Cpk (Process Capability Index)

Cpk is calculated as: Cpk = min[(USL − μ) / 3σ, (μ − LSL) / 3σ], where USL and LSL are the upper and lower specification limits, μ is the process mean, and σ is the process standard deviation. The 'k' in Cpk stands for 'centering' - Cpk penalizes a process that is centered off the spec midpoint.

Interpretation in plain language:

• Cpk < 1.0 - process is producing out-of-spec parts. Not acceptable.
• Cpk = 1.00 - barely capable, ~0.27% out-of-spec. Risky.
• Cpk = 1.33 - automotive minimum, ~63 ppm out-of-spec. Industry standard for routine features.
• Cpk = 1.67 - required for safety-critical features, ~0.6 ppm out-of-spec.
• Cpk = 2.00 - Six Sigma 'capable' threshold, ~0.0019 ppm.

Cpk requires a stable process (in statistical control) and normally distributed data over at least 30 measurements (preferably 100+). For unstable or short-run processes, use Ppk (process performance index) instead - same formula, but no requirement that the process be in control. PPAP submissions typically require initial Ppk for first-shipment, then Cpk after production stabilizes.

Common mistakes: calculating Cpk on out-of-control data (the result is meaningless), using too few samples (fewer than 30 produces wide confidence intervals), confusing Cpk with Cp (Cp ignores centering - a process with Cp 2.0 but centered far off-spec can still produce defects), or chasing high Cpk on a non-functional characteristic at the cost of more important features.