Process Capability Study

Six Sigma – iSixSigma Forums General Forums Tools & Templates Process Capability Study

This topic contains 5 replies, has 3 voices, and was last updated by  Matt Savage 2 years, 11 months ago.

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
  • #55476


    In looking into a new process, how can one go about running a capability study?

    The resources I have say that a Cp/Cpk study is useless unless the process is in control. I am exploring a completely new process for producing parts, so sampling the old process is useless.

    Do I need to monitor the control of the process before the Cp/Cpk analysis?
    Say I do a machine runoff, and sample 100 parts. I check that the process is in control. Once process is in control, can I use same sample parts to evaluate Cp/Cpk?



    Quick answers to your questions…
    – No, you don’t need to look at the old process if you are creating an entirely new process. However, it is nice to know how the old process was performing to show the old vs new when telling the story of what work you’ve completed. That’s a good way of showing what value your work brought to the organization.
    – Yes, once the new process is in control you can run a capability to validate it is meeting the customer’s specifications.


    Mike Darrish

    Yes, you can use the sample parts. The idea behind calculating Cp/Cpk only for a process that is in statistical control is to eliminate distortion from special cause variation, which should be investigated and eliminated, leaving only common cause variation, which represents the actual process capability. Reducing common cause variation requires studying the process to understand the causes of that variation and typically a process improvement project.

    I agree with @jpmarth that you should benchmark the old process to prove to yourself and to aausfahrtstakeholders that they should use the new process.


    Steve McMullen

    Don’t know if you have Minitab or not, but there a tool for capability that evaluates stability and normality along with capability information your requesting. One should always consider the normality when using Capability analysis as well.


    Chris Seider

    Consider this thought. If the process isn’t under control, it’s hard to have confidence to a customer what your capability is, so why quote a Cpk and associated ppm defective with an out of control process.


    Matt Savage

    If the measurements you obtain from the 100 parts are in a state of statistical control–no special causes, for example–then you have met one of the requirements prior to performing a capability analysis. If the process is in control, this implies that the future will be similar to the past, unless there is a process change. Do you expect the next 100 or 1,000 parts to be similar (mean and standard deviation) to the first 100 parts? If yes, this implies your process is stable. If you do not expect the next series of measurements to be similar, this is an indication that the process isn’t stable. We want stability /predictability before we can assess acceptability via Cp, Cr, Cpk, Cpm.

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

You must be logged in to reply to this topic.