Sampling by Variables (ANSI/ASQ Z1.9, formerly MIL-STD 414)
-
Speaker : WILLIAM LEVINSON
-
When : Wednesday, May 21, 2025
-
Time : 01 : 00 PM EST
-
Add To Calendar
Refer a Friend
William A. Levinson, P.E., FASQ, CFPIM is the principal of Levinson Productivity Systems, P.C. and the author of several books on quality, productivity, and management.
ANSI/ASQ Z1.9 (formerly MIL-STD 414) is an acceptance sampling plan that uses the average measurement of a sample, and its range or standard deviation, to determine whether a production lot should be accepted. This webinar will show how to select and use sample plans.
Attendees will receive a pdf copy of the slides and accompanying notes, and an Excel spreadsheet that converts a traditional sampling plan into a sequential plan.
Learning Objectives
1. ANSI/ASQ Z1.9 is an acceptance sampling plan for variables (real number data). The underlying principle is that, given an estimate of the lot's mean and standard deviation, we can estimate the nonconforming fraction in the lot from the sample statistics (average and standard deviation or range—in general, the standard deviation method is better).
- The general principle is that more distance between the sample average and the specification limits, with distance being measured in sample standard deviations, is better. The standard quantifies this with acceptance and rejection criteria.
2. A prerequisite for this plan is that the quality characteristic follows a normal (bell curve) distribution. If it doesn't, the underlying assumptions are invalidated and we cannot be confident in the results.
- Production lots should ideally come from identical conditions (time, materials, environment, and procedures) as opposed to having items drawn from production under differing conditions.
3. Switching rules for normal, tightened, and reduced inspection. The specific details appear in the plan and will be described in the webinar. In general,
- If too many lots are rejected, we must switch to a tightened (more stringent) sampling plan. A specified number of lots must then be accepted to go back to normal inspection.
- If enough consecutive lots are accepted under normal inspection, we can go to reduced (less stringent) inspection; rejection of a lot means we have to go back to normal inspection.
4. Selection of the sampling plan, including the sample size
- Is the sample variation known or unknown? Different systems are available for both situations.
- Is the specification one-sided or two-sided? Both can be accommodated by ANSI/ASQ Z1.9.
- Use the lot size, inspection level, and acceptable quality level (AQL) to select the sampling plan.
5. Calculate the quality index from the sample statistics (larger is better) to determine whether the lot meets the acceptance criteria. We can also estimate the nonconforming fraction (estimated lot percent defective) from the sample statistics. This is for a one-sided specification limit.
- For a two-sided limit, we must estimate the nonconforming fraction at each limit and also the total nonconforming fraction (the sum of the two former quantities).
6. Awareness of the sample range method, which uses the range rather than the standard deviation to estimate the variation.
- The standard deviation method is, however, better. The range method might date back to the days when hand calculators and computers were not available, and the inspector on the shop floor could save a lot of time by just subtracting the smallest measurement from the largest.
7. If the variation is known (e.g., from long experience with the process), the plan is simplified considerably, and smaller samples can be used. The operating characteristic (OC) curve, which shows the chance of acceptance as a function of the nonconforming fraction, when the variation is known.
This webinar will also present some of the rationale behind the procedure, and this background does not appear in the standard itself. An appendix will provide some of the underlying details for those who are interested.
Who Should Attend
Quality managers, engineers, technicians, and others with responsibility for acceptance sampling activities
-
$199.00
-
