useful information can be sought about how the new-found quality improvement may be sustained. This is an extremely valuable opportunity not be missed, as continuous improvement is now the cornerstone of modern quality management.
[1]
Nicholas L. Squeglia,et al.
Zero Acceptance Number Sampling Plans
,
1994
.
[2]
Thong Ngee Goh,et al.
Some procedures for decision making in controlling high yield processes
,
1992
.
[3]
G. Rex Bryce.
Statistical quality control for manufacturing managers
,
1987
.
[4]
Thong Ngee Goh.
A Charting Technique for Control of Low‐Defective Production
,
1987
.
[5]
John S. Oakland,et al.
Statistical Process Control
,
2018
.
[6]
D. Coleman.
Statistical Process Control—Theory and Practice
,
1993
.
[7]
Lloyd S. Nelson.
Statistical Quality Control, 6th Ed.
,
1991
.
[8]
Thong Ngee Goh,et al.
Statistical monitoring and control of a low defect process
,
1991
.
[9]
Barrie Dale,et al.
Some problems encountered in the construction and interpretation of control charts
,
1990
.
[10]
Thong Ngee Goh,et al.
New approach to quality in a near-zero defect environment
,
1994
.
[11]
Thomas P. Ryan,et al.
Statistical methods for quality improvement
,
1989
.
[12]
T. Calvin,et al.
Quality Control Techniques for "Zero Defects"
,
1983
.
[13]
Eugene L. Grant,et al.
Statistical Quality Control
,
1946
.
[14]
Josef Schmee.
Attribute Sampling Plans, Tables of Tests, and Confidence Limits for Proportions
,
1984
.
[15]
A. R. Crathorne,et al.
Economic Control of Quality of Manufactured Product.
,
1933
.