Need to check the effect of all parameters like pressure, flow rate, power, deposition time by changing +-10% individually, keeping other parameters constant.

[#1]

I want check the effect of parameters like first I want to change pressure by +-10% keeping other parameter a fixed value.

And the change in flow rate by +-10%, again keeping other parameters constant.

How would I do that?

[#2 · Accepted solution]

Welcome to the forum.  You won't like may advice, but you are doing OFATs (One factor at a time manipulation).  This is an inefficient and often ineffective method for understanding factor relationships.  Primarily:

1. Narrow inference space

2. Inability to estimate interaction effects

3. Unrealistic (you can't possibly hold ALL factors constant and change 1)

4. Sub-optimal (you select optimum settings are conditional (i.e.,  based on settings of other factors).  If you find better settings for other factors, your conclusions about the first factor are moot))

5. Costly (number of runs to optimize is significantly greater than DOE)

I suggest you design a multi-factor experiment instead.  You have described a 4 factor experiment. This could be accomplished in as little as 8 treatments (Res IV).

[#3 · Accepted solution]

Hi @Satyam,

Welcome to the Community!

I couldn't agree more with @statman, you should run a DoE to have more information and certainty about the effect of parameters. I'm not sure about your objective : do you want to screen some important parameters effects, or have you already found a optimum and you want to check that this optimum is robust regarding small variations of your factors ?

If you have already found an optimum and would like to check the robustness of this optimum with small variation of the factors around it, I would suggest creating a simple design like @statman was suggesting, but adding several centre points to test the experimental variability of your optimum. This would enable to test the pure experimental variability of your process thanks to centre points replicates, as well as the impact of your factors.
See some idea about robustness designs in this conversation : Robustness design 

You may also find some information in these ressources :
https://www.nist.gov/document/robdesgn-1pdf
https://www.itl.nist.gov/div898/handbook/pri/section1/pri12.htm

Hope this answer will help you,

[#4]

Yes @Victor_G , I have already optimized the conditions. I just need to check the effects of each important parameters individually, which can affect the outcome. For now, I'm not wondering for the interactions of two parameters. 

I don't care whether it's inefficient or not. I just want to know that can we do this in JMP in a single chart? @statman 

I don't want to make different charts for each parameters. 

I would really appreciate the guidance. Thank you

[#5]

Ok, then the links provided in my previous post about robustness designs may help you. You don't need to run a complex design to check robustness and adequacy of your optimum, as @statman wrote, a simple 8-runs fractional factorial design could do the job (+ centre points to confirm your optimum and check its reproducibility).

Of course, you can always continue your work with an OFAT approach in JMP, by creating your datatable manually, fixing all parameters but one and then changing the parameters values one by one.
But as written previously, OFAT approach will lead you nowhere with several factors. The biggest risk you may face with OFAT when checking your system would be to falsely believe you have found the optimum when it's not. OFAT can only confirm your opinion, not contradict your understanding of your system, leaving no possible option for improvement. OFAT is not appropriate for several experimental reasons listed by Statman as well as representativeness of the experiments done in the experimental space studied.

If you're new to DoE, there are several ressources listed here to help you get started : https://community.jmp.com/t5/Design-of-Experiments-Club/DOE-Resources/td-p/635455

Hope this answer may help you,