Conversion of AQL to Sigma

Six Sigma – iSixSigma Forums Old Forums General Conversion of AQL to Sigma

Viewing 13 posts - 1 through 13 (of 13 total)
• Author
Posts
• #27018

sreedher
Member

Hello Tom, and others

Consider the following,

As per AQL 2.5 acceptance sampling system which is supplied by our customer,

Lot size Inspect Accept
501 – 1200 32 2

Let me assume that there exists 106 opportunities for a unit to deviate, the unit being a Basic shirt which you wear.

so total opportunities for 32 units being inspected would be 3392 of which he would accept 2 only. In which case it would be 589.62 in million opportunities. This indicated towards, a Z-bench of 4.75 thereby expecting a sigma of more than 4.75 for the process. Now this varies as the acceptance number & inspection number varies. And thiese numbers are dependent upon Lot size or quantity being audited. Is my analogy O.K.

Further more the Lot size are usually expressed as range values one of which is expressed above. This demands a little depth into the analogy.

Since lot of conventional quality systems are still running with Accepted Quality Level kind of systems, this is a question needs to be resolved first.

Hope to hear from you and others

sreedher

0
#65709

Ropp
Participant

You should understand how an acceptance sampling plan works.
When you take a sample of 32 items and find two defectives , this says that there is no reason to believe the total defectives in the lot EXCEEDs 2.5% of the whole, but it may well equal it. Now, if we do find two or fewer defects, we can only say that the total defective in the lot is probably less than 2.5% or 25,000 PPM.
The other end of the OC curve is called the LTPD or Lot Tolerent Percent Defective. This is the percent defective that wil be rejected, i.e. > two found, 90% of the time. For this plan that is around 15 % defective.
Overall, this is probably somewhere around 3 sigma performance and is typical of the performance of this sort of plan.
To find true process performance, take a large enough sample for long enough and count the number found; or “p” chart it.

0
#65711

sreedher
Member

That is true. I will make my problem a little more clear,

Our customer accepts the goods we supply at AQL 2.5, now that I am analyzing the situation based upon DPMO, on an average there are 106 opportunities in an unit to go wrong. Your guess is right, we are doing 100% inspection at the end of assembly line for discrete parameters while sampling inspection for variables.
Because of availability of high opportunities I am getting a larger value for sigma. For example

# inspected = 75
# passed = 60
# failed = 15 (deviations in 15 opportunities)

Total number of opportunities available = 106 x 75
Defects per opportunities = 0.001887
Defects per million opportunities = 4.37

So if I consider that my customer say demands somewhere around 3 sigma, our process is performing at 4.37 sigma.

Say from the same kind of a lot our customer picks up 2 deviations he orders for rescreen. This is where I am puzzled.

On the other hand what seems to me is, 15 failed out of 75 directly gives the process rate as 2.33 sigma, this has more relevance to practical applications which I am currently using.

If I consider the AQL table supplied by the customer on the basis of accept and reject numbers converted to opportunities, the expectations come to an average of 5.5 sigma.

This value has a practical relevance to my above said value of 4.37. Since I am very well aware that we are not up to the mark with customer expectations about quality these were my conclusions, tell me if you agree upon the given details.

0
#85772

ithrees
Participant

pls send the details,sir

0
#101103

shahidiqbal
Member

hi plz send me material on siz sigma

0
#101104

shahidiqbal
Member

hi plz send me material on siz sigma

0
#107967

Ejaz
Participant

Plz sir send to me the detail.
thank u

0
#107968

Ejaz
Participant

Plz sir
Send me the detail.
thank u

0
#131790

sopia
Member

Hi Sreedhar & others,
As per my opine, Lot size Inspect Accept 501 – 1200 32 2
above will have a sigma of 3.03, as 2 are the defectives and not defects in this case.
correct me if iam wrong. In sigma solutions we cannot have assumptions on the number of defect per unit.
Sophia

0
#145459

Member

Hi,
pls suggest me
I have working in garment industry from 1999to2006.According to my knowledage if lot size 501-1200 than check the qty 800 as per 2.5,acceptable defect 5pcs
if lot size is 1201to3200 as per 2.5 standard check the qty125,accept  7pcs.
if lot size is 3201to10000 as per 2.5 standard check the qty200,accept 10pcs

0
#156703

Scott
Member

from the customers standpoint you have exactly one opportunity per unit for a defect to occur as the customer looks at the complete product.
How many opportunities you have in your process is irrelevant.
So an AQL of 2.5 is really about a 2 sigma level

0
#158365

yoganandam.N
Member

Dear sir,
tell me how can we select the carton sample
for example lot is   5000 pcs
No.of cartons is 100
as per AQL, I have to select  200 pcs, how  i can pull the 200 pcs from how many carton
explain me
thanks
yoga

0
#177536

Idalia Valdez
Participant

Would this mean that if I have an AQL of 0.01% with 0 defectives that my lot would be 100 PPM?
If not, how could I convert my level of acceptance 0.01% into PPM?

0
Viewing 13 posts - 1 through 13 (of 13 total)

The forum ‘General’ is closed to new topics and replies.