iSixSigma

Pp and Ppk

Viewing 13 posts - 1 through 13 (of 13 total)
  • Author
    Posts
  • #43896

    Wilson
    Member

    I would like to know when to calculate Pp and Ppk ( for long term capabilities ) , do we need to remove the special causes ?

    0
    #139883

    dianel
    Participant

    Pp and Ppk can be used when porcess is not stable, so special cause is not need to remove.

    0
    #139894

    Singh
    Member

    As far a my knoledge goes, these are the indices associated with the short term capability studies. Cp & Cpk are the indices for long term capability studies.
    For Pp & Ppk  studies, we require continuously manufactured 30 ~ 50 data points. The standard deviation is calculated considering the part to part variation through RMS methodology. For Pp value, process mean is not considered, i.e. process tolerance is compared against the total process variation (in terms of 6 óobs). For Ppk value calculation, process mean is considered. These indices provide potential process capability.

    0
    #139902

    Steve SVK
    Member

    Hello
    When Im not mistaken, Pp and Ppk are used for short term capabilities – they represent a snapshot of your process. You take the sample and calculate the process capability from sample data. But basically if your process is not under control (you dont have the required standard deviation from a controled process), you use Pp or Ppk. (for example:pre production)
    I wouldnt remove special causes they could be of valuable information to you. If you remove them you are loosing information.
    Best regards.

    0
    #139935

    Wilson
    Member

    Pp and Ppk are long term capability right ? My question is fro Cp and Cpk I know special causes should be removed , but how about Pp and Ppk ?Do we really need to remove or not ?

    0
    #139937

    Haugen
    Participant

    Wilson,What are you using the data for? Short-term data gives you a directional read on how good the equipment you are running off is. We use it to determine whether to ship the equipment from the suppliers floor, or require additional work tuning in the equipment before shipping. Normally, the short-term capability requirements are tighter than the long-term. If you have enough parts and time, address the special cause(s), then run additional parts to verify. If you don’t have time/parts, then remove the data and determine if the Cp/Cpk meets the required spec with the special cause(s) removed.If you don’t have convincing analysis that shows that the data you want to remove is a special, removable cause – you should not remove the data, because you have no proof that it is special cause data.Keeping data in the analysis that you know is not reflective of the true process does nobody any good. But you MUST inform the customer any time you remove data – the customer must agree with your analysis.

    0
    #140168

    Hongkonger
    Participant

    There are two voices about Pp or Ppk. One party says it is long term (P there means Production, i.e. long term), another party says it is short term (P there means Preliminary, i.e. prototype production). Therefore, make sure you are talking on the same platform. No matter which side you stand for, special causes (assignable causes) should be eliminated to represent the inherent capability of a process. Since special causes contributed to atypical vairation in process output, counting them into Pp or Ppk does not help. Most important for you is to eliminate the special causes first, and then re-assess your process capability. That will become more meaningful. Hope that helps.

    0
    #140172

    Samir Palsule
    Member

    Hi Wilson,
    Sorry for the delayed reply. There is a minor confusion in Pp & Ppk indices.
    In automotive industry, Pp & Ppk indices are related to preliminary process capability, which is a snapshot of the process keeping 5Ms constant, i.e. Man, Machine, Method, measurement & Material (composition & other properties). The parts are manufactured and critical dimensions are checked against specification. In this study we try to measure the part to part variation through RMS methodology for standard deviation. There are certain criterias for acceptance of study like x% data points should lie in 3 sigma zone, y% data points should lie in 2 sigma zone, there should not be any trends for the data points. Out colleagues from automotive industry can further elaborate the point. These criterias are laid down because we want to check whether we are using full tolerance zone (TO ENSURE THAT WE ARE NOT ENCURRING HIGH MANUFACTURING COSTS). Off course, special causes need to be identified and eliminated (only after working on the root cause analysis).
    In six-sigma world, Pp & Ppk indices are related to process performance, which talks about long term.
    Hope this clarifies your query.
    – Samir.

    0
    #140175

    Hemanth
    Participant

    Hi Wilson
    Normally Pp and Ppk are used to validate the capability of a new process or product (predominantely used in automotive industry). The study is a short term one (usually spanning a day or two with few setting changes). This is done to avoid any impact of special causes creeping into the system. Also, the formula used considers standard deviation to be calculated as RMS value of all readings.
    Having understood this, for Pp and Ppk it is assumed that there is no confounding of data due to special causes and hence you do not remove any point (assuming it to be because of special cause).
    Coming to your situation.. are you doing a similar study as above? If not then you may want to use Cp and Cpk (after you have stabilised the process).
    In either case I would discourage you to remove any point from your calculations. Hope this is helpful.
    RegardsHemanth

    0
    #140182

    Mikel
    Member

    Wrong

    0
    #140198

    Samir Palsule
    Member

    Stan;
    I do not say I am 100% right. I have tried to reply on my own experience.
    Just saying WRONG does not help anybody. Please come up with explaination Why you feel it is wrong?
    – Samir.

    0
    #158484

    Ang
    Participant

    What determines acceptable when using PP and PPK?
    What determines Random when using PP and PPK?

    0
    #158488

    GrayR
    Participant

    If you have special causes in your data, you do not have a stable process.  Stability means that the process operates with approximately the same average and variation over time.  Therefore, if you don’t have stability, then the average and variation can/will vary overtime, and you cannot determine the process capability. 
    You have to establish process stability before you measure capability.

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

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