Process Not Centered

Cp is healthy but Cpk lags — the mean has drifted off target. What is causing the drift, and how do I prevent it from returning after a one-shot offset?

Cp ≥ 1.33 but Cpk is significantly lower (typically by 0.20+).

1. Compute the offset magnitude and direction relative to nominal — not to the nearest spec limit.
2. Overlay the last 5 shifts on one control chart to detect a ramp or step pattern.
3. Pull the tool-life counter; correlate mean drift with cumulative cycles.
4. Check whether the drift resets after machine warm-up, tool change, or shift start.
5. Verify the first-piece verification rule in the Control Plan is being executed.
6. Inspect fixture datum surfaces and clamp forces for wear or chip build-up.
7. Apply the one-shot offset only after the drift mechanism is understood.
8. Install the appropriate long-term control: wear compensation, warm-up cycle, probing, or setup rework.
9. Repeat the capability study on a fresh window; require Cpk ≥ 0.90 × Cp.
10. Update Control Plan reaction rules and PFMEA detection rating.

• Tool-wear compensation disabled or set to zero.
• First-piece verification skipped because "the machine was already running".
• Fixture locator worn on one side, biasing the datum.
• Thermal growth of the spindle or fixture during warm-up.
• Operator adjusting offsets between subgroups (hand steering).
• Mixed material lots with differing hardness offsetting cut depth.

• Tool wear usually affects centering long before it affects variation — a rising mean with stable range is the fingerprint.
• Operators re-center on gut feel far more often than logs suggest. A control chart that "looks too flat" is often the result of hand steering.
• An offset that must be re-applied every shift is not a fix — it is a workaround. Find why the process cannot hold nominal.
• Warm-up drift is under-reported in machining; the first-hour parts often bias the study upward.

• Do not apply centering logic when Cp itself is below the target — you will only trade one problem for another.
• Do not offset a process that is out of statistical control; you are chasing noise.
• Do not use Cpk to justify hand steering — the resulting chart is not a real process signal.

• Cp − Cpk gap: Direct measure of miscentering severity. Cpk = Cp × (1 − k).
• Cpm: Penalizes deviation from target and reveals miscentering that Cpk alone under-reports on wide tolerances.
• Ppk vs Cpk: If Ppk < Cpk sharply, the drift is between-subgroup — centering is time-varying, not static.

1. Apply a one-shot offset to re-center.
2. Check setup procedure for first-piece verification.
3. Audit tool-wear compensation rules.
4. Use the free Process Capability Calculator to compute Cp, Cpk, Pp, Ppk, sigma level, and predicted PPM from your measurement data — no sign-up required.

• Never accept a one-shot offset as corrective action — it is containment.
• Never let the operator adjust more than what the reaction plan permits; log every adjustment.
• Never re-center during a running capability study; complete it, then correct.

• offsets, thermal growth or tool condition — Machining conditions provide practical centering causes.

• ✓Cp − Cpk gap quantified and attributed
• ✓Drift mechanism identified (wear / thermal / setup / operator / material)
• ✓Long-term control implemented, not just a one-shot offset
• ✓First-piece verification rule enforced
• ✓Capability re-run on a stable, drift-free window
• ✓Cpk ≥ 0.90 × Cp achieved and sustained over 25 subgroups
• ✓Control Plan and PFMEA updated

• Process condition: statistical stability is required.
• Distribution assumption: use a distribution model justified for the data.
• Confirm process stability and measurement-system adequacy before interpreting a capability index.
• Use a justified distribution model or non-normal method when the normal model is unsuitable.