In semiconductor and semiconductor-adjacent data, we often have multiple sources of variation in our data. For example, lot-to-lot variation, wafer-to-wafer variation, and site-to-site variation. In creating control charts, one point is often one lot. That one lot is usually an average, with the multiple points measured from that single lot average being across multiple wafers and/or sites. In traditional Xbar control limits calculations, the within-subgroup variation is used to calculate the limits for the Xbar chart. But wafer-to-wafer and site-to-site variation can be utterly unrelated to the lot-to-lot variation. So control limits calculate from wafer-to-wafer plus site-to-site variation can look absolutely ridiculous on an Xbar control chart where each point is 1 lot. I have seen real data examples using traditional Xbar limit calculations with almost all of the points being marked as OOC because the limits were so tight.
It would be much better if the limits on the Xbar control chart in this example situation were based on the lot-to-lot variation.
As far as I know, this is only currently possible in JMP17 using Analyze>Quality and Process>Legacy Control Charts>Presummarize. I am afraid that the Presummarize is on the chopping block ("Legacy") for a future version of JMP. Either keep it as a platform in JMP, or create a new platform that has the same capability of calculating control limits from the variation on the same chart.