Use this tool to determine if your production process meets specified quality limits. It calculates your ability to produce parts within tolerance by comparing the actual variation of your process to the requirements. Apply this analysis only after confirming that your measurement system is reliable and your process is stable.
Evaluating Process Performance
Cpk process capability provides a snapshot of how well your process output aligns with customer or engineering requirements. By focusing on the relationship between the process spread and the specification limits, it quantifies how much of your product meets the necessary standards.
This analysis distinguishes between potential capability and actual performance. By comparing short-term variation—derived from grouped samples—against long-term variation, you can quickly identify if your process is suffering from significant drift or instability.
Understanding the Workflow
To ensure reliable results, this method follows the strict engineering sequence: evaluate the measurement system, ensure the process is in control, and finally, calculate capability. Skipping these steps often leads to metrics that do not reflect true process reality.
Lattice automatically flags concerns if your data shows non-normal patterns, insufficient sample sizes, or if your capability scores fall below standard targets. These flags help you decide whether to focus on reducing variability or re-centering your process mean.
Interpreting Capability Indices
The tool calculates four primary indices to help you diagnose production health. While Cp and Pp look at the total spread allowed by your specifications, Cpk and Ppk account for the centering of your process within those limits. This is vital for identifying whether your process is 'tight' but biased, or simply too variable.
By utilizing your specific USL and LSL values, the platform calculates an estimated DPM (defective parts per million). This translation turns abstract statistical ratios into a clear business metric, allowing you to prioritize interventions based on actual yield risk.
1 · Intent → method
An LLM picks spc_cpk from a fixed catalog.
2 · Method → numbers
Deterministic Python engine runs the math. Same input → same output.
3 · Numbers → plain language
A second LLM translates the result into your domain’s vocabulary.
Why do I need to perform SPC before calculating Cpk?
Cpk process capability assumes the process is operating in a stable state. If your process is drifting or has special cause variation (detected by SPC), the Cpk value will be misleading because the underlying variation is not predictable.
What is the difference between Cpk and Ppk?
Cpk uses short-term variation (based on subgroup averages) to show the potential capability of your process when it is centered. Ppk uses long-term variation (the standard deviation of all data) to show the actual performance you are currently delivering over time.
What should I do if my Cpk is below 1.33?
A result below 1.33 indicates your process lacks sufficient capability. You should investigate whether the issue is caused by the process being off-center or by excessive variation. Check your control charts to see if special causes need to be addressed before making adjustments.
Tool input schema
Schema for spc_cpk not exported yet (run pnpm export:registry).