Carillon Technologies Limited

 Capability Studies

Background

Why?

  • To predict the ability of a machine or process to produce consistently to specifications.
  • To act as a problem discovery technique for operations running outside specification limits.

    • How?

  • By statistical analysis of data gathered from a representative sample.
  • By graphical, calculator or computer techniques.

    • When?

  • Prior to taking delivery of new equipment.
  • As part of a production approval process.
  • As production starts, to establish process capability (equipment, tooling, material , operator combinations).
  • As a quality assurance audit.
  • When out-of-specification conditions are found.

    • Where?

  • In your equipment supplier's plant.
  • In the company plant.
  • In outside supplier's plant.

    • What?

  • All characteristics that require control.

    • Who?

  • Manufacturing or production engineering.
  • Production.
  • Quality Control.

    • The Need For Capability Studies

    1. The aim of SPC is to help bring about continuous improvement to all processes, helping achieve customer satisfaction.

    2. To assure continued customer satisfaction, strict conformance to all customer requirements must be a goal of a company.

    3. It is more economical to produce to requirements than to sort and rework in order to meet requirements.

    4. Consistent performance requires consistent machine and process capability.

    5. Process capability = machine capability (inherent variability) + process control.

    6. To measure this conformance, capability studies should be performed.

    Principles of Variation

    1. No two items are exactly alike.

    2. Tolerance result from the fact that it is impossible to make two items exactly alike.

    3. The differences between these two items can be measured and the reasons for the variability can be determined.

    4. There are two major types of variation : Assignable Variation & Chance Variation

    5. Machine Capability is the study of chance variation.

    6. Process capability is the study of all chance variation and assignable variation normally present in the controlled process.

    Capability & Variable Control Charts

    The best way to determine process capability is with the use of control charts on processes that are in statistical control. A study utilizing this method will provide the basis for a Cpk index.

    An alternate method for the determination of machine and process capability is the use of run charts, histograms and descriptive statistics.

    The process must be consistent over time as indicated by the run chart.
    The process must be normally distributed as indicated by the histogram.
    The number of data points collected should number at least 30, in order to provide a sufficient number of values for determining the descriptive statistics.
    This method will provide the basis for a Ppk index.

    Capability

    Definition: Capability is the inherent reproducibility of a process. In statistical terms, a process's capability falls within a range of 6s.

    This 6 standard-deviation spread will include 99.7% of the product being created by the process.

    Note the capability does not indicate whether the process is producing what your customers want. It is only indicating what the process is capable of doing.

    Capability Measures

    Capability Ratio (CR) and Capability Index (Cp)

    The capability ratio is a fraction that predicts what portion of the total tolerance is used by 99.7% of the population.

    The capability index is the reciprocal of the capability ratio

    The s is based on the "inherent variability" of the process - based on a process in control.

    Performance Ratio (PR) and Performance Index (Pp)

    The performance ratio is a fraction that predicts what portion of the total tolerance is used by 99.7% of the population.

     

    The performance index is the reciprocal of the performance ratio

    The s is based on the total variation of the process.

    Cpk

    The s is based on the "inherent variability" of the process - based on a process in control.

    Ppk calculates the Zupper, Zlower, and minimum the same as the Cpk.

    The s is based on the total variation of the process

     
    Comparsion

    Prob. Value\Index

     CR

    Cp

    Cpk (centered process)

    9.9 x E -10 
    		 .50  2.0  2.0

    2.9 x E - 7 
    		 .60   1.66  1.66

    3.2 x E - 5 
    		 .75  1.33  1.33

    1.35 x E - 3

     1.00

     1.00

     1.00

    1.22 x E - 2

    1.33

    .75

    .75

    Return