Capability inputs
Result
This result is an estimate based on your inputs and a normal-distribution approximation. PPM/DPMO and sigma level do not guarantee production acceptance or shipment quality.
Histogram and specification markers
Index details
Method notes
- Cp = (USL - LSL) / (6 sigma). Cpk is the smaller of Cpu and Cpl, so it includes centering.
- With pasted measurements, within sigma is estimated from average moving range divided by 1.128, and overall SD is the sample standard deviation.
- Pp and Ppk use overall SD. Summary mode does not show Pp/Ppk because one entered SD cannot prove both within and overall variation.
- One-sided specifications report the applicable Cpu or Cpl as Cpk. Cp and Pp require both specification limits.
Related statistics tools
Use the standard deviation calculator before entering summary inputs, descriptive statistics for a broader measurement summary, normal distribution for general tail probabilities, and Likert survey summary for aggregated survey counts.
FAQ
What is the difference between Cp and Cpk?
Cp compares the two-sided specification width with process spread. Cpk also considers how close the process mean is to the nearest specification limit, so an off-center process can have a lower Cpk than Cp.
What is the difference between Cpk and Ppk?
This calculator uses moving range divided by 1.128 as the within sigma for Cp and Cpk when measurement data is pasted. Pp and Ppk use the overall sample standard deviation from all pasted measurements.
How are one-sided specifications handled?
For an upper-only or lower-only specification, the page reports the applicable side index, Cpu or Cpl, as Cpk. Cp and Pp are not defined because they require both LSL and USL.
How are PPM and DPMO estimated?
The estimate uses a normal distribution approximation with the overall sample SD when raw measurements are pasted, or with the entered SD in summary mode. It is an estimate, not a guarantee of actual defect rate.
Why can this result differ from Minitab or another calculator?
Capability results can differ because of the within-sigma estimator, subgroup handling, normality assumptions, one-sided specification rules, and display rounding. This MVP uses individual moving ranges for within sigma and sample SD for overall sigma.