Using JMP to Streamline Six Sigma DMAIC Process
Created:
| Last Modified:
Six Sigma Projects require many files to Define, Measure, Analyze, Improve and Control (DMAIC) your process. Use JMP for Six Sigma process improvement and manage your data, results and all related Six Sigma files within a JMP Project.
See how to:
- Use a JMP Project (.jmpprj) file to organize a Sigma Sigma project using the DMAIC methodology.
- Link external files in a JMP Project for easy access.
- Include images and notes so that you can remember your reasoning for what was done, and you are prepared for a toll gate review at a moment's notice
- Understand the DMAIC acronym (Define, Measure, Analyze, Improve, Control)
- Define the problem and goal
- Measure the process
- Optionally incorporate Measurement Systems Analysis into the design
- Set up and analyze a classic two-level factorial design and fit a model to the data.
- Use Monte Carlo simulation within JMP to estimate the long-term process results while considering the variation in the controllable factors
- Estimate the improved process capability based on simulation results using a JMP Capability Explorer Add-In
- Compare initial process to improved process
Questions answered by @MarilynWheatley , @JerryFish and @scott_allen at the webinar:
Q: JMP calculates Cpk ,so why did you need to calculate it on the side?
A: JMP will calculate all the different capability indices depending on the spec limits that you have set in the column properties.
Q: How did you create the fishbone diagram in JMP?
A: You can find it in JMP under Analyze>Quality and Process>Diagram. JMP Help explains how to use it.
Q: Does the EMP MSA still exist in JMP 17 & 18?
A: Yes.
Q: Can you summarize the Improve steps for me?
A: I first fit a model to the data that I collected as part of my DOE. Then I reduce the model to include only terms that are important for controlling the moisture loss. And then using my fitted model, I use the red triangle and ask JMP to maximize my desirability.
Q: How do you define desirability?
A: When we set up this design in JMP, I entered my response is the moisture loss. My objective was to minimize, or find the lowest values that I can get for, moisture loss. So, when I maximized desirability, I am using the Prediction Profiler to identify controllable input values that are going to get me as close as possible to my objective - lowest moisture loss. Behind the scenes there's an equation that's associated with this and JMP tries to maximize that equation to maximize desirability and show it to you on the interactive Profiler.
Q: How did you get the 9 or data points of AFTER?
A: I could have baked more cakes, but for some problems that is not the best solution. JMP lets you use a Monte Carlo simulator. That generated my AFTER results. I told JMP what was going to vary in my inputs and indicated that I want 10 rows simulated to make a new table, then added the results to the table with the BEFORE data.
Q: What is the meaning of Block and Center Point and when can we use them?
A: Typically you use a block to account for variations attributable to nuisance factors. In my example, I will bake 12 cakes over 2 days, and during this time there may be a difference because of the moisture in the air or the ambient temperature on different days. I used a block to account for any difference that could be attributable to the days. You might introduce a center point run to test if there is curvature in the response surface. My factors were set at a high and a low level, so with 2 values, I can just draw a straight line. But, if I have some observations in the middle of that line that are higher or lower, then I would need more of a curve to model that. And so that's why I introduced the center points. See documentation.
Q: Can we see the equations for the Prediction Profiler?
A: Yes. You can see the prediction expression for the equation by looking for it in the Red Triangle menu. Or you can save the equation to the data table and view the formula.
Q: How do you launch a new Project?
A: File>New>Project
Q: The Add-In in requires the use of a spec limit which is usually user defined. But JMP calculated control limits based on the moving average. Can we use the calculated spec limit?
A: And spec limits are not the same as control limits. Control limits are calculated from the data, whereas specification limits are defined by you or your customer.
Q: How is the Recent Observations Chart generated?
A: The recent observation chart was built using Graph Builder and then appended to the report.
See how to use the Capability Explorer Add-In:
Resources