iSixSigma

Capability Analysis In Receiving Inspection

Six Sigma – iSixSigma Forums Old Forums General Capability Analysis In Receiving Inspection

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • #41050

    Tierradentro
    Participant

    Dear All,
    I would like a few opinions on this question –
    Is it reasonable and customary to calculate a Cpk based on data generated in a receiving inspection operation?  If so, is it reasonable and customary to request a corrective action when all samples conform to specifications, but the calculated Cpk is < than specified minimums?
    Please assume that the gage R&R is acceptable before responding.
    Thanks in advance for your opinions.
    Yours Truly,John

    0
    #128965

    hzypbh
    Participant

    Very good question(s).  The first answer is maybe.  This depends on how close the parts were to the specification limits.  As you may or may not know, the Cpk number your expecting is likely more limiting to the part acceptability than the spec.  Ex.  Say your part spec is +/- 5 inches and your desired Cpk is 2.0, thus you expect that your parts typically fall between +/- 2.5 inches.  This does not mean that they are out of spec when they approach +/- 5, it just means that they are not what you expect from your process. 
    Also keep in mind that Cpk is sometimes referred to “short term” or “within” subgroup.  Therefore, if the process is sampled over the long haul and your gathering between shift, between days, etc…  You may wish to use a “long term” “between” measure such as Ppk.  Typically, the sampling of your process dictates the measure you’ll use.
    Hope this helps

    0
    #129033

    Huzi
    Participant

    Between Supplier and Customer, the customer sets the spec limits and accepts everything within the spec band. If you ‘specify’ that the supplier needs to be between +-5 mm AND have a capability greater than 2.0, then you are effectively telling him to keep the specs at +-1.2 mm, although +-5 mm will not hinder your product build. It would be in good interest to you and your supplier if he starts off with controlling his process between +-1.2 (if he can and if it is cost effective). There is a point where the goal of achieving capability AND spec together crosses the ‘on the floor’ capability of his machines and processes. A good ‘plant floor + stats guru’ guy can figure out the optimum balance without loosing money on ‘too much quality’.
    I work in a automotive stamping plant and I am a Dimensional control Engineer. I am facing similar problems. I have been given measurables for sheet metal stamping to be within spec of +-0.7 mm and also achieve 1.33 capability. That means, I have to control my process between +-0.28 mm. With 30 year old presses and incoming material having variation that is out of my hands to control, I physically find it difficult and costly to be within those limits. Now, the question is, does a +-1.5 mils tolerance bother the customer? Nope. They are fine with it. But some nerd on a leather chair, who didnt really see the reality on the plant floor and got the wrong customer satisfaction metrics drove the tight tolerances and capability justification in. Waste of money.
    Huzi
     
     

    0
    #129385

    jayant
    Participant

    Dear Friend ,
    Samples confirming to your specification is not a sign of a good process capability , as you confirmed the required CPk is less then the expectations , it means lot of variation in the results obtained showing no control over the process and is tend to go out at any moment due to arise of any assignable cause .
    Thus asking corrective measure to limit the variation will ensure lesser process variation and better process capability.

    0
    #166755

    Matson
    Participant

    Careful on this approach.  Not knowing the order of production can greatly mislead your capability intentions.  The supplier may provide product from a process which they 100% sort.  So, the CpK can be misleading.
    One approch would be to give the supplier a protocol to follow where they select consecutive sample sub-groups over a period of manufacturing [without sorting].  This can give you a good view of the supplier process.  The sorting process is another story.

    0
    #166774

    H. Jones
    Participant

    John,
    Firstly forget about the myths of Gage R&R :
    http://www.spcpress.com/pdf/good_data_%20bad_data.pdf
    Cpk can be very misleading.  Use control charts.
     

    0
Viewing 6 posts - 1 through 6 (of 6 total)

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