Sampling Raw Materials using ANSI/ASQ Z1.4

[JanW2406]

Hello,

I have a question about the use of the AQL tables. We receive raw materials into our facility and I want to be able to use the AQL Tables to choose the quantity that need to be sampled/tested. The problem I have is when we go to actual sample and test product from the previously chosen buckets of raw materials; they are tested for micro, assay and sometimes other requirements. How can I establish Accept/Reject levels when a) there are many tests that are conducted for each sample and b)in general, do Accept/Reject levels apply if we are doing testing?

 

[Mike S.]

Not sure if I understand your question, but I'll give it a try. Only you and the experts within your company, possibly with customer input, can determine an acceptable sampling method based on risk analysis, costs, etc.

Lets say you have a bucket of 500 pieces of steel. Your sampling plan that you carefully selected specifies the sample size is 20, Ac = 2 Re = 3. You test each sample for micro, assay, and other. Lets say all pass except there are 3 micro failures. The lot is rejected. You could decide to 100% the lot for micro or disposition in some other appropriate way (scrap, use-as-is, rework, etc.)

 

[Bev D]

It really depends on what you and your Customer need (although frequently your Customer might not be concerned with incoming inspection, only release testing of finished material/component.)

remember how the probability works to make your choice. Let’s take Mike S’s example: n=20, Ac=2, Re=3
1. Characteristic A has zero failures, Characteristic B has 1 failure and Characteristic C has 2 failures. Now the lot passes as each characteristic (treated independently) passes. Btu remember that the lot itself fails for the total number of failures. Is that OK? Only you can answer that.
2. If the different characteristics have different severities, (in terms of cost, probability to escape to customer and severity of failure at Customer, etc.) then you are probably better off treating each characteristic separately. If they all have the same or similar severity then you are probably better off treating them all together…

 

[Steve Prevette]

The standard does allow for minor, major, and critical defects, each of which have different rejection criteria and do that all with the same sample.

 

[JanW2406]

Not sure if I understand your question, but I'll give it a try. Only you and the experts within your company, possibly with customer input, can determine an acceptable sampling method based on risk analysis, costs, etc.

Lets say you have a bucket of 500 pieces of steel. Your sampling plan that you carefully selected specifies the sample size is 20, Ac = 2 Re = 3. You test each sample for micro, assay, and other. Lets say all pass except there are 3 micro failures. The lot is rejected. You could decide to 100% the lot for micro or disposition in some other appropriate way (scrap, use-as-is, rework, etc.)

Thanks, but to clarify each sample that is sampled/tested would fail if anyone of the attributes (micro, assay, etc) failed, so how do I apply the Accept/Reject for each sample sampled/tasted.

 

[JanW2406]

The standard does allow for minor, major, and critical defects, each of which have different rejection criteria and do that all with the same sample.

Thanks, but to clarify each sample that is sampled/tested would fail if anyone of the attributes (micro, assay, etc) failed, so how do I apply the Accept/Reject for each sample sampled/tasted.

The standard does allow for minor, major, and critical defects, each of which have different rejection criteria and do that all with the same sample.

Thanks, but to clarify each sample that is sampled/tested would fail if anyone of the attributes (micro, assay, etc) failed, so how do I apply the Accept/Reject for each sample sampled/tasted.

 

[Steve Prevette]

You can set up the reject criteria (based upon the associated AQL) such that, as an example, a given sample may fail on one critical defect, two major defects or four minor defects.