♦
160
♦
WaveRIDER
®
The equations used to calculate the index values are as follows:
.)
.
(
*
6
Dev
Std
LSL
USL
Cp
−
=
.)
.
(
*
3
Dev
Std
x
USL
Cpk
−
=
OR
.)
.
(
*
3
Dev
Std
LSL
x
Cpk
−
=
, whichever is less
As can be interpreted from the above equations, Cp gives an indication of how narrow
the data distribution is relative to the width of the specification limits. Essentially, it
indicates how well the process would be able to stay within the specified limits if the
data were perfectly centered between those limits.
Cpk compares the widest half of the data distribution to the appropriate specification
limit. It indicates whether the process is capable of meeting the specification as
indicated by the “worst half” of the measurements. Unlike Cp, the Cpk index measures
process capability without assuming the data is well-centered.
The figures below give a graphical representation of the concept of Cp and Cpk. Notice
that in each graph, the same upper and lower specification limits (USL, LSL) are used.
The values of Cp and Cpk will differ according to the data that is compared with those
specifications.
Depending on the particular process being monitored, the desired value for Cp and Cpk
may differ. In general, however, a Cp and Cpk of 1.33 or above is desired. This
assures that the process is not only capable of meeting the required specification limits,
but also has a built-in margin for error that may be needed in special circumstances. In
addition to targeting a certain minimum Cp and Cpk, it is also desirable to have these
two values equal one another. This indicates that the process is well-centered between
the specification limits.
