Six Sigma Process Capability Calculator

About Six Sigma Process Capability Calculator

The Six Sigma Process Capability Calculator measures how well a stable process meets its specification limits. Quality engineers, Six Sigma Black Belts and reliability teams use the indices it computes — Cp, Cpk, Pp, Ppk, sigma level, DPMO and yield — to answer one of the oldest questions in operations: given what we are making, how often will we make something the customer cannot accept?

This tool stands out from a plain Cpk calculator in three ways. First, it accepts two input modes: you can either paste raw measurements (it computes everything for you) or skip straight to entering the mean and sigma. Second, it correctly separates within-subgroup sigma (used for Cp and Cpk) from overall sigma (used for Pp and Ppk) — many free calculators conflate them. Third, it visualises the result: a Normal-distribution overlay with USL, LSL and target markers, a capability dial, and a 1σ–6σ scale showing exactly where your process sits.

The four capability indices, demystified

Cp measures potential capability: how well the process could fit between the specs if it were perfectly centred. It compares the spec width (USL − LSL) with six standard deviations of process variation. A Cp of 1.0 means the process spread exactly fills the spec window — no margin. A Cp of 1.33 leaves 33 % margin; 2.0 leaves 100 %.

Cpk measures actual capability. It takes the smaller of (USL − μ) / 3σ and (μ − LSL) / 3σ, so it penalises a process that drifts away from the target. Cpk ≤ Cp always, and the gap between them is a centring problem you can usually fix with adjustment rather than redesign.

Pp and Ppk are the long-term cousins. They use the overall standard deviation — computed from every measurement in the study — so they include drift between subgroups, tool wear, shift-to-shift changes and any other slow movement. If Pp ≪ Cp, your process is not as stable as it looks moment-to-moment.

Sigma level and DPMO

The sigma level is shorthand for "how many σ separate the process mean from the nearest spec limit". A process at 6 σ short-term, after the conventional 1.5 σ shift, is associated with 3.4 defects per million opportunities (DPMO) long-term. This calculator reports both the short-term sigma level (what you would see on a control chart) and the DPMO and yield computed directly from the Normal distribution.

How to Use This Tool

1. Pick an input mode. Choose Summary statistics if you already have μ and σ; choose Raw measurement data if you want to paste readings.
2. Enter spec limits. Provide USL, LSL or both. Target is optional but appears on the chart.
3. Provide the data. In summary mode, enter the mean and sigma. In raw mode, paste at least two numbers (separated by commas, spaces, or new lines).
4. Submit. The report shows Cp, Cpk, Pp, Ppk, sigma level, DPMO, yield and a plain-language verdict — with a Normal-curve overlay, capability dial and step-by-step working.

What does a "good" Cpk look like?

• Cpk < 1.00 — not capable. Defects expected during normal operation.
• 1.00 ≤ Cpk < 1.33 — marginal. Small shifts will produce defects.
• 1.33 ≤ Cpk < 1.67 — capable. The classic industry benchmark.
• 1.67 ≤ Cpk < 2.00 — excellent. Comfortable margin to spec.
• Cpk ≥ 2.00 — world-class. A true Six Sigma process.

Worked example

A bottling line targets 500 mL per bottle with specs LSL = 497 mL and USL = 503 mL. The process produces μ = 500.4 mL with σ = 0.62 mL. Cp = (503 − 497) / (6 × 0.62) ≈ 1.61, Cpk = min((503 − 500.4) / (3 × 0.62), (500.4 − 497) / (3 × 0.62)) = min(1.398, 1.828) ≈ 1.40. The process is comfortably capable (Cpk ≥ 1.33), and the slight off-target mean shows up as Cpk being noticeably less than Cp.

Frequently Asked Questions

What is the difference between Cp, Cpk, Pp and Ppk?
Cp/Cpk use the within-subgroup σ (short-term, R̄/d₂) and tell you how capable the process could be at its current spread. Pp/Ppk use the overall σ (long-term, including drift) and tell you how it actually performed. Cp and Pp ignore centring; Cpk and Ppk penalise off-target processes.

How is sigma level related to DPMO?
Sigma level is the short-term Z value — the distance, in σ, from the mean to the nearest spec limit. DPMO is the long-term defect rate per million units, computed from the Normal-distribution tail areas beyond the specs. The classic Six Sigma table maps a short-term 6 σ level to 3.4 long-term DPMO, after a 1.5 σ shift convention.

What is the 1.5 sigma shift?
An empirical observation that processes drift by about 1.5 σ between short-term studies and long-term operation. By convention, long-term sigma level ≈ short-term sigma level − 1.5. That is why a process measured at 6 σ short-term is associated with 3.4 DPMO long-term, not the much smaller true 6 σ tail probability.

Can I use this with only one spec limit?
Yes. Leave the unused limit blank. Cp and Pp need both limits and will be marked n/a, but Cpk and Ppk are computed as a one-sided index — for example, Cpk = (USL − μ) / (3 σ) for upper-only specs.

Which sigma is used for what?
Within-subgroup σ (R̄ / d₂) feeds Cp and Cpk. Overall σ (sample standard deviation with n − 1) feeds Pp, Ppk and the DPMO calculation. The two are equal only when the process is perfectly stable; the bigger the gap, the more drift you have between subgroups.

Why does my Cpk differ from Pp?
Cpk uses within-subgroup σ and is the minimum of the upper and lower one-sided indices. Pp uses overall σ and ignores centring. So Cpk falls when the process is off target; Pp falls when long-term variation is high. Compare them: a big Cp/Pp gap signals instability over time, while a big Cp/Cpk gap signals an off-target mean you can usually adjust away.