Goals for continual process improvement
Specification limits represent the ideal measurement requirements that should be met in order for a part or portion of a process to operate optimally. Specifications are boundaries, usually set by management, engineering, or customers, within which a system must operate. They are sometimes called engineering tolerances. Specification limits are used to determine if the product will function in the intended fashion. As such, they represent the goal in striving for continual improvement and increased quality.
Control limits represent statistical characteristics of a process. They refer to the boundaries in which sample statistics can be expected to vary due simply to the randomness of the sample used. Specifically, control limits represent the dividing lines between random deviations from the mean of the distribution and non-random deviations from the mean of the distribution. Control limits are used to determine if the process is in a state of statistical control (i.e., is producing consistent output). They are the upper and lower limits, representing the outer values for 3 standard deviations above and below the center line process – the process average. In this way, the span between these values represents values that will occur in the process 99.7% of the time. Therefore, any process with values falling outside these limits will be displaying special cause variability – i.e. this value would only occur 3 in 1000 times – a very rare occurrence very unlikely to show up in the observations due to normal chance.
A control chart is a time-ordered graphical plot of sample statistics. Control charts include two control limits that separate the random variation and non-random deviation - the larger value or the upper control limit (UCL) and the smaller value or the lower control limit (LCL). A sample statistic that falls between these two limits suggests randomness and predictability, while a value outside either limit suggests non-randomness.
A process can be in statistical control without being in compliance with its specification limits. This would mean that the process is stable and predictable, but not yielding results that meet the design criteria. In this case, the center line process value would be different than the center line specification. The goal would then be to make adjustments to the system that bring the center line process to the center line specification, while still keeping the process in statistical control.
Often control limits and specification limits are mistakenly used interchangeably. Control limits and specification limits are completely different values and concepts. Typically, there is no relationship whatsoever between control limits and specification limits. Control limits are calculated from process data for a particular control chart.
Specification limits are chosen in numerous ways. They generally apply to the individual items being measured and appear on histograms, box plots, or probability plots.
Confusing control limits with specification limits leads to mistakes. The most common mistake is to use specification limit values instead of control limit values on an X-bar chart or an Individuals chart. Using specifications on an X-bar is the most egregious error: the specifications are in one unit (items) while the chart is in another (average of several items). Unless the specification and control limit values are identical, one of two errors occurs: (1) the control limits are set too tightly which leads to over-adjustment and tampering with the process. Tampering adds to process variation, resulting in lower quality and higher costs’; and/or (2) the control limits are set too loosely. Signals of process change are ignored and opportunities for process improvement are missed. The result is additional avoidable variation, lower quality, and higher costs (Chase, Jacobs & Aquilano, 2005).
Reference
Chase, R. B., Jacobs, F. R. & Aquilano, N. J. (2005). Operations management for competitive advantage, 11th ed., New York,. NY: McGraw-Hill/Irwin