Skip to content
LeanSuite

Free template

Updated

Process capability (Cpk) study sheet

A process capability study measures how well a stable process fits inside its specification limits. You take readings of one characteristic in small subgroups over time and compare the spread and the centering of the process with the lower and upper specification limits (LSL and USL). Cp = (USL − LSL) ÷ 6σ says how many times the process spread fits in the tolerance; Cpk = the smaller of (USL − μ) and (μ − LSL), divided by 3σ, also accounts for where the mean μ sits. Which σ you use matters: Cp and Cpk use the short-term σ within subgroups, usually R̄ ÷ d₂, while Pp and Ppk use the overall σ of every reading, which includes drift between subgroups. The PDF has a data sheet for 25 subgroups of 5 readings with sums, means and ranges, and a page with every formula, the control limits, a worked example and a histogram grid. The Excel version takes subgroups of 2 to 5, works out the x̄ and R control limits and flags subgroups outside them, and calculates Cp, Cpk, Pp and Ppk (one-sided when only one limit is given), with a histogram and a chart of the subgroup means.

Free to use: print it, copy it and edit it for your team. Enter your name and work email once to download.

Page 1 of the process capability study sheetFields for part, characteristic, nominal, LSL, USL, gauge and operator, a grid of 25 subgroups with a date and time and five readings each and rows for the sum, mean and range, and a box for events during the study.
25 subgroups of 5Means and rangesControl limitsσ within and σ overallCp, Cpk, Pp and PpkWorked exampleHistogram

When to use it

When to use a process capability study

  • For a new process or a new part, before production approval, often for the special characteristics in a PPAP.
  • After a change to a machine, tool, material or method, to confirm the process still fits.
  • To decide whether a characteristic can move from 100% inspection to sampling.
  • When a customer asks for capability data on a characteristic.

How to fill it in

  1. 1

    Check the gauge first

    A gauge with poor repeatability adds its own spread to the result. Check it with a gauge R&R before the study.
  2. 2

    Collect subgroups

    Take 5 consecutive parts at regular intervals, 25 times: 125 readings. Record the time and anything unusual, such as a tool change or a new batch.
  3. 3

    Work out means and ranges

    For each subgroup, the mean (sum ÷ 5) and the range (largest − smallest). The Excel version does this as you type.
  4. 4

    Check stability

    Mean chart limits are x̿ ± A₂R̄, range chart upper limit D₄R̄. A subgroup outside the limits means a special cause: find it before quoting capability.
  5. 5

    Calculate the indices

    σ within = R̄ ÷ d₂ (2.326 for subgroups of 5); σ overall is the standard deviation of all readings. Then Cp, Cpk, Pp and Ppk.
  6. 6

    Read the result

    Compare with your customer's requirement. A low Cpk with a good Cp means the process is off center; a low Cp means the spread itself is too wide.

A filled-in example

Illustrative, not a benchmark

An example: a shaft diameter of 10.00 ± 0.10 mm (illustrative numbers).

Why the example's Cpk is lower than its Cp

  • Mean to USL: 0.08 mm = 4.0σ (the nearer limit)
  • Mean to LSL: 0.12 mm = 6.0σ

Cp = (USL − LSL) ÷ 6σCpk = (gap from the mean to the nearer limit) ÷ 3σ

Nearer gap: USL − mean = 10.10 − 10.02 = 0.08 mm (the LSL gap is 0.12 mm).

  1. Cp =0.20 ÷ (6 × 0.02)= 1.67
  2. Cpk =0.08 ÷ (3 × 0.02)= 1.33

A normal curve with mean 10.02 millimeters and sigma 0.02 between limits 9.90 and 10.10. The mean is 4 sigma from the upper limit and 6 sigma from the lower one, so Cpk is 4 divided by 3, 1.33, while Cp is 1.67.

Cp only looks at the width of the curve. Cpk looks at the nearer limit, here the USL, 4σ away, so 4σ ÷ 3σ = 1.33.
  • LSL 9.90 mm, USL 10.10 mm. 25 subgroups of 5, no subgroup outside the control limits.
  • Mean μ = 10.02 mm, σ within = 0.02 mm.
  • Cp = 0.20 ÷ (6 × 0.02) = 1.67.
  • Cpk = smaller of (10.10 − 10.02) and (10.02 − 9.90) = 0.08, divided by 3 × 0.02: 0.08 ÷ 0.06 = 1.33.

The spread would give a Cpk of 1.67 if centered, but the mean sits 0.02 mm high, so Cpk is 1.33. Moving the machine offset down to the nominal is the fix, not a new machine.

Common mistakes

  • Quoting Cpk for an unstable process

    Capability describes a stable process. With points outside the control limits, the number does not predict anything.

  • Not saying which σ you used

    Cpk from short-term σ and Ppk from overall σ give different numbers when the process drifts. Report both, labeled.

  • Too few readings

    A Cpk from 20 parts has a wide margin of error. 25 subgroups of 4 or 5 is a common study size.

  • Ignoring the shape

    The formulas assume roughly normal data. A histogram with two peaks or a long tail needs a look at the process first.

Download the template

Free to use: print it, copy it and edit it for your team. Enter your name and work email once to download.

Run this template in LeanSuite

In LeanSuite quality checks run as digital checklists with number limits, an out-of-spec answer can raise a tag or start a CAPA automatically, and KPI Builder tracks quality KPIs against targets.

See it in a 45-minute demo

FAQ

Process capability (Cpk) study sheet: common questions

  1. Cp = (USL − LSL) ÷ (6 × σ). Cpk = the smaller of (USL − mean) and (mean − LSL), divided by (3 × σ). For both, σ is the short-term standard deviation within subgroups, usually the average range divided by d2.

  2. They use the same formula with a different σ. Cpk uses the short-term σ within subgroups; Ppk uses the overall σ of every reading, which includes shifts and drift between subgroups. When the process is stable the two are close.

  3. A common minimum is 1.33 for an established process, and 1.67 is often asked for new processes and special characteristics. Your customer's requirement is the one that applies. Below 1.00, some output falls outside the limits.

  4. A common study is 25 subgroups of 4 or 5 consecutive parts, 100 to 125 readings, taken over enough time to see normal variation. Fewer readings give an index with a wide margin of error.

  5. Cp needs both limits, so it is not calculated. Cpk uses the distance from the mean to the one limit: (USL − mean) ÷ 3σ or (mean − LSL) ÷ 3σ. The Excel version does this when one limit is left blank.

More free templates

All templates
Browse all 278 templates

Can't find the template you need?

Tell us what you're looking for. We'll make it and send it to your inbox, free.

A shift handover sheet, a red tag log, a changeover checklist for one machine: if your team uses it on the floor, ask.

By sending this form you agree to our Terms of Service and Privacy Policy.

Pass it on

We want LeanSuite to be the best place on the internet for lean help. If this was useful, send it to someone on your team or in your network who needs it.