This month’s question:
When a defect is found during interval sampling in a continuous sampling plan (CSP), is it required to inspect the uninspected units between the discontinuance of the 100% inspection and the first detection of a defect? Or is that one defect corrected and 100% inspection resumed without considering the quality of the uninspected units? Does the answer to this question depend on whether the production process is an assembly of parts or production of parts in a machining process?
If a defective unit is found during the sampling of CSP, the defective unit is either reworked or replaced with a good unit. The simple CSP1 procedure, described in the flowchart in Figure 1, assumes the process is stable with a lot proportion defective (p). The average outgoing quality limit (AOQL) achieved is determined by reading the operating characteristic (OC) graph of sampling parameters, clearing interval (i) and sampling of fraction units (f) for a given proportion defective (p).
Per normal procedure, it’s not expected that you go backward and screen the uninspected units. This is because the sampling procedure and the related parameters are statistically designed to ensure a certain quality level. If you are suspicious of the uninspected units between the discontinuance of the 100% inspection and the first detection of a defective unit, you may conduct a 100% screening of those uninspected units in special situations.
If you found additional defectives among the uninspected units, this indicates that one of our assumptions is incorrect. It may either be the process is unstable, the proportion defective (p) is larger than assumed, or the inspection methods are simply inconsistent due to inspector judgment variability from time to time. This could be attributed to human fatigue and other ergonomic factors. If automatic inspection equipment is being used, variability could be attributed to malfunctioning electronics or insufficient lighting, for example.
Other considerations when using a CSP include:
- The process is producing—or capable of producing—materials, and the process quality level is stable (for example, proportion defective or defects parts per million).
- The inspection is relatively quick and easy—for example, verification of visual defects by a person or automatic inspection through a programmed tester. This could be a visual aspect or go/no-go dimension check.
- The inspection is nondestructive because the procedure incorporates 100% screening at the beginning and after the discovery of defects during sampling.
- The sampling procedures can apply to defective units or defects (there may be multiple types of defects being inspected in a given inspection unit). A defect may be individual or grouped in classes.
To answer the second part of the question, this is applicable to the continuous production process. The production units may be arriving through an inspection stage one by one. For example, an automotive manufacturing assembly moves units from one stage to another through an assembly line and reaches final inspection, or mail sorting at a post office through automatic sorter.
Typically, there is no periodic stopping or frequent restarting of the process. In some industries, a process is stopped for every batch and retooled, and process parameters adjusted. Different types of sampling are applicable to these types of processes.
There are several continuous sampling plans (CSP1, CSP2 and CSP 3, for example) and others for special applications. There are differences in procedure and treatment for replacing defective units versus not replacing them. Based on your specific situation, it may be beneficial to review the sampling standards to understand and apply.
Stephens, Kenneth S., The Handbook of Applied Acceptance Sampling Plans, Procedures and Principles, ASQ Quality Press, 2000.
Suresh, K.K., and V. Nirmala, "Comparison of Certain Types of Continuous Sampling Plans (CSPs) and Its Operating Procedures—A Review," International Journal of Science and Research, March 2015, Vol. 4, Issue 3, www.ijsr.net/archive/v4i3/SUB152048.pdf.
U.S. Department of Defense, Military Standard 1235C: Single- and Multi-Level Continuous Sampling Procedures and Tables for Inspection by Attributes; Functional Curves of the Continuous Sampling Plans, March 15, 1998, www.variation.com/files/standards/milstd1235c.pdf.
This month’s response was written by Govind Ramu, senior director, global quality management systems, SunPower Corp., San Jose, CA. Visit them at www.us.sunpower.com.