Statistical Process ControlSPC

Statistical Process Control — SPC — is the method that uses the control chart to watch the process vary and to tell the common-cause variation from the special-cause, so the operator acts on the signal and ignores the noise. The common-cause variation is the inherent randomness of the process, the spread the stable system produces; the special-cause variation is the assignable, the non-random signal that something changed — the tool wore, the supplier switched, the temperature drifted — and the control limits at plus or minus three sigma are calculated from the process data itself, not from the specification, because the limit that flags a change is the one that reflects the process, not the tolerance the customer set.

The discipline of SPC is the reaction rule. The point outside the limit, the run of seven on one side, the trend — these are the patterns that trigger the investigation and the corrective action, and the operator who reacts to the common-cause variation by adjusting the process makes the variation worse, the over-control that Shewhart and Deming warned against. The control chart is the operational measure underneath the nonconformance: the SPC catches the drift before it produces the out-of-tolerance part, the way the FMEA catches the failure mode before it produces the defect, and the two run on the same process. SPC is a core IATF 16949 requirement and a standard tool in the pharmaceutical, food, and aerospace quality systems. See Six Sigma, NCR, and FMEA.