1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
Dismiss Notice
You must be a registered member in order to post messages and view/download attached files in this forum.
Click here to register.

Simple Explanation for AQL VS LTPD

Discussion in 'Sampling, Standards and Inspection' started by Pongsakorn, Jul 7, 2016.

  1. Pongsakorn

    Pongsakorn Active Member

    Joined:
    Aug 10, 2015
    Messages:
    72
    Likes Received:
    16
    Trophy Points:
    7
    I use LTPD Sampling plan for the Incoming Material Inspection and use AQL Sampling plan for Outgoing and in-process in accordance with sampling plan theory.
    However, customer always ask why LTPD is used at Incoming Inspection instead of AQL. Customer does not like LTPD since the sampling plan is fixed regardless lot size unlike AQL.

    Please advise what should be the "Simple" explanation to make customer understand clearly.
     
    Pongsakorn, Jul 7, 2016
    #1
  2. Bev D

    Bev D Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    633
    Likes Received:
    686
    Trophy Points:
    92
    Location:
    Maine
    First remember that there are TWO defect level 'anchors' in a sampling plan: the AQL (Acceptable Quality Level) and the RQL (Rejectable Quality Level). The AQL is the maximum defect rate that is ACCEPTABLE to ship to your Customer. The RQL is the minimum Defect rate that not acceptable to ship; in other words it is the rate that you want to detect and reject. A LTPD plan is an RQL based plan. LTPD stands for Lot Tolerance Percent Defective and is generally the defect rate that will be rejected 90% of the time.

    Frankly your Customer is fairly ignorant of the details of acceptance sampling if they are asking you this question. They are stuck in 'check the box' mode. If they did understand acceptance sampling they would be asking you to perform an RQL sample plan at outgoing inspection

    The whole change in sample size vs. lot size is a result of negotiations a long time ago (1950s) and were biased to not over burden industry. There is very little statistical validity to it. Unless the lot size is very small relative to the sample size there is no need to adjust the sample size based on the lot. The ability of any sample size to 'accurately' represent the lot is based on the defect rate in the lot and the homogeneity of the defects in the lot.

    Check out my resource "Inspection Sampling Plan for Categorical Data - RQL and AQL Based Plans" in the Resources section of this forum...
     
    Bev D, Jul 7, 2016
    #2
    Marc Smith, Pongsakorn and Atul Khandekar like this.
  3. Pongsakorn

    Pongsakorn Active Member

    Joined:
    Aug 10, 2015
    Messages:
    72
    Likes Received:
    16
    Trophy Points:
    7
    Hello Bev, really thanks for your reply. I will take time to study on what you posted as well as the one that I have downloaded from Resources Section.
    Thanks again.
     
    Pongsakorn, Jul 8, 2016
    #3
  4. Pongsakorn

    Pongsakorn Active Member

    Joined:
    Aug 10, 2015
    Messages:
    72
    Likes Received:
    16
    Trophy Points:
    7
    Hello Bev,
    By the way, please also help to give some simple explanation on "How to Determine LTPD sampling plan", the Control Plan that I use was established long time ago by the person who is no longer with the company.
    The sampling plan is 20% LTPD, Ac/Re=0/1, I do not know how was it determined. I think it might be "Rule of Thumb" or the sampling plan determined to be suitable for inspection capacity (manpower).
    So, I need to really know how LTPD sampling plan is theoretically determined since customer always ask and I have no idea how to provide proper answer.
    For AQL at Outgoing Inspection, I have no problem since it is the agreement made between customer and my organization for Outgoing Quality Level. Usually the AQL 0.065%, Ac/Re-0/1 is used.
    However, I still need to know why customer determined this AQL. Please also advise if you have any comment.

    Thanks.
     
    Pongsakorn, Jul 9, 2016
    #4
  5. Bev D

    Bev D Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    633
    Likes Received:
    686
    Trophy Points:
    92
    Location:
    Maine
    LTPD plans work the same way other plans do. For LTPD, remember that the selected defect rate is the one that you want to REJECT if it is present. You also select the probability of detecting that defect rate. (LTPD is a specific term for a 90% probability of detecting and therefore rejecting the rate. but you can select any probability you want) I typically use 95%...the spreadsheet I referenced above will take your input and calculate the sample size and accept reject numbers. one of the tabs is a simple c=0 plan for the "LTPD". Note that this spreadsheet doesnt' adjust for lot size as that isn't really necessary unless you have really small lot sizes...
     
    Bev D, Jul 11, 2016
    #5
  6. Ravi Khare

    Ravi Khare Member

    Joined:
    Jul 30, 2015
    Messages:
    24
    Likes Received:
    27
    Trophy Points:
    12
    Location:
    Pune, India
    Like Bev said above, it's really in the interest of the customer that you follow the LTPD sampling plan. The AQL based sampling plans aim to contain the error of incorrect rejection of a lot deemed good (producer's risk) whereas the LTPD (or RQL) based sampling plans contain the risk of incorrect acceptance of a lot deemed not good. If you explain to the customer that you are indeed following a plan that will contain their risk, they should be happier with the LTPD plan than the AQL one.
    To take the discussion a little deeper and to talk about the current status on sampling plans standards, the updated ANSI Z standard for sampling has an interesting take on the definition of AQL. It used to be the acronym for 'Acceptable Quality Level', implying the maximum defect rate that you can tolerate in the lots that you ship to the customer. If you ask any Quality Manager or buyer in plain English as to what level of defects would be deemed an acceptable level, the natural answer would be zero. Under this circumstances 100% inspection would be the only solution ( this too subject to measurement system errors) and statistical sampling would be impossible.
    The ANSI Z standard therefore now defines AQL as 'Acceptance Quality Limit'. It is defined as that maximum defect rate that is accepted by the sampling plan most of the times. The 'most of the times' is also quantified by the confidence level which is often 95%, implying that the lot with the AQL level fraction defectives would be accepted by the sampling plan 95% of the times.
    All said, the sampling plans still are based on the AQL (containment of producer's risk) rather than the LTPD (containment of the customer's risk). IMHO they really ought to be the other way around.
     
    Ravi Khare, Jul 12, 2016
    #6
    Bev D likes this.
  7. CARLOTA

    CARLOTA New Member

    Joined:
    Nov 4, 2017
    Messages:
    1
    Likes Received:
    0
    Trophy Points:
    1
    GOOD DAY

    am new in AQL and our company developing a quality manual. our product is plastic polybag. packed in a bundles, ex. 200,000pcs (200 bundles@ 1000pc) what is the lot size 200,000pcs or 200bundles?
     
    CARLOTA, Nov 4, 2017
    #7
  8. Bev D

    Bev D Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    633
    Likes Received:
    686
    Trophy Points:
    92
    Location:
    Maine
    You could use the number of bundles IF you inspect every bag in the bundle...and so your sample size is number of bundles.
    Otherwise your lot size is number of bags AND your sample size is the number of bags you will inspect. The difficulty is that you will have to randomly select the bags from among all of the bundles...
     
    Bev D, Nov 5, 2017
    #8