AQL (Acceptable Quality Level) for Safety Critical Parts

[StorkSPA]

One of our suppliers wants to use a sampling plan for a critical part. He suggested an AQL of 2.5%. The production batch size is 30.

It seems to me that, for a safety critical part, an AQL of 2.5% is pretty high.

What would be an acceptable AQL for safety critical parts?

 

[Bev D]

Re: AQL for critical parts

zero

as long as the test isn't destructive.
and the safety issue is serious (not just a hang nail)


seriously the AQL should be very low but at your lot sizes that will translate to 100% pretty quickly.

 

[Sturmkind]

I think it would depend on why they want to use a sampling plan. If this is for final product or dock audit, AQL 2.5 may be fine.

Otherwise, the critical characteristics (non-destructive as Bev said) should be either 100% inspected or process error-proofed/in-line inspected at 100% to ensure that non-conformances are never shipped.

 

[Kales Veggie]

One of out suppliers wants to use a sampling plan for a critical part. He suggested an AQL of 2.5%. The production batch size is 30.

It seems to me that, for a safety critical part, an AQL of 2.5% is pretty high.

What would be an acceptable AQL for safety critical parts?

For 30 pieces, I would do 100% inspection. It is all about risk. Risk is high.

What to inspect is even more important.

 

[StorkSPA]

The part is critical for a landing gear, and not the final product but a mono part. It contains about 70 dimensions (of which I'm sure not all are functionally critical).

So I can see why the supplier does not want (unless we pay 15% more) to measure 100%.

My suggestion would be: critical dimensions 100%, others by sampling.

But then, the question remains: how to determine the right AQL? (must be based on more than "my opinion is .15; 2.5; 1.5")

I dug into internet sources for hours now, but can't find a specific answer or simple tutorial. But I am certain I wil need the answer for this and other products.

 

[StorkSPA]

As an example, I know surgical rubber gloves have AQL 1.5 (government requirement). And that, if you cut off 4 fingers, the AQL must be 0.4.
For both products the critical function is to create a barrier to fluids.

 

[Stijloor]

One of our suppliers wants to use a sampling plan for a critical part. He suggested an AQL of 2.5%. The production batch size is 30.

It seems to me that, for a safety critical part, an AQL of 2.5% is pretty high.

What would be an acceptable AQL for safety critical parts?

What was agreed upon during the contract review phase? I would assume that because of the criticality of the part, appropriate verification methods would be discussed and agreed upon prior to acceptance of the order. Too many times, this becomes a serious issue between customer and supplier.

Stijloor.

 

[StorkSPA]

What was agreed upon during the contract review phase? I would assume that because of the criticality of the part, appropriate verification methods would be discussed and agreed upon prior to acceptance of the order. Too many times, this becomes a serious issue between customer and supplier.

This issue came up when we were examining the suppliers MPS - which is a requirement on the order before start of production. But the finding did kick the price up and I assume this does not make my procurement collegue happy .

So yeah this is an issue we have to embed in our quality requirements. Same is true for process capability - which must be known before sampling is used - and leaves the supplier no choice but to measure 100% for the first batch(es).

 

[Bev D]

Instead of using an AQL, you might consider using SPC for the non critical characteristics.

Depending on the process and how you might subgroup, you could always just reduce the sampel size down to 1-3 units and have the supplier give you the control charts.

 

[StorkSPA]

It's awfully quiet with responses to this question. Is really no one experienced in determining AQL levels?