AQL

HI Every Body,
sorry because i have lot of questions

what is the reference for the AQL tables we see on the web, and how to use them ?


thanks in advance

 

They refer to an "Acceptable Quality Level" for sampling plans originally as part of MIL-STD-105 (and others) but now incorporated into ANSI/ASQ Z1.4 (for attribute sampling) and Z1.9 (for variable sampling). The AQL generally refers to about what the highest amount of defects within a population will allow the lot to be regularly accepted (about 95% of the time). You need to read the standards for the specific details but generally based upon the lot size and prescribed AQL you get a sample size and accept/reject criteria used to sentence the lot.

David

 

Thanks David

 

Do you have a link for the latest issue of AQL standard ?

 

actually, i don't know, but i feel it is not clear or easy to understand when to use the Special Inspection levels and the general inspection levels. also why do we need to use AQL (0 for critical defects, 1.5 for major defects and 4.0 for minor defects).
i guess the tables are more flexible to use, don't they ??


any one has a more detailed explanation about this point ??!!

 

You can get MIL-STD-105 from the original source: http://quicksearch.dla.mil

AQLs of 1.5 for majors and 4 for minors seems pretty lax to me as well but it may be commodity/customer dependent. I've generally seen majors in the 0.25% to 0.4% range and Minors in the 0.65 to 1.0 range. As Bev says, there isn't a "0" AQL; I've usually seen "100%" be put in historically for criticals to mean inspect them all but perhaps your customer uses some other notation where "0" means something special, like 100% inspection.

David

 

Whether to use General or Special inspection levels is a balance of risk and level of effort. Whatever level is used, it is still the same AQL; the further on the Special side you go, the smaller the sample size (and smaller acceptable number of defects) and the greater level of producer and consumer risk.

I recall often seeing General-2 as the usual inspection level for inspections but for destructive/functioning tests using Special inspection levels.

Try playing with a couple examples and it will help illustrate the differences, especially if you also go and look at the corresponding O-C curves.

David

 

Another commonly used sampling for attributes is ZERO ACCEPTANCE NUMBER SAMPLING PLANS (C=0), the parameters for these plans are lot size-N, sample size-n, associated AQL ( value agreed upon between producer and the consumer). In a Zero acceptance sampling plan, the acceptance number is "0" regardless of the lot size or the sample size c=0, a zero acceptance sampling plan could be derived from a general sampling plan such as ANSI/ASQ z1.4.





THE ACCEPTANCE NUMBER IS ALWAYS "0" IN THIS PLAN. For further reading a good book reference is Zero Acceptance Number Sampling Plans, Fifth Edition by Nicholas L. Squeglia ( ASQ Press)

 

Other common C=0 (also known as Accept on Zero or AOZ) sampling plans are MIL-STD-1916 and its commercial equivalent ISO 21247. They were derived as the Department of Defense and other industries looked to head away from AQLs and the known acceptance of nonconformances without corrective action and instead move to prevention, process control, continuous improvement, corrective action, etc.

Because of the nature of AOZ plans and their OC curves they don't provide good statistical matches to their AQLs from a traditional meeting of AQLs (5% alpha/producer risk). In a "true" AQL sampling system, the plans are well matched within a specific AQL (column); increased lot sizes lead to increased sample sizes with corespondingly larger numbers of acceptable defects. In Sqeglia's plan, increasing lot sizes have larger sample sizes but the same number of allowed defects (0) so the sampling plans within an AQL become more severe.

The 1916 & 21247 standards abandoned the AQL terminology and use verification levels instead and don't attempt to tie specific inspection levels to specific quality levels.

David