Inspection is a mandatory but non-value-adding activity whose purpose is to protect an internal or external customer from poor quality. Less is therefore better, as long as we give the customer the necessary level of protection.
The ANSI/ASQ Z1.4 standard for acceptance sampling (formerly MIL-STD 105) offers double and multiple sampling plans that reduce the average sample number (ASN), while sequential sampling reduces it even further. A zero acceptance number (c=0) plan makes the sample size even smaller, and any ANSI/ASQ Z1.4 plan can be converted into a c=0 plan that provides identical or superior protection against poor quality down to the rejectable quality level (RQL).* Nothing is however free in industrial statistics, and the trade-off is a much higher risk of rejecting lots at the acceptable quality level (AQL) which means these plans should be used only when quality far exceeds the AQL. If this is not the case, the previously mentioned alternatives such as double, multiple, and sequential sampling should be used.
* ANSI/ASQ Z1.4 plans do not have formal RQLs but we treat the nonconforming fraction at which there is a 10% consumer's risk of acceptance as the RQL to define c=0 plans as well as sequential sampling plans and narrow limit gaging plans.
Webinar Objectives
Attendees will learn how to convert any ANSI/ASQ Z1.4 normal or tightened sampling plan into a zero acceptance plan and generate an operating characteristic (OC) curve to assure the customer that it provides equal or superior protection against poor quality up to and including the RQL.
- Know the basics of how to use ANSI/ASQ Z1.4. This will be a review for those already familiar with it. This includes selection of the sample size n, acceptance number c, and also use of the switching rules.
- Know how to define the rejectable quality level (RQL) as the nonconforming fraction for which the lot has a 10% consumer's risk chance of acceptance, and use this to define a zero-acceptance sampling plan with sample size n.
- Know how to calculate the operating characteristic (OC) curve that shows that the c=0 plan has an equal or greater chance to reject the lot at nonconforming fractions up to the RQL, thus assuring the customer that the plan offers more than adequate protection against poor quality.
- If the quality level is not sufficiently good to support zero acceptance sampling, know which alternatives can be used to reduce the average sample number.
Webinar Highlights
- Selection of the sample code letter, sample size n, and acceptance number c under ANSI/ASQ Z1.4 given the inspection level, lot size, and acceptable quality level (AQL)
- Conversion of any ANSI/ASQ Z1.4 normal or tightened inspection plan into a c=0 plan. A formula is given for calculation of the RQL (which is also needed for sequential sampling plans and narrow limit gauging plans).
- Generation of the operating characteristic curves for the original ANSI/ASQ Z1.4 plan and the c=0 plan
- The c=0 plan will have the smallest average sample number (ASN) as compared to ANSI/ASQ Z1.4 single, double, and multiple sampling plans, as well as sequential sampling plans. The tradeoff is a much higher producer's risk of rejection at the AQL. This could invoke ANSI/ASQ Z1.4's switching rules (if in use) to require tightened inspection and even 100% inspection.
- When quality is not sufficiently good to support zero acceptance sampling, alternatives are available.
- ANSI/ASQ Z1.4 double and multiple sampling plans are designed to reject very bad lots very quickly, and also accept extremely good lots very quickly. This reduces the average sample number. The presentation will give a brief overview of how this work.
- Sequential sampling plans have even lower average sample numbers. A brief overview will be provided of how they work, although the specific mechanics are beyond the scope of this 1-hour webinar.
- Narrow limit gauging offers enormous reductions in sample sizes but has the prerequisites that (1) the quality characteristic be a real number that follows the normal or bell curve distribution, (2) the nonconforming fraction is solely a function of shifts in the process mean, and (3) go/no-go gages can be set to specific real number dimensions.
Reduced inspection under ANSI/ASQ Z1.4 requires two decisions; whether to accept the lot, and whether to remain with reduced inspection or switch back to normal inspection. The simplest solution is probably to stay with the traditional reduced inspection plan to avoid administrative complexity, noting that the required sample size is already smaller than under normal inspection.
Who Should Attend
Quality managers, engineers, and technicians, and others with responsibility for acceptance sampling activities
