SYENIT
Free tool · No sign-up

Cp / Cpk Calculator

Paste your measurements, set the spec limits, and get Cp, Cpk, Pp and Ppk instantly — with a capability histogram. AIAG / ISO 22514 formulas.

Built aroundAIAG SPC 2005ISO 22514-2
Separate values by space, newline or semicolon. Decimal comma or point both work.n=0
Leave one limit blank for a one-sided spec.

What Cp, Cpk, Pp and Ppk tell you

Process capability compares what a process actually produces with what the customer allows. The specification limits (LSL and USL) define the permitted band; the spread and centering of your measurements show how much of that band the process uses. The four indices below condense this into single numbers that suppliers and customers use to agree whether a characteristic is under control.

Cp: potential capability

Cp compares the width of the tolerance with the natural spread of the process, taken as six standard deviations. It ignores where the process is centered, so it shows what the process could achieve if it were exactly on target.

Cp = (USL − LSL) / (6 · σ within)

Cpk: capability including centering

Cpk measures the distance from the mean to the nearer specification limit in units of three standard deviations. It is never larger than Cp and falls as the mean moves toward a limit. With a one-sided specification only the existing limit is used, and Cp is not defined.

Cpk = min(USL − μ, μ − LSL) / (3 · σ within)

Pp and Ppk: performance indices

Pp and Ppk use the same formulas, but with the overall standard deviation of all values. They therefore include shifts and drift between subgroups. Cp and Cpk use the standard deviation within subgroups; for single readings it is estimated from the moving range.

Pp = (USL − LSL) / (6 · σ overall)

Ppk = min(USL − μ, μ − LSL) / (3 · σ overall)

How to read the result

The thresholds below are the ones commonly quoted in AIAG SPC and used in customer requirements:

  • Below 1.00: not capable. The process produces nonconforming parts with the current setup.
  • 1.00 to 1.33: marginal. The process can meet the specification only with tight control; improvement or extra inspection is usually required.
  • 1.33 and above: the usual minimum requirement for a capable, ongoing process.
  • 1.67 and above: often required for special characteristics and for new processes.

Agreed limits differ between customers, so always check the requirement in the purchase specification or quality agreement. A Cpk of 1.33 means the mean is about four standard deviations away from the nearer limit.

When to use Ppk instead of Cpk

Use Ppk when the process has not yet been shown to be stable, for example in a short initial run, a launch or a pre-production sample. Ppk describes what happened in the data you collected, including drift. Use Cpk when the process is in statistical control, checked on a control chart, and you want to describe its short-term potential. If Ppk is clearly lower than Cpk, there is variation between subgroups or drift that the within-subgroup estimate does not see, and the cause should be investigated before the number is quoted.

Assumptions and amount of data

The indices are only meaningful if the following conditions hold:

  • Stable process: no special causes, confirmed on a control chart (I-MR or X̄-R).
  • Approximately normal distribution. Skewed or bounded characteristics such as flatness or runout need a non-normal method; do not apply normal-theory Cpk blindly.
  • Enough data: at least 25 subgroups, or about 100 individual values, collected over a representative period.
  • A capable measurement system. Check the gauge with a Gage R&R study first, otherwise measurement noise is counted as process variation.
  • Correct specification limits taken from the drawing, in the same unit as the measurements. This calculator needs at least 4 values; with so few values the result is only indicative.

Worked example

The example above uses 30 consecutive measurements with LSL = 9.85 and USL = 10.15. The mean is 10.019, which gives Cp = 1.695 and Cpk = 1.486. The overall spread is slightly larger than the within-subgroup spread because the values drift upward, so Pp = 1.584 and Ppk = 1.389. Cpk and Ppk are both above 1.33 but below 1.67: the process meets the usual minimum, and the lower Ppk is a hint to look at the trend before promising more.

Frequently asked questions

What is a good Cpk value?

Most customers ask for at least 1.33 for ongoing production and 1.67 for special characteristics or new processes. Below 1.00 the process cannot meet the specification reliably. Always confirm the value in your customer's requirement.

What is the difference between Cpk and Ppk?

Cpk uses the standard deviation within subgroups and assumes a stable process. Ppk uses the overall standard deviation of all data and therefore also shows drift and shifts between subgroups. Ppk suits initial studies; Cpk suits stable, running production.

How many measurements do I need?

At least 25 subgroups of 4 to 5 parts, or about 100 individual values from a representative period. Results from 10 or 20 values have a wide uncertainty and should be treated only as an early indication.

Can Cpk be negative or very large?

Cpk is negative when the mean lies outside a specification limit. Very large values mean the spread is tiny compared with the tolerance; check the resolution of the gauge and whether the data were rounded.

Does Cpk work with a one-sided specification?

Yes. With only an upper or only a lower limit, Cpk uses the single available distance to the limit, and Cp is not defined. Leave the other limit blank in the calculator.

Save the result as a PDF report and work with studies

Keep your measurements, run complete studies and export the result as an audit-ready PDF report.

Start free trial