Process Capability - Determining the process capability of certain equipment

jammy

Registered
Hi;
I have a question regarding determining the process capability of a certain equipment. So this equipment will tend to classify the raw materials based on diameter. At first, someone sent me the result but Cpk was determined based on the individual diameter reading of which I said it must not be. Since the equipment doesn't control the diameter of the raw materials it will just segregate it. The Cpk should be anchored on how will the machine was able to segregate it. So the process capability should be run based on % of being able to correctly segregate by the machine. Am I correct on this?
I attached here, the data based on the % accuracy in classifying by the machine. My question is even if all the % accuracy is within the allowable specs (95% to 100%), still the Cpk is below 1. So is this possible, that even if I have actual performance within 95-100% but still the process is not capable? I am confused on this.
Thanks for the help!
 

Attachments

  • data.xlsx
    10.2 KB · Views: 174

kuyakut

Involved In Discussions
IMO, determining the process capability of the machine is not feasible. The cpk result is not from the machine, it is a cpk result of the raw materiaI. The appropriate approach test method qualification.
 

AMIT BALLAL

Super Moderator
Hi;
I have a question regarding determining the process capability of a certain equipment. So this equipment will tend to classify the raw materials based on diameter. At first, someone sent me the result but Cpk was determined based on the individual diameter reading of which I said it must not be. Since the equipment doesn't control the diameter of the raw materials it will just segregate it. The Cpk should be anchored on how will the machine was able to segregate it. So the process capability should be run based on % of being able to correctly segregate by the machine. Am I correct on this?
I attached here, the data based on the % accuracy in classifying by the machine. My question is even if all the % accuracy is within the allowable specs (95% to 100%), still the Cpk is below 1. So is this possible, that even if I have actual performance within 95-100% but still the process is not capable? I am confused on this.
Thanks for the help!

Your post seems confusing. I thought it is about gage capability (Cg/Cgk), hence didn't post earlier.

There 3 different indices possible. Let me know which one you are referring to:
1. Machine capability (Cm,Cmk): To check whether is machine is capable of producing quality products
2. Gage capability (Cg,Cgk): To check whether the gage is capable for measurement
3. Process capability(Cp,Cpk): To check whether the process is capable of producing quality products

Now after re-reading I am assuming it is regarding process capability.
How did you calculate process capability for this data?
If your process is generating upto 5% defectives, the process is not capable. A capable process with 1.67 Cp can have 0.023% defectives that is 233 PPM or DPMO.
 
Last edited:

Ron Rompen

Trusted Information Resource
My understanding of the initial post is that you are trying to determine if the gauge is capable of correctly classifying (sorting) raw material, based on the OD. This would appear to me to be something which an MSA study is more appropriate than a capability analysis. The machine/gauge does not 'make' the product, only tries to detect and segregate it.
 

Brakeman

Involved In Discussions
I think the guys are right. I changed the data around to report percent mis-sorted and the 1 sided presumed normal chart shows the problem.
Process Capability - Determining the process capability of  certain equipment
 

Matt Savage

Trusted Information Resource
...So is this possible, that even if I have actual performance within 95-100% but still the process is not capable? I am confused on this.
Thanks for the help!
Adding my comments: Cpk is valid if the specs are correct (In this example, you are 'bounded' by 100, so I do not think it is appropriate to enter the USL = 100)
And yes, it is possible to have a good / great Cpk yet data is out-of-spec and conversely a poor Cpk, yet all data is in-spec.
Lastly, Cpk is only a valid statistic if the process is stable (in statistical control)
 
Top Bottom