What are Group Charts and How To Create One
Group charts are a family of unique control charts that allow several product characteristics to be monitored on the same chart. Group charts are excellent tools to screen locations of high process variations from several possible sources. The group chart achieves its flexibility by grouping together data samples from different process sources but plotting only the maximum and minimum plot points from the group. The author has found that the greatest benefit of the group chart is the tool’s ability to quickly isolate problematic sources of variation among several possible candidates across a part or within a process. This ability does not rely on control limits; therefore, control limit calculations for group charts will not be covered in this section.
Steps for Creating a Group Chart
1. Select a logical number of product characteristics to be monitored at the same time on the same group chart.
(Typically these are measurements of the same feature at different locations across a part. For instance, a cut depth, a web thickness, or a contour. Best results are obtained if the number of features on one group chart is six or less.) The figure below shows a torsion rod with four measurement points.
2. Select a subgroup size.
(If the subgroup size is one, the chart would be a Group Individual X – Moving Range Chart. If the subgroup size is greater than one, the resulting chart would be a Group 17x̄ and R chart.) In our example, we will select a subgroup size of 3.
3. Collect measurements of each characteristic from several parts.
You can log them on a datasheet, as illustrated below. In our example, the measurements are recorded as deviation from nominal.
4. Circle the largest and smallest average and range.
5. On a graph, plot the high and low averages on an x chart, and the high and low ranges on a Range chart.
Label the plot points with the appropriate measurement location.
6. Continue plotting the high and low averages and ranges.
After several plot points, look for locations that repeat as plot points. According to the example below, location B is consistently above nominal. Location C seems to be consistently on the low side, but according to the group range chart, location C also seems to be the most stable. Group charts are extremely versatile. The group chart also works with the short-run x and R data transformations, which will allow you to plot several different types of product characteristics on the same chart. As you experiment with the special charting techniques found in this chapter, you will continue to see SPC applications where you once may have thought impossible.