256936_en-US

liliconcom

Plot PC1, PC2,... vs. Experimental Run Number

Hello Team

 

I have data from 10 Experimental Runs of Semiconductor (Plasma) Process. How can I get Numerical PC1, PC2, ... values and plot them vs. Run Number? I want to see how PC1, PC2,...are trending (going up or down or flatten) at the end of last (for example 10th) run. I am using JMP 14.1.

 

Appreciate your help.

Thanks.

Pramod

Re: Plot PC1, PC2,... vs. Experimental Run Number

Several workflows come to mind. I'm presuming your responses are ordered in the data table in the experimental execution order AND you do not have a run order column in the data table. If so, try Analyze -> Control Chart -> Run Chart. Then put your responses in the Y window, click OK.

Re: Plot PC1, PC2,... vs. Experimental Run Number

I am having a bit oof trouble understanding what you want?  First what are PC1, PC2, etc?  Are these principle components?  Or Y's?  Are the experimental runs different treatment combinations?

Re: Plot PC1, PC2,... vs. Experimental Run Number

Hi Statman

Sorry. I thought PC is a standard language for JMP Community. Yes, PC means "Principal Component".  I want to plot PC1, PC2,... means they are "Ys".  Since I want to plot them against "Run Numbers" on x-axis to see the variation in PC1,..., I will be running same experiment/recipe in the Plasma Chamber again & again keeping every thing else constant. So the "Run-1" will have PC1-1 value, "Run-2" will have PC1-2  and so on. 

Hopefully, it is clear now. 

Thanks.

Pramod

 

Re: Plot PC1, PC2,... vs. Experimental Run Number

Clear as mud. I wouldn't assume anything on a public forum, especially acronyms. To make sure I understand... You have a time series of 10 runs within treatment combination of an experiment (repeats of the treatment).  What measures (y's) did you use to create the principle components?  If you use all of the measures from the 10 runs, you will only get 1 value for each PC for each treatment.  Perhaps you could the JMP data table?

 

As you can tell, I clearly do not understand what you are doing...

Re: Plot PC1, PC2,... vs. Experimental Run Number

Each experimental run will generate one Data Table with 100s of rows/columns and I can calculate PC1, PC2, ...values for each row of data table using JMP Analysis>Principal Component feature. I think I can manage using JMP  Graph Builder & other related feature. Thank for your help any how.

  

Re: Plot PC1, PC2,... vs. Experimental Run Number

OK, yeah none of that was explained up front.  Sorry for my confusion.  You can do simple IR charts (Individual, Moving range) for the within treatment analysis of stability, runs, trends, shifts, etc.