R&R Analysis Using Graphical Techniques

Graphs of measurement results provide a visual means to analyze measurement systems.

Appropriate graphical techniques include:

• Control charts used to monitor test equipment variation and appraiser variation.
• Scatter plots to look at how consistent different appraisers are to each other as well as their consistency taking measurements of the same part or sample.
• Whiskers charts to view how consistent each appraiser is in measuring the same part as well as to check the consistency between appraisers.

Control Charts for MSA

• The x-axis is used differently.
  • When used to analyze the repeatability and reproducibility of the measurement system, the x-axis of the control chart does not represent time.
  • The parts or samples tested by each appraiser are tracked on the x-axis.
• As usual, check the Range chart first.
  • Any point outside the upper control limit on the range chart or any unusual patterns of variation (such as all of the range values for one appraiser being below the average range) are a sign that the measurement system is out-of-control. It should be investigated and improved.
• With the Averages (X-bar) chart, we “want” to see points outside of the control limits.
  • This is obviously different than our goal with a standard averages chart. With standard averages charts, we want all points to be in-control. We do not want that here.
  • An X-bar & R control chart on the measurement system analysis results must have range values for each part/sample and operator in-control. But unless the majority of the values on the X-bar chart are outside of the control limits, then the measurement system is not adequate.
  • Unless more than half the averages are outside of the control limits, the measurement system has considerably more variation than the total part, or process, variation so it cannot adequately discriminate between good and bad parts.
  • The X-bar chart should have roughly the same pattern for all of the appraisers. It the patterns of the appraisers are very different, there might be some differences in their techniques that should be investigated.
• Checking for appraiser bias.
  • A separate control chart for the grand averages for all appraisers should also be set up to show test differences between the appraisers. Only the averages chart is used for this view of the measurement system.
  • If any of the grand averages fall outside the control limits, the we must investigate the differences between that appraiser and the others.

Scatter Plots for MSA

• With Scatter Plots, each measurement taken in the study is plotted by the part or sample and by appraiser.
  • This is a variation of the standard Scatter Diagram. We can use these plots to look at how consistent the appraisers are relative to each other and how consistent they were in measuring the same part/sample.

Whiskers Charts for MSA

• A Whiskers Chart plots the range of the measurements for each part and for each appraiser.
• This provides a view of how consistent each appraiser is in making measurements on the same part/sample. It can also be used to look at consistency between appraisers.