DoE - testing parameters from different process stages

[#1]

Hello,

A DoE has been performed on a process. The process included step A, step B, step C and step D.

The DoE tested two parameters from step B and two parameters from step C in one single design. All parameters were continuous.

I am wondering if next time it would not be better to do 2 separate DoEs: one on step B and one on step C.

Is there any principle in DoE that says that one should avoid mixing parameters from different steps of a process?

Thank you for your support.

Best regards,

Claire

[#2 · Accepted solution]

Hello @cbaril,

Separating 2 process steps in 2 designs may decrease the information and hide the interactions between your process steps.

Mixing different parameters from different steps of a process sounds a good idea, as the input of step C might be influenced by the output of process B. Due to dependancies between your process steps, taking into account interactions between your process steps and the different sources of variability sounds reasonable.

This Mastering webinar may help you : Optimizing Processes Using Designed Experiments - JMP User Community
Hope this first answer will help you, 

[#3]

Thank you Victor for your prompt response! I was wondering because the results could not be successfully verified, and as idea had to reduce the complexity of the design to get simpler models.

[#4]

Why/how wasn't you results successfully verified ?

Do your process steps have too much variance to screen significant effects ? What was the problem ?