An Overview of Six Sigma Statistical Process Control
Six Sigma is a quality control methodology that relies heavily on statistical theory to improve the quality and efficiency of business processes. This is done in order to decrease the defects in the end result of the processes; a business’ product or service.
The Six Sigma quality system relies heavily on the statistical analysis and statistical process control (SCP). Six Sigma statistical process control tools allow you find out whether the business process in questions is manageable and stable, or if it is trending towards variability which could lead directly to errors in the end product.
Limits are categorized into lower and upper. The lower control limit (LCL), would be set three sigma levels under the mean, while the upper control limit (UCL) is usually set at three sigma levels, over the mean. Since around 99 percent of average process unpredictability will take place within plus or minus three sigma, if the procedure is managed, it should approximate a standard distribution around the mean, and every data point needs to be inside the pre-defined limits.
In order to compute restrictions, you should first identify the mean. Begin with a trial of 30 or more procedure observations, for instance the altitude of a solder bump on top of a circuit board, calculated in thousandths of an inch. Compute the mean through adding up all the values and dividing by the number of evaluation or observation. If the trial size is 30 and the total of the experimental values is 173, the method would be 173/30 = 5.8.
The standard deviation (SD) is easier to compute with the use of the programmed standard deviation calculator in a numerical analysis program or the STDEV function that can be found in a spreadsheet program. For instance, let’s presume the SD is 1.8.
The formula to compute the UCL is (3 x SD) + (Process Mean) = UCL. In this example, this would appear to be (3 x 1.8) + 5.8 =11.3. The LCL would be computed as (Process Mean)-(3 x SD) = LCL. Going back to the example, this would appear to be 5.8 – (3 x 1.8) = 0.3. To sum up, the process mean for this trial should be 5.8, and should be exactly focused between the LCL of 0.3 and the UCL of 11.3. These values will be applied in the next section to produce charts.
A control chart is simply a line chart presenting chronological measurements of a procedure characteristic, such as the size of a machined part, with additional lines to illustrate the lower and upper limits. Statistical software packages will have automated chart functions that will create these charts for you. When you assess a chart, you’re looking for signals that the procedure could be unmanageable or trending toward being unmanageable.
If any of the following caution signs are present, the procedure could be unmanageable or is trending toward becoming unmanageable:
• A particular point that is outside either of the control limits
• Two out of three points in a line that are on the related side of the middle line and two sigma or more away from it
• Four of five consecutive points on one side of the middle line and higher than one sigma from it
• Eight or more points in a line that are moving the same way.
While the measurements might still be inside the acceptable ranges, if the procedure isn’t manageable, it is time to make a move because you’ll soon see faulty units generated by the procedure. Six Sigma statistical process control is not easy to understand, even though it is crucial to the entire Six Sigma Methodology and Goals. It is best to perform these tests within the capable hands of a trained and certified Six Sigma Black Belt.