What it measures
It compares process variation against customer specification limits and shows whether output is stable, centered, and commercially acceptable.
When to use it
Use it during validation, line release, supplier reviews, corrective actions, and continuous improvement projects across machining, filling, molding, and assembly.
Why it matters
A strong capability index helps cut scrap, reduce complaints, improve delivery confidence, and support meaningful quality targets.
Enter your process inputs
Sample production run
This example shows a centered process with moderate variation and realistic interpretation targets.
| Lot | Observed Mean | Within Sigma | Overall Sigma | LSL | USL | Cpk | Ppk |
|---|---|---|---|---|---|---|---|
| A-101 | 50.03 | 0.148 | 0.161 | 49.50 | 50.50 | 1.06 | 0.97 |
| A-102 | 49.98 | 0.121 | 0.127 | 49.50 | 50.50 | 1.32 | 1.27 |
| A-103 | 50.08 | 0.095 | 0.101 | 49.50 | 50.50 | 1.47 | 1.39 |
| A-104 | 49.92 | 0.082 | 0.091 | 49.50 | 50.50 | 1.71 | 1.47 |
Core process capacity equations
- Cp = (USL - LSL) / (6 × within sigma). Cp measures potential capability when the process is perfectly centered.
- Cpk = min[(USL - mean) / (3 × within sigma), (mean - LSL) / (3 × within sigma)]. Cpk adjusts for off-center performance.
- Pp = (USL - LSL) / (6 × overall sigma). Pp uses long-term variation across the entire process history.
- Ppk = min[(USL - mean) / (3 × overall sigma), (mean - LSL) / (3 × overall sigma)]. Ppk reflects delivered performance.
- Cpm = (USL - LSL) / [6 × sqrt(overall sigma² + (mean - target)²)]. Cpm penalizes drift away from the target value.
- Z lower = (mean - LSL) / overall sigma and Z upper = (USL - mean) / overall sigma. The smaller side becomes the Z Bench.
- Yield = 1 - outside probability. Estimated defect rates and parts per million come from the normal distribution tails.
Practical workflow
- Choose Raw measurements if you have individual readings from the line. Paste them with commas, spaces, or new lines.
- Choose Summary statistics if you already know the mean and sigma values from another quality system or study.
- Enter the lower and upper specification limits exactly as defined on the drawing, control plan, or customer requirement.
- Add a target value when nominal centering matters. This enables Cpm, which is useful for loss-based quality thinking.
- Press Calculate Process Capacity. The results appear above the form, directly below the page header.
- Use the export buttons to save a CSV table or a PDF summary for meetings, audits, supplier reviews, or improvement records.
FAQs
-
1. What is a good Cpk value?
Many manufacturers treat 1.33 as a common minimum for capable processes. Higher-risk products often require 1.67 or above, especially during validation or launch stages. -
2. Why are Cp and Cpk different?
Cp assumes the process is centered. Cpk shows how close the mean is to the nearest specification limit, so it drops when the process drifts. -
3. Why calculate Ppk as well?
Ppk uses overall sigma, so it reflects real long-term behavior. It often reveals hidden instability that short-term capability alone can miss. -
4. What does Cpm add?
Cpm includes target centering in the denominator. It is useful when hitting the nominal target matters, even if all parts still fall inside specifications. -
5. Does the calculator assume normality?
Yes. Yield, tail probability, and parts-per-million estimates are based on a normal distribution. If your data are strongly non-normal, confirm results with additional analysis. -
6. Can I use raw data from one-piece flow?
Yes. Raw mode is well suited to individual observations. It estimates within sigma from the average moving range, which is common for individuals charts. -
7. What if my process has one specification limit?
This page is built for two-sided specifications. For one-sided studies, evaluate the relevant side directly and interpret Cpu or Cpl separately. -
8. Why might Cp be high while Ppk is low?
That pattern usually means the process can perform well short term but shifts, drifts, or widens over time. Stability improvement should come before celebration.