Control limits

Control limits are statistically calculated boundaries used in statistical process control (SPC) to determine whether a manufacturing or business process operates in a state of statistical control. These limits define the expected range of variation for a process and help distinguish between common cause variation (inherent to the process) and special cause variation (resulting from external factors). Walter A. Shewhart developed the concept at Bell Labs in 1924, and his methodology remains fundamental to quality management practices worldwide^[1].

Origins and historical development

The control chart was invented by Walter Andrew Shewhart (1891-1967), an American physicist, engineer, and statistician who worked for Bell Telephone Laboratories. On May 16, 1924, Shewhart wrote an internal memo that introduced the control chart as a tool for distinguishing between common and special causes of variation. This date marks a turning point in industrial quality practices.

Before Shewhart's innovation, industrial quality control was limited to inspecting finished products and removing defective items. His approach transformed the discipline. Rather than reacting to defects after production, manufacturers could now monitor processes in real-time and prevent quality problems before they occurred.

Shewhart published two influential books: Economic Control of Quality of Manufactured Product in 1931 and Statistical Method from the Viewpoint of Quality Control in 1939. These works established the theoretical foundation for modern quality management systems^[2].

Statistical basis

Control limits are typically set at three standard deviations (3-sigma) above and below the process mean. The upper control limit (UCL) and lower control limit (LCL) together with the centerline form the basic structure of a control chart.

Shewhart selected the 3-sigma limits based on Chebyshev's inequality, which states that for any probability distribution, the probability of an outcome greater than k standard deviations from the mean is at most 1/k squared. At three standard deviations, this captures approximately 99.73% of all common cause variation in a normally distributed process.

The formulas for calculating control limits depend on the type of data being analyzed:

X-bar charts (for means): UCL = X-bar + A2*R-bar; LCL = X-bar - A2*R-bar
R charts (for ranges): UCL = D4*R-bar; LCL = D3*R-bar
p charts (for proportions): UCL = p-bar + 3*sqrt(p-bar(1-p-bar)/n)

The constants A2, D3, and D4 are derived from statistical theory and vary based on sample size.

Types of variation

Shewhart identified two fundamental types of process variation:

Common cause variation is inherent to the process itself. It represents the natural variability that occurs even when the process operates as designed. Reducing common cause variation requires fundamental changes to the process structure, equipment, or methods.

Special cause variation arises from external factors not normally present in the system. Machine breakdowns, operator errors, or defective raw materials can trigger special cause variation. When data points fall outside control limits, this signals the likely presence of special cause variation that requires investigation^[3].

Application in quality management

Control limits serve several practical purposes in management and organizational contexts:

Process monitoring: Operators can track production in real-time and detect problems before they produce defective output
Decision making: Points outside control limits trigger investigation and corrective action
Process improvement: Long-term data collection reveals patterns that guide improvement initiatives
Capability analysis: Comparing control limits to specification limits shows whether a process can meet customer requirements

Motorola's Six Sigma methodology, developed in the 1980s, extended Shewhart's concepts. The approach aims to reduce process variation until control limits fall well within specification limits, achieving defect rates of 3.4 per million opportunities.

Control limits versus specification limits

A common misconception confuses control limits with specification limits. Specification limits are determined by customer requirements or engineering tolerances. Control limits are calculated from actual process data. A process may be in statistical control (all points within control limits) but still produce output that fails to meet specifications^[4].

This distinction matters for quality management decisions. A process in control but not capable requires fundamental redesign. A capable process showing out-of-control signals requires investigation of special causes.

{{{Concept}}} Primary topic {{{list1}}} Related topics {{{list2}}} Methods and techniques {{{list3}}}

References

Shewhart, W.A. (1931). Economic Control of Quality of Manufactured Product. D. Van Nostrand Company.
Shewhart, W.A. (1939). Statistical Method from the Viewpoint of Quality Control. Graduate School of the Department of Agriculture.
Wheeler, D.J. & Chambers, D.S. (1992). Understanding Statistical Process Control. SPC Press.
Montgomery, D.C. (2009). Introduction to Statistical Quality Control. 6th ed. John Wiley & Sons.

Footnotes

<references> <ref name="one">Shewhart's original 1924 memo established the foundational concept of distinguishing between controlled and uncontrolled variation.</ref> <ref name="two">The 1931 publication remains a cornerstone text in quality management education.</ref> <ref name="three">The distinction between common and special cause variation determines the appropriate management response to quality issues.</ref> <ref name="four">Process capability indices (Cp, Cpk) quantify the relationship between control limits and specification limits.</ref> </references>

{{a|[[User:Slawomir Wawak|Slawomir Wawak]}]}