Components of Variation not within or between

Brad Gover

Involved In Discussions
Hi all I have an interesting question. I am trying to determine how the variation in thickness in a machinied part has on a coating process's thickness. Both vary the total or overall thickness. Would this be as simple as
Coating variation Std Dev = √Total Std Dev2 – Machine Std Sev2

This works with within and between

Any thoughts?
 

Stijloor

Leader
Super Moderator
Hi all I have an interesting question. I am trying to determine how the variation in thickness in a machined part has on a coating process's thickness. Both vary the total or overall thickness. Would this be as simple as: Coating variation Std Dev = √Total Std Dev2 – Machine Std Sev2 ?

This works with within and between.

Any thoughts?

A quick bump!

Can someone with this expertise help?

Thank you very much!!

Stijloor.
 

Stijloor

Leader
Super Moderator
Hi all I have an interesting question. I am trying to determine how the variation in thickness in a machined part has on a coating process's thickness. Both vary the total or overall thickness. Would this be as simple as
Coating variation Std Dev = √Total Std Dev2 – Machine Std Sev2

This works with within and between

Any thoughts?

Another quick bump.

Can someone help Brad?

Thank you!!

Stijloor.
 

Jim Wynne

Leader
Admin
Hi all I have an interesting question. I am trying to determine how the variation in thickness in a machinied part has on a coating process's thickness. Both vary the total or overall thickness. Would this be as simple as
Coating variation Std Dev = √Total Std Dev2 – Machine Std Sev2

This works with within and between

Any thoughts?

Could you rephrase the question? "I am trying to determine how the variation in thickness in a machinied part has on a coating process's thickness" seems to have a word or words missing. Are you trying to determine how the variation in thickness of the part affects the coating thickness? Are you trying to establish some kind of correlation between part thickness and coating thickness? Can you give a specific example or examples?
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
Souds like you are trying to study the variation in the original thickness, plus the variation in the coating thickness to get the variation on total thickness.

An equation like you suggested in the original post would ONLY work if the original thickness and the coating thickness are independent of each other. If for example, a bump in the original piece makes an even thicker bump to form of coating, then you've got to consider the covariance between the two thicknesses. Or even the coating may "smooth out" bumps which would also be a dependence.

Keep in mind the equation would be:

(std dev of original thickness) squared + (std dev of coating) squared = (std dev of total thickness) squared

for the independent case.

How much data do you have? I'd look at SPC to see if the three thickness measures are each in control, and also a scatter plot of the original thickness vs the coating thickness.
 
Top Bottom