Process Capability Calculator (Cp, Cpk)
Paste measurements to get Cp, Cpk, Pp, Ppk, a histogram and expected defect rates
Loading the tool…
Enter measurements and spec limits to get the process capability indices Cp, Cpk, Pp and Ppk, a histogram and the expected defect rate in PPM. Use it to see how comfortably a process meets its specification, or to report capability.
How to use
- Paste spreadsheet values into "Measurements": one column for individual values, several columns for one subgroup per row. A "Subgroup size" cuts a single column into subgroups.
- Enter the "Lower spec limit (LSL)" and/or "Upper spec limit (USL)". A "Target" adds Cpm.
- Read the within (Cp) and overall (Pp) indices, expected PPM and sigma level, with the histogram and normality test.
- Keep the results with "Copy results" or "Save chart as PNG".
How it works
- Cp = (USL − LSL) ÷ (6 × within standard deviation). Cpk = the smaller of (USL − mean) and (mean − LSL), ÷ (3 × within standard deviation). Pp and Ppk use the overall standard deviation.
- "Automatic" estimates the within standard deviation as R-bar/d2 for subgroups up to 10, S-bar/c4 above that, and MR-bar/1.128 for individual values.
- PPM and indices assume a normal distribution. The 95% confidence intervals are approximate, and fewer than 30 values give a warning.
- 1.00, 1.33 and 1.67 are common benchmarks only; the required level depends on the industry and customer.
Questions and answers
- What is the difference between Cp and Cpk?
- Cp compares only the specification width with the spread; Cpk also counts how far the mean is off centre. With USL 20, LSL 8, mean 16 and standard deviation 2, Cp is 1.00 but Cpk is 0.67.
- What is the difference between Cpk and Ppk?
- Cpk uses the within standard deviation from variation inside subgroups; Ppk uses the standard deviation of all data. If the mean shifts between subgroups, Ppk is smaller.
- What does a Cpk of 1.33 mean?
- The nearer spec limit is 4 standard deviations from the mean (1.33 × 3 = 4).
- What if my data are not normally distributed?
- Interpret Cpk and PPM with care. Check the normality test, the histogram and the count of values actually outside the limits.