I have made this letter longer than usual, only because I have not had the time to make it shorter.
- Acceptable Quality Level
The maximal percent of nonconforming items (or the maximal number of nonconformities per 100 items), which is considered, for inspection purposes, as a satisfying process mean.
The AQL is generally specified by the authority responsible of sampling. Different AQLs may be designated for different types of defects. It is common to use an AQL of 1% for major defects, and 2.5% for minor defects.
Values of AQL that are 10% or less are suitable for percent nonconforming or nonconformities per 100 items. Values of AQL over 10% are only suitable for nonconformities per 100 items.
- Acceptance Limit
The upper limit on the number of non-conforming items in a sample, that would still lead to the acceptance of the entire lot. If the number of non-conforming items in the sample exceeds this number, the entire batch must not be accepted.
- Alarm Zones
The zones in an ordinary 3-sigma control chart are beyond the upper control limit (3,∞), and below the lower control limit (-∞, -3).
- To specify the alarm zone as the area between the warning limits and control limits, enter a=2, b=3.
- In a control chart with 0.001 probability control limits (3.09 "sigma") and 0.025 warning limits (2.24 "sigma"), the rule "two consecutive points between the control and warning limits" is given by: k=2, a=2.24, b=3.09.
- The zones are usually symmetric around the center line. For example: [-3,-2] and [2,3].
- Average Outgoing Quality Limit
The highest/worst possible average percent of non-conforming items in the process, after employing some inspection scheme. This measure is usually used in rectifying inspection, where the inspection procedure changes the outgoing rate of non-conforming items in the batch or process, relative to the incoming rate. For example, by removing the non-conforming items that are encountered during inspection. Note that this is only the worse possible average percent of non-conforming items, and therefore there is still a possibility that the percent non-conforming of a single batch will exceed this limit.
- Average Run Length
The mean (average) of the run length. This is the average number of samples that are taken until an alarm is signaled by the control chart.
A batch is a collection of items from which a sample will be drawn, for deciding on its conformance to the acceptance inspection. A batch should include items of the same type, size, etc. and that were produced under the same production conditions and time.
The batch size is the number of items in a lot or a batch.
- Clearance Number
The number of consecutive items (or batches, in Skip lot sampling) that must be found conforming, in order to quit the screening phase (100% inspection) when applying continuous sampling.
- Consumer Risk
The chance that a batch with quality LTPD or worse is accepted.
- Inspection Levels for Military Standard 105E (MIL-STD-105E)
Levels I, II, and III are general inspection levels:
- Level II is designated as normal.
- Level I requires about half the amount of inspection as level II, and is used when reduced sampling cost are required and a lower level of discrimination (or power) can be tolerated.
- Level III requires about twice the amount of inspection as level II, and is used when more discrimination (or power) is needed.
The four special inspection levels S-1,S-2,S-3,S-4 use very small samples, and should be employed when small sample sizes are necessary, and when large sampling risks can be tolerated.
- Inspection Levels for Military Standard 414 (MIL-STD-414)
Levels I, II, III, IV, V are general inspection levels:
- Level IV is designated as normal.
- Level V requires a larger amount of inspection than level IV, and is used when more discrimination (or power) is needed.
- Levels I, II, III require less inspection than level IV, and are used when reduced sampling costs are required and a lower level of discrimination (or power) can be tolerated.
- Lot Tolerance Percent Defective
The consumer’s required per-batch quality.
- Maximal Run Length Value
The largest number on the horizontal axis, in the run length plot. Or, the largest value of t on the plot for which P(RL=t) is plotted. For example, selecting 500 will give a probability plot of run-lengths in the range 1,2,...,500.
- Nonconforming Items
The nonconformity of an item is expressed as the percent of nonconforming items. When each item can contain more than one defect, the nonconformity of an item is expressed as the number of non-conformities (defects) per 100 items.
- Percent Non-Conforming
The percent or proportion of non-conforming items in a batch or in a process. In many cases this is unknown, but it is used to learn about scenarios for different values of p.
- Producer’s Risk
The chance that a batch from a process with quality AQL is rejected.
- Production Interval
A period of production under continuous sampling assumed to consist of essentially homogeneous quality. It is normally a single shift. It can be a day if it is reasonably certain that shift changes do not affect quality of product, but shall not be longer than a day.
- Rejection Limit
The smallest number of non-conforming items in a sample that would lead to the rejection of the entire lot. In most cases (besides reduced sampling) this value is equal to the acceptance limit -1.
- Run Length
The run length is the number of samples taken until an alarm is signaled by the control chart.
- Sample Size
The number of items that should be randomly chosen from a batch.
- Sampling Fraction
The proportion of items (or batches, in Skip lot sampling) that are inspected during some phase, when applying continuous sampling. f is between 0 and 1. There are three ways to sample with a fraction of f:
- Probability Sampling: Each item/batch is sampled with probability f.
- Systematic Sampling: Every 1/f 'th item/batch is sampled. 1/f must then be a natural number (e.g., every 3rd item is inspected, when f=1/3).
- Block-Random Sampling: From each 1/f consecutive items/batches, one is chosen at random. 1/f must then be a natural number (e.g., in each block of 3 items one is chosen, whenf=1/3).
- Shift Size
The purpose of using a control chart is to detect a shift in the process mean, of a specific size. To detect a shift of two standard-deviations-of-the mean, enter the value 2.
Statistical Process Control. See here.
- Type of Inspection
There are three types of inspection:
- Normal inspection is used at the start of the inspection activity.
- Tightened inspection is used when the vendor's recent quality history has deteriorated (acceptance criteria are more stringent than under normal inspection).
- Reduced inspection is used when the vendor's recent quality history has been exceptionally good (sample sizes are usually smaller than under normal inspection).
- Verification level
Prescribes the level of significance or utility of a characteristic to the user. The amount of effort to assure conformance can be allocated on the basis of importance to the user. (Major characteristics will require more verification effort than minor characteristics.) VL-VII requires the highest level of effort, and the effort decreases as the VL decreases to the lowest level, VL-I