Capability Analysis

Capability Analysis

In today’s competitive manufacturing environment, it's important that HTI creates a high-quality product the first time, every time. That means controlling variation. Controlling variation is the key to producing consistent high-quality products that meet or exceed customer requirements. To identify and eliminate processing problems that create variation, we gather extensive data from each process and analyze it from a statistical standpoint. This level of analysis lets us understand our ability to effectively control or eliminate most sources of variation. We call this study capability analysis.

Causes of Variation

The causes of variation in a process are numerous, but can be categorized into two groups - common cause variation or special cause variation.

Common Cause Variation
Common causes of variation are all small factors beyond the control of the machine or process operator, things simply inherent to the production environment. These include changes in temperature or humidity, repeatability of equipment, fluctuations in power, and many other things. In combination, these small variations add or subtract from a process’s ability to consistently produce identical parts. Common cause variations are, by nature, uncontrollable, but they tend to be consistent and predictable over time. Thusly, they can be observed, documented, analyzed, understood, and accounted for during production using charts, histograms, and more. A process operating with only common cause variation is declared to be in a state of statistical control.

Special Cause Variation
By contrast, a special cause variation is a variation introduced to a process that's identified as uncommon or unpredicted. Special causes of variation are things like broken cutting tools, a faulty valve, or out-of-spec raw materials. These factors tend to be singular, and are most often eliminated by fixing a specific problem or item. As a manufacturer, we can’t determine a process’s capability when special causes of variation are present - it must be eliminated before making a capability analysis. Once the special causes of variation are eliminated from the process, it can be assessed and declared to be in a state of statistical control.

Capability Analysis

Operating in a state of statistical control, a process can then be accurately compared to the given specification limits to make a capability analysis. To do this, we must gather sufficient data and determine its standard deviation. Standard deviation is a statistical term used to describe the dispersion of a set of data, or the average distance of each data point from the mean. For a process running in a state of statistical control, almost all points (99.73%) will fall within +/- 3 standard deviations from the mean. By taking the standard deviation multiplied by 6 (+3 and -3 from the mean), we can determine the Natural Tolerance of the process. The process’s potential capability is expressed as a Cp Value, determined by dividing engineering tolerance by the natural tolerance.

An Example
Let’s say we're measuring a part length with a given size and tolerance of 1.25" +/- .01". After determining the Natural Tolerance (6 Std. Dev.) to be .008", we can calculate the Cp as .010/.008, or 1.25. A Cp of 1.25 means that the observed process spread (variation) will fit 1.25 times within the given tolerance.

After determining that the potential process variation fits within the engineering tolerances, we must learn if the process was centered within the engineering tolerance. To do this, we compare the mean point of the data to the middle of the engineering tolerance. If the mean of the data set, referred to as the X-bar, is equal to the nominal dimension, the process is perfectly centered. However, it's more likely that the mean of the data will be some measure above or below the nominal dimension.

If the process is not perfectly centered within the engineering tolerance, we'll qualify our process capability statement by taking into consideration the distance from nominal that our process is centered. This value is expressed as the Cpk Value and takes into account both the process spread and the process centering. Cpk value is determined by comparing the X-bar to the upper and lower spec limits individually then dividing the difference from each by 3 standard deviations. If the X-bar of the process is closer to the upper spec limit, the Cpk for the upper limit is used. If the process X-bar is closer to the lower spec limit, the capability to the lower limit is reported as the Cpk value.

The HTI Way

HTI Plastics designs products with manufacturability in mind. Through superior testing, we design our products to maximize capability, providing our customers a high degree of confidence that their products will meet or exceed their expectations for quality. We use state-of-the-art video and touch-probe measuring combined with QC Calc statistical software to measure and analyze each critical feature for control and capability. We target a minimum Cpk of 1.33 or more on critical features.

For every product we make, HTI will deliver full First Article Reports and Process Validation data to communicate our production capabilities to our customers and as a reference point for any process or tooling changes. That way, you can see and understand our test results and production processes until your product is finalized.