Gage R&R nested on Minitab

[ptitevy]

Hello!

This is my first time on this forum.
Let introduce myself.
I am a student and I am currently doing an intership in a international company. I am in the quality department.

I am here to implement SPC at least to start to understand how it works ) I found this forum after long research on Gage RR, and how to interprete Data. I already found many answers )

I have now data on a "destructive test".
We took one order and 10 samples were cut by 3 operators before being weight. This test was repeated 3 times, that means each operator cut 30 samples.

I have now the data, and I am wondering how I should enter them in Minitab 14.

Should I consider 10 same samples measured 3 times ?
Should I consider 30 samples measured 1 time ?


Thanks in advance for any helps.

[Jim Wynne]

Originally Posted by ptitevy

I have now data on a "destructive test".
We took one order and 10 samples were cut by 3 operators before being weight. This test was repeated 3 times, that means each operator cut 30 samples.

Welcome to the Cove!! Please explain "order," what is being cut, why it's being weighed, and what the value is that's being measured.

[ptitevy]

Originally Posted by JSW05

Welcome to the Cove!! Please explain "order," what is being cut, why it's being weighed, and what the value is that's being measured.

Hello!

Ok sorry for being so vague.

Order is laminated paper : a big laminator machine cover some paper with some Polyethylene (PE). The point is about reducing the waste of PE on the edge of the rolls.

So the edge of roll is taken to make the samples, that means that we have a strip composed of {paper with PE} and {only PE}. I don't know if I am really clear , but the width of the {only PE} (ie the waste of PE) should be as low as possible.

The long strip is split into small strips by the engineer.
Each trip is given to the operator who will cut the sample in order to keep the {only PE} and then to weight it !

Finally the value is the a weight of the waste of PE.
3 operators measured 3time 10 samples. But you can easily understand that the 90 are different even if they should be the same.

So how should I know enter the data ?

Thank in advance for any clarification!!

[D.Scott]

Welcome to the Cove.

I am wondering if you are actually doing a Gage R&R or if you are trying to determine the consistency of the process in an effort to reduce the waste PE.

The process you have described seems to me to end with the weight measurement of various pieces. If that is so, the test would not be destructive as the same pieces are still available to weigh (nothing happens to them after the first weight).

If, as I suspect, you are more interested in the actual weight of the waste PE samples, the process used to obtain the samples would be of more interest than the weight gage itself. This would indicate a need for a process study to determine the variation in the overlap of the PE. This would show if the excess were consistent enough to reduce the width while still ensuring there is always enough to cover the edge of the paper.

The experiment you are now conducting could establish the weight of the wasted PE but could do little about reducing the width of the PE with statistical confidence that it will actually cover the paper.

A deeper explanation of what you are trying to do would probably be helpful.

Dave

[ptitevy]

Hello!

Thank for your reply.

As I mention, I am a student in this company so forgive if I don't know deeply the subject. At first the company would like to implement SPC on package lenght, this is the reason I am here. Then it came out that I have to study about how gage RR works generally (and not only on package lenght where I was able to measure the gage RR crossed), in order to be able to implement on other parameters of the packaging.

That is why, I was given those data to interpret them.

To come back to the PE, it is not really a destructive test, but the operator is supposed to cut the sample themself. Yes of course nothing happens to the sample, they could be weight once 10 times, it should be the same.
May in this case it is more about measuring the consistency of the process AND the operator who will readjust the sample {PE + paper} to have the sample {only PE} which is weight.

------------
This is what we have now (data collected couple of time ago):
1/ 3 operators
2/ one lane
3/ 90 samples = 10 samples * 3 times * 3 operators
Am I clear ?
-----------
So may be those data are appropriated, and may a gage RR is not needed in this case...

I don't really know anymore, it sems that I am confusing now...

[D.Scott]

I don't think you are confused. It sounds like you are well on your way. I am having a slight problem understanding the need for a Gage R&R at this time. I am still thinking the consistency of the cutting process (how far from the paper is the PE cut before it is trimmed by the operators) is where you need to look to determine how you could save PE. Whatever amount is trimmed by the operator is waste so the project should be to determine how to reduce the excess PE - BEFORE - the samples are cut.

Am I correct in this assumption or have I misunderstood?

Dave

[Jim Wynne]

Originally Posted by ptitevy

Hello!

Ok sorry for being so vague.

Order is laminated paper : a big laminator machine cover some paper with some Polyethylene (PE). The point is about reducing the waste of PE on the edge of the rolls.

So the edge of roll is taken to make the samples, that means that we have a strip composed of {paper with PE} and {only PE}. I don't know if I am really clear , but the width of the {only PE} (ie the waste of PE) should be as low as possible.

The long strip is split into small strips by the engineer.
Each trip is given to the operator who will cut the sample in order to keep the {only PE} and then to weight it !

Finally the value is the a weight of the waste of PE.
3 operators measured 3time 10 samples. But you can easily understand that the 90 are different even if they should be the same.

So how should I know enter the data ?

Thank in advance for any clarification!!

It sounds like you're trying to evaluate the accuracy and repeatability of the weighing process before moving ahead in trying to reduce waste? If that's the case, it sounds like you're on the right track, or at least have your tasks in the right order, if you have reason to suspect that the weighing process might be a problem.

I still have a few questions. Is accuracy of the cutting of samples an issue, or only the weighing after the samples have been cut?
You wrote,

The long strip is split into small strips by the engineer.
Each trip is given to the operator who will cut the sample in order to keep the {only PE} and then to weight it !

What does cutting the long strip into smaller strips accomplish? Don't you want to know the total weight of the PE on the end of the roll? In production, who will be doing the sampling?
If you can provide some more information I think we can come up with an answer that will help with what you're trying to achieve.

[ptitevy]

Hello!

First thank for all the replies to help me

Originally Posted by JSW05

It sounds like you're trying to evaluate the accuracy and repeatability of the weighing process before moving ahead in trying to reduce waste? If that's the case, it sounds like you're on the right track, or at least have your tasks in the right order, if you have reason to suspect that the weighing process might be a problem..

Yes you are right. At the moment there is not any specifications to measure this waste of PE.
As far as I understood, the point is to reduce the waste, but it is not possible to remove totally this waste because of the edge effect (the {PE} is thicker).

Originally Posted by JSW05

I still have a few questions. Is accuracy of the cutting of samples an issue, or only the weighing after the samples have been cut?
You wrote,
What does cutting the long strip into smaller strips accomplish? Don't you want to know the total weight of the PE on the end of the roll? In production, who will be doing the sampling?
If you can provide some more information I think we can come up with an answer that will help with what you're trying to achieve.

Cutting into small strip by the engineer certifies the width of the strip.
Then the operator recut it and the point is "how far to the paper to cut?" We are here testing repeatability.

As I mentionned this measuring is not common, we need to test this method before do it in a regular frequency. That is why we were thinking about the gage RR nested...