The response is the thickness of the vinyl that is produced. You want to maximize thickness. A lower limit for thickness values is 10.
The whole plot factors are the rate of extrusion (extrusion rate) and the temperature (temperature) of drying. These are process variables and are hard to change.
The subplot factors are three plasticizers whose proportions (m1, m2, and m3) sum to one. These factors are mixture components.
1.
Select DOE > Custom Design.
2.
Double-click Y under Response Name and type thickness.
To add factors manually, follow step 4 through step 11. Or, to load factors from a saved table, select Load Factors from the red triangle menu next to Custom Design. Open the Vinyl Factors.jmp sample data table, located in the Design Experiment folder. If you select Load Factors, skip step 4 through step 11.
4.
Type 2 next to Add N Factors.
5.
Click Add Factor > Continuous.
6.
Rename these factors extrusion rate and temperature.
7.
Click Easy and select Hard for both extrusion rate and temperature.
This defines extrusion rate and temperature to be whole plot factors.
8.
Type 3 next to Add N Factors.
9.
Click Add Factor > Mixture.
Responses and Factors Outlines
11.
Click Continue.
12.
Click Interactions > 2nd.
13.
Click OK to dismiss the informative message.
14.
Type 7 next to Number of Whole Plots.
15.
Type 28 next to User Specified.
Note: Setting the Random Seed in step 16 and Number of Starts in step 17 reproduces the exact results shown in this example. In constructing a design on your own, these steps are not necessary.
16.
(Optional) From the Custom Design red triangle menu, select Set Random Seed, type 123686, and click OK.
17.
(Optional) From the Custom Design red triangle menu, select Number of Starts, type 10, and click OK.
18.
Click Make Design.
Design Outline
Note that the whole plot factors, extrusion rate and temperature, are reset seven times in accordance with the levels of the factor Whole Plots. Within each level of Whole Plots, the settings for the mixture ingredients, m1, m2, and m3, are assigned at random.
The Vinyl Data.jmp sample data table contains experimental results using a design created in a previous version of JMP.
1.
Select Help > Sample Data Library and open Design Experiment/Vinyl Data.jmp.
Fit Model Window
The factor Whole Plots has the Attribute called Random Effects (&Random). This specifies that the levels of Whole Plots are random realizations. They have an associated error term.
The analysis method is REML (Recommended). This method is specified precisely because the model contains a random effect. For more information about REML models, see Restricted Maximum Likelihood (REML) Method in the Fitting Linear Models book.
3.
Click Run.
Split-Plot Analysis Results
The Parameter Estimates report shows that the three mixture ingredients, as well as the extrusion rate*m3 interaction, are significant at the 0.05 level.
The REML Variance Component Estimates report indicates that the variance component associated with Whole Plots is 2.476748. This is 38.838% of the total variation. It follows that the error term associated with whole plot replication is smaller than the residual (or within-plot) error term.

Help created on 9/19/2017