Spec AQL

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
  • #49434


    Have a customer who has provided specifications of acceptance for products we are manufacturing for them. Within the contract and specifications they indicate upto 1% acceptance of a particular defective. Within our facility we have a constant product defective rate of less than 1% and actually at 0.07% (products manufactured in a clean room environment). These products are manufactured, inspected, packaged, and sealed all within the clean room environment and never opened until received by the customer in their clean room environment. Customer continues to return batch of products when they identify one (1) contaminant within a batch of 50,000 components (contaminant is not a defective product but debris or a hair upon the product). Identified quantity of debris is well within 1% acceptance… Yet customer requests we implement procedures and equipment to improve this debris rate. Can you recommend a response as I feel this would not be cost effective and may have to increase product cost to implement such practices and equipment.



    Hi john,
                There are usually two kinds of specs comming from the customer.
    1st) Specs for the feature (this hole must be made within specs of 5mm +0-0.5).  <=== This specs will be made known to you.
    2nd) Condition for the feature (cracks on the walls?, burrs ard the edge?)  <=== This "specs" might not be made known to you BUT u will have to do something to minimise these from happening.
    In your case, what your customer mean could be ” The product is in specs but some debris is found on them. I wish that something can be done to minimise the quantity of debris”
    In my opinion, the procedure for the clean room might not be comprehensive enough. If a hair can be found on a product that has it’s whole cycle time in the environment that would mean that the procedures for the employees are not enough. (A hair that is found when every1 is supposed to be covered from head to toe.. (Where does the hair come from???))..
    A review of the procedures should solve the issue…. Only implement new procedures when the old 1 is “too broken” to be fixed…….
    Hope the above does not confuse you………



    Well understood, but I feel you miss my point that we are very much within desired specification of less than 1.00% defectives and that we are in fact running at 0.07% (seven hundreths of one percent) or 99.93% effective…
    The customer continues to write NCR’s for issues and/or findings well withing specifications. Wishing that something be done to minimize this debris is already considered and has been dealt with to a level of acceptability in accordance with their specifications.
    My question is “should they continue writing NCR’s even though we’re within requested AQL levels?” Statistics indicate that improvements to reduce to less than 1.00% is typically not cost effective, and in our particular industry, making improvements to values lower than 0.07% would cost many thousands in investments that would not provide monetary gain in tens of years to come. i.e.: to find and remove 2 specks of dust 0.015″ in diameter for every 50,000 components I find quite ludicrous and still not guaranteeable as I feel we would need a vacuum chamber as a workspace to even come close to achieving this level.



    I’m honestly surprised that nowadays there are industries that are still using the AQL system to determine the “acceptable” level of defects.
    In my experience working with customers SQE’s -Supplier Quality Engineers- I made sure that we both communicate in terms that we both clearly understand because like in your case, I was dealing with a customer that in every rejected lot they made reference to the so-called 1.6 AQL… so when I asked him, show me how did you arrive to the conclusion that 1.6 had to be used to determine the level of defectives and why 1.6? and tell me how many samples are you using to find a defective part. The response? There was no response. All he said is that that number have been used for years before he got in that company. And in regards to sample size, they were almost doing 100% inspection!
    I’m not saying that we should disagree when the customer finds a defective parts or lot that they cannot use, but to use system and terminology that allows both sides to understand the magnitude of the problem, because in many cases only 1 single defective part is needed to reject the entire lot.
    My recommendation use PPM or 6-sigma to quantify the defects -make sure to have boundaries samples for visual defects- and yes, like the customer is suggesting, look at the process again to see how can the hairs be completly eliminated from getting into the product.


    Rahul Deshmukh

    John Good Question, First things first when you say it is 1% acceptable. Do you have a defination of the defect. I mean for the debri…Boundary sample signed by customer, Acceptance size gauge, Picture, Color of debri, Location of debri.
    1% or 2 % or 650ppm etc etc does not make sence if defect definations are not correctly identified.
    The customer is always Right? You can fight the battle by ” DATA”. By showing that this particular issue is a is defect but per our definition it is an acceptable defect.
    Implemeting new automation/personnel/equipement will only increase your cost of quality. ( non value added cost)
    Rahul Deshmukh ( ASQ CSSBB)



    John,I have seen this with fastener specs and customer expectancies. Fastener flushness was within spec but being wrote up internally by inspectors for customer expectations. You have two several options:1. Change your spec using the ABC measurement system
    2. Keep your spec and ensure your customer knows what they agreed to.
    3. Clean your process
    4. Find out why the customer is unhappy and how it is affecting their process… it doesn’t matter how good your process is if they can’t use it or won’t use it. HF Chris Vallee

Viewing 6 posts - 1 through 6 (of 6 total)

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