I need Statistical KNOWLEDGE assitance and FAST! 4

[Sheep]

Hello all;
I was just posed with what may seem a riduculous question (and it is) but I am required to arrive at an answer THIS MORNING! :O
I on had a lot of (new product, no historical data available) of 3600 units and one was to inspect but 6 of those unit and they were ffound to be acceptable, to WHAT AQL would one working? This one is so far off the scales that I can't even BEGIN to imagine! Is this even a viable question? Please keep in mind that I am currently working in a totally out-of-control company. Both the answer AND the formula from which it is derived would be GREATLY appreciated! Thanks in advance.
Sheep

 

[Sheep]

Hello all;
I was just posed with what may seem a riduculous question (and it is) but I am required to arrive at an answer THIS MORNING! :O
If on had a lot of (new product, no historical data available) of 3600 units and one was to inspect but 6 of those unit and they were found to be acceptable, to WHAT AQL would one working? This one is so far off the scales that I can't even BEGIN to imagine! Is this even a viable question? Please keep in mind that I am currently working in a totally out-of-control company. Both the answer AND the formula from which it is derived would be GREATLY appreciated! Thanks in advance. I apologize for the typos earlier, I had 5 people talking at once as I was trying to compose the question. HELP, PLEASE!
Sheep

 

[dingy2]

Sheep
AQL's are not used in automotive but at one time, we used to use MIL 105 D for attributes and not variables in a non-automotive atmosphere.

The standard has sampling plans but I really don't know if they are used anymore. I thought it was quite good but it was only used in a company that imported product from Taiwan using AQLs.

Good luck!

Dave D.

http://www.qmsi.ca

 

[Sheep]

Thanks! I really DO appreciate the response. Perhaps you didn't quite get my drift however; the problem is that this number (sampling number) was selected by my supposed "Director of Quality" and it is SO far off the 105 (or any other) sampling tables that it is ludicrous. I was really just wondering if anyone knew HOW to figues out what the AQL would BE, given this number (6 of 3600) I meely extrapolated it mathmatically and came out with something near 21) Thanks again!

 

[eddiewdi]

hey i might be alittle too late in response buuttt
acording to ISO 2859-1:1999
the answer to your question of AQL is 2.5
page 20 table 2-A lot size 3201 to 10000
with S-1 inspection level sample size of 5 AQL 2.5
Accept 0 findings reject lot with 1 finding
or General inspection level 1 sample size j lot size 80 AQL 2.5 accept lot with 5 findings reject lot on 6 findings. Personnely I would inspect 6 instead of 80

NOW THE REAL POINT IS YOU AS THE MANUFACTURE, SUPPLIER OR
CUSTOMER ESTABLISH THE AQL FOR THE PRODUCT AND THE TABLES GIVE US GUIDELINES ON HOW TO SAMPLE. (did not mean to shout) AQL means the quanity of non-conforming parts you are willing to accept based on 100 lot. in this case 2.5 out of 100

 

[AmeristarQA]

There is also C=0 sampling plan that was derived by a guy named Squeglia (not sure if thats spelled right), which is based on the Mil-105. His tables show different sample sizes based on the lot size, and the Certainty you are looking to achieve. I think this is what eddiewdi is refering to.

 

[kenvlach]

ISO 2859-1:1999 is probably based upon MIL-STD-105E whose tables have exactly the same values given above by eddiewdi.

http://assist.daps.dla.mil/quicksearch/basic_profile.cfm?ident_number=35496

I've never heard of anyone using such a low quality level as AQL 2.5. The anodizing specification MIL-A-8625F has a default AQL of 1.5, with sampling per Inspection Level II. However, neither the sampling size or AQL are acceptable for most customers. 100% inspection is typical for the visual inspection while a reduced sampling size is typical for measuring coating thickness. As MIL-STD-105E is unintelligible to most workers (illegible too if one only has the pdf download), I derived the following table:

SAMPLING AND ACCEPTANCE TABLE BASED UPON MIL-STD-105E FOR INSPECTION LEVEL II WITH ACCEPTABLE QUALITY LEVEL (AQL) 1.5

[tt]LOT SIZE SAMPLE SIZE ACCEPTANCE NO. REJECTION NO.
________ ____________ ___________ ______________
2 to 8 2 0 1
9 to 15 3 0 1
16 to 25 5 0 1
26 to 50 8 0 1
51 to 90 13 0 1
91 to 150 20 1 2
151 to 280 32 1 2
281 to 500 50 1 2
501 to 1200 80 2 3
1201 to 3200 125 3 4
3201 to 10000 200 5 6
10001 to 35000 315 8 9
35001 to 150000 500 12 13
150001 to 500000 800 18 19
500001 and over 1250 18 19[/tt]
__________________________________________________________

Comments:
This Table gives the sampling and acceptance/rejection criteria of MIL-STD-105E for normal sampling (Inspection Level II) with AQL 1.5.
It is permissible to examine a greater number of parts, but not fewer.
Regardless, customer inspection requirements must be met. Some customers require 100 % parts inspection.