Mixture profiler not working

[#1]

I generated a mixture design DOE with three ingredients and 2 responses. With the responses, I made a fit model, where I used the macro "mixture surface response" and everything else as default. I saved the formulas in a new column, and I'm trying to look at a mixture profiler to see which portion of the space gives me the desired values for both responses. But nothing appears in the window, I just get the red triangle, no error message, nothing. even the debugger doesn't say anything. 

To make sure I'm not doing these (completely) wrong, I generated a mixture DOE where I entered dummy data manually, generated the fit models in the same way. The mixture profiler was working in this way.

Any idea what could be the issue? 

I'm running JMP 16.2.0

[#2]

I'm not sure what steps you are following, but I have attached your data table with the fit model script showing the surface.  When you run Fit Model and use the factors as Mixture Response Surface and both Y's, when you get the output, click on the red triangle associated with the response.  Select Factor Profiling>Mixture Profiler.  I added the script to your data table.  The profiler is at the bottom of the output.

[#3 · Accepted solution]

Actually, @statman 's profiler does not work quite right either since it does not maintain the mixture constraint of the ingredients adding to 1.

The problem is that you did not create your design in JMP. So, the metadata (the data about the data) is not stored in the table. Specifically, the A, B, and C components are not identified as mixture components in the Column Info. I added the Mixture column property to each column. That will ensure that the Mixture Profiler will appear and the Prediction Profiler will maintain the mixture sum of 1. Running your table scripts should verify that. 

If you create the design in JMP, you will also have a Design Role property that is set to Mixture. I did not add it here as it is not needed for the Mixture Profiler or the Prediction Profiler. 

[#4]

Thank you for the quick reply.

Indeed I did not follow the usual DOE design steps. I was not aware of this "Mixture" column property. It's good to see the solution is so simple.

[#5]

Thanks @statman 

good to have this option too. I didn't know I could run two fit models at the same time and also generate the mixture profiler in the analysis. It sure is a nice way to automate things!