Gage R&R Studies For Small Set of Data With Varied Known Value

I

ikhsansiro

Hi

I am conducting an experiment to determine how good/precise one of my machine (that basically measure the coating thickness) measuring our parts. It’s comes to a confusing when the result that I got from Excel is far different compare from the result by using Minitab. I realized that’s maybe because of Minitab using the methods of ANOVA in calculating GR&R while the spreadsheet (template that I got from this forum, Thanks!) is use different method.
At one point, I am wondering whether the data that I have in hand is sufficiently meets the minimum requirements to do GR&R studies. To give you the clearer picture of my experiment, here is the background


Background
No of appraisers: 2
No of trials: 3

I want to carried out an experiment which known value (spec) is varied from one to another. There is total of 10 spec being follow (250u”, 300u”, 330u”, 350u”, 380u”, 400u”, 430u”, 450u”, 500u”, 530u”) and every spec consist of 2 samples. I know that’s impossible to do GR&R that gathered all this data into one since the spec. limit is different from one to another, so I made a decision to do GR&R by each spec which I think more practical and the problem is I have lack of data’s. Please refer attach for my GR&R. result (I am not sure how good/correct the result that I got)

Ex; For spec 250u”
Appraiser: 2
Trials: 3
No of part: 2
Total Data: 12


  • Is it really good enough to do GR&R?

  • What is the minimum requirement before doing GR&R studies?

  • What is the best method to be followed if we are only having small sample size of data?

  • Besides doing GR&R, what is the best method in analyzing data to measure the capability of the machine/Gage (especially for a set of data with varied spec limit as I got)?
 

Attachments

  • GRR & Linearity Ni-P.xlsx
    1.3 MB · Views: 236
I

ikhsansiro

Hi

I am conducting an experiment to determine how good/precise one of my machine (that basically measure the coating thickness) measuring our parts. It’s comes to a confusing when the result that I got from Excel is far different compare from the result by using Minitab. I realized that’s maybe because of Minitab using the methods of ANOVA in calculating GR&R while the spreadsheet (template that I got from this forum, Thanks!) is use different method.
At one point, I am wondering whether the data that I have in hand is sufficiently meets the minimum requirements to do GR&R studies. To give you the clearer picture of my experiment, here is the background


Background
No of appraisers: 2
No of trials: 3

I want to carried out an experiment which known value (spec) is varied from one to another. There is total of 10 spec being follow (250u”, 300u”, 330u”, 350u”, 380u”, 400u”, 430u”, 450u”, 500u”, 530u”) and every spec consist of 2 samples. I know that’s impossible to do GR&R that gathered all this data into one since the spec. limit is different from one to another, so I made a decision to do GR&R by each spec which I think more practical and the problem is I have lack of data’s. Please refer attach for my GR&R. result (I am not sure how good/correct the result that I got)

Ex; For spec 250u”
Appraiser: 2
Trials: 3
No of part: 2
Total Data: 12


  • Is it really good enough to do GR&R?

  • What is the minimum requirement before doing GR&R studies?

  • What is the best method to be followed if we are only having small sample size of data?

  • Besides doing GR&R, what is the best method in analyzing data to measure the capability of the machine/Gage (especially for a set of data with varied spec limit as I got)?


Anyone? :confused:
 
Top Bottom