01 Distinguishing Process Limits from Customer Expectations
In Statistical Process Control (SPC), a fundamental source of confusion is failing to distinguish control limits from specification limits. Control limits represent the empirical voice of the process. They are calculated mathematically from process observations (typically three standard deviations around the mean). These limits indicate how the process naturally behaves when only common-cause variation is present.
Specification limits, by contrast, represent the voice of the customer or design engineering. They set target boundaries for individual physical parts, dimensions, or service levels. Specification limits are established independently of the production line based on product usability, mechanical assembly fit, or safety regulations.
"Control limits reveal what your process is currently capable of delivering. Specification limits specify what your customer demands. Conflating the two creates misaligned quality goals."
Process Engineering Review
Discuss your statistical variation observations with an atlas specialist.
02 Why Specification Limits Do Not Belong on Control Charts
Superimposing upper and lower specification limits directly onto standard control charts leads to operational misinterpretation across engineering and production teams:
- Comparing subgroup means to individual specs creates mathematical errors, as sample averages exhibit tighter dispersion than individual measurements.
- A process point inside specification limits may still reflect an out-of-control condition, hiding early signals of equipment degradation or shift changes.
- Attempting to adjust a stable process simply because values approach a specification boundary introduces tampering, which actually increases overall process variance.
03 Connecting Stability to Process Capability
Evaluating whether a stable process fulfills technical specifications requires process capability indices such as Cp and Cpk. Capability metrics compare the 6-sigma statistical width of common-cause variation against the allowable specification range defined by engineering.
A process can be strictly in statistical control yet consistently produce non-conforming product if its natural spread exceeds specification limits. Conversely, an out-of-control process might temporarily meet tolerances while remaining inherently unpredictable. Achieving control limit stability is the indispensable baseline prior to performing capability assessments.
Need Help Structuring Control Chart Baselines?
Consult with our statistical methodology specialists to properly define subgrouping strategies and separate process limits from engineering specifications.
Consult an SPC SpecialistFrequently Asked Questions
Yes. Control limits reflect actual historical variation in the process. If natural process variation is wide, the calculated 3-sigma control limits will fall outside customer specification limits, signaling that the process cannot meet specifications without fundamental redesign.
X-bar charts plot averages of subgroups rather than individual data points. Due to the Central Limit Theorem, subgroup averages have a smaller standard deviation than individual values. Plotting individual specification limits on an average chart creates a misleading comparison.
Control limits should be recalculated only when a deliberate process change—such as new tooling, raw material upgrades, or physical process modifications—permanently shifts the baseline variation. Routine data additions to a stable process do not require recalculating limits.