Acceptable Quality Levels (AQL) - What is the usage of AQL 100 to 1000?

W

WLN

Acceptable Quality Levels (AQL)

In the sampling plan master tables there are AQLs ranging from 0.010 to 1000.
I wonder what is the usage of AQL 100 to 1000? Under what circumstances are they used?

Marc

Fully vaccinated are you?
Any Sampling Plan gurus out there?

A

Avi

I am not a guru but:
If you use percent deffective the AQL values are 0.01% to 10%.
If you use number of deffects per 100 units the AQL values are from 0.01 to 1000 defects

S

Shasta Ell

How can you have 1000 defects per 100 units? As in more than 100 defective per 100 units?

Al Rosen

Super Moderator
Shasta Ell said:
How can you have 1000 defects per 100 units? As in more than 100 defective per 100 units?
It is the number of defects not the number of defectives. Of course you can have more than one defect per unit.

N

NormBlack

WLN said:
In the sampling plan master tables there are AQLs ranging from 0.010 to 1000.
I wonder what is the usage of AQL 100 to 1000? Under what circumstances are they used?

This is not a specific answer to you question since I have no clue why AQLs of above 4 (much less 100 and above) would be used.
I found an interesting site for plugging in desired AQLs to get sample size. I have not verified it's accuracy, but you might find it of interest. It is: https://www.sqconline.com/

Norman

Govind

Super Moderator
Super Moderator
Re: Acceptable Quality Levels (AQL)

In the sampling plan master tables there are AQLs ranging from 0.010 to 1000.
I wonder what is the usage of AQL 100 to 1000? Under what circumstances are they used?

I got this same question from a student in SSBB class. “ How can AQL be greater than 100?” I am researching an appropriate answer.
Can some one from Cove provide an answer why?

Thanks,
Govind.

Tim Folkerts

Trusted Information Resource
As was pointed out earlier, a single "item" could have more than 1 defect. For example, the "item" might be a crate of apples. If it acceptable to have 2 bad apples in the crate, then AQL would be 200, i.e. two defects per item.

Or the "item" might be the paint job on an appliance. If you can live with 4 minor flaws in the paint, then the AQL would be 400.

Tim F

Last edited:

Govind

Super Moderator
Super Moderator
As was pointed out earlier, a single "item" could have more than 1 defect. For example, the "item" might be a crate of apples. If it acceptable to have 2 bad apples in the crate, then AQL would be 200, i.e. two defects per item.

Or the "item" might be the paint job on an appliance. If you can live with 4 minor flaws in the paint, then the AQL would be 400.

Tim F
Thanks.
Even the definition in Single Sample plan & Double Sampling plan says defectives. MIL STD 105 E Page 16: Sections 4.10.1.1, 4.10.1.2

The table title does not say this explicitly.
Example: Table II A, II B, II C, III A, III B, III C.

Example Table VI A Limiting Quality – Says explicitly Percent defective.

This is what got me thinking.

While the AQL % in the table from 100 to 1000% suggests otherwise.

Your explanation & previous contributor (Al) replies the question.

Regards,
Govind.

Note: I also concurrently asked this question to a Sampling expert Dr. Wayne Taylor (www.variation.com)
"The tables are for both defectives and defects. Table X is most revealing. When you look in the table at the column headers, AQLs above 10% are always labeled defects. For AQLs of 10% and below, there may be columns labeled both defectives and defects.
The tables can be used for all AQLs when reporting defects. When defectives are reported, only AQLs of 10% and below should be used".

Last edited:
T

Not sure if anyone still has questions regarding this topic, but I came across it in a Google search and thought I share some thoughts:

AQL refers to the % defective for normal inspection. That said, the lot acceptance number (or the maximum number of rejects to accept the lot) does not also represent the AQL. For example, if I have an AQL of 4.0, a sample size of 125, and a lot size of 1201 to 3200, 10 defects are allowed before the lot would get rejected. However, the ten rejects is actually 8% of the total sample size, not 4. So in summary, the AQL does not reflect the number of acceptable defect per 100.

Last edited by a moderator: