I am using ProcessBook. I want to create what I know as an XmR chart. With this chart the control limits are calculated from the moving range for each datum in the set (subgroup = 1). This unlike an X-barR chart that uses subgroups of size greater than one.
A minimum of two points (n=2) is needed to calculate a moving range. When n=2, the moving range is calculated for each point as the difference between itself and the reading preceding it. For n=3, the moving range is the range of values of the current point and the two preceding points.
The average of those moving ranges becomes the basis for the control limits on both the chart of the individual data as well as the moving range chart. The value of n determines the values of multipliers for the average moving range that ultimately determine the individuals control limits (d2) and the moving range control limits (D3 and D4).
When I edit the chart definitions to specify the chart type, I am confused to find Individuals and a Moving Range types listed independently. I learned you always use them together, so the fact you would have to create them separately is strange.
I forged ahead nonetheless, creating two SQC charts of each type against the same tag. I now have other questions with the chart definitions:
- (Sample tab) My calculation basis is time. I am using the first reading of each calendar day, so my Calculation Period is 24:00:00 and my Start Time of Sample is 00:00:00. The Sample Size is not editable and is equal to 1 and thus the Sample Period is also not editable. I therefore conclude sample size parameter is the subgroup size.
- (Control Parameters tab) I have selected to calculate control limits using Average Range Estimate. However, I note I cannot specify the value of n. Why not? The help file acknowledges there are different values for d2, D3, and D4 depending upon the size of the moving range, and explicitly provides the d2 values for a two-period moving
range. What value is it using for n and d2?
- (Data Filter tab) I find I am able to specify the minimum number of data points to generate a value Sigma. What does it mean, minimum? Does that mean it is going to keep recalculating the control limits with the addition of each new point? Starting when? Starting with my Plot Start Time? What if I want to display older data against
the current control limits? How do I exclude them from the calculation while still displaying them? What if I want it to stop calculating new limits but continue to plot new data?
Moving Range Chart:
- (Sample tab) I am unable to select time as a calculation basis to match the individuals chart. I can only use an event calculation basis. Therefore, I cannot properly pair this mR chart with my individuals chart.
- (Sample tab) The default sample size is two, but is editable higher. I guess for this type of chart this becomes n, which means it would affect the values of D3 and D4. The help file recommends a value of eight (8). However, since I can’t do the same with the individuals chart I cannot create properly paired charts.
- (Control Parameters tab) The method of sigma calculation is not editable because this chart type is the Average Range Estimate method by definition.
- (Data Filter tab) Ditto Item #3 for Individuals Chart.
- (The chart itself) If I specify an eight-period moving range (n = 8), I should not see a value for mR until the eighth point in stream. However, I see points starting immediately on the left edge of the chart. I guess the sample size on the Sample tab is not actually n. If not, what is it? If so, why does the chart show values for a moving range where n=8 before a full eight readings are available?
In summary, calculation of control limits for an individuals chart requires analysis of the moving range and the calculation of the average moving range, which also determines the control limits for the moving range chart. Why doesn’t the individuals chart come with its matching moving range chart? How do I create the pair?