Interpreting Minitab Gauge R&R Results

[velan26]

Hi,

Any body will help me to intrepret the Guage R&R results.

 

[Marc]

Re: Intrepreting Minitab Guage R&R Results

You will have to post your data/results if you want anyone to be able to help you.

 

[velan26]

Re: Intrepreting Minitab Guage R&R Results

Hi marc,


From the attached result, what i could conclude?


Advance Thanks....Velan

 

[Barbara B]

Re: Intrepreting Minitab Guage R&R Results

Some clues drawn from the plot:

Bar chart on the upper left side:

• Variation due to measurements is way too high: green bar Gage R&R approx. 70% (should be below 20-30%)
• Repeatability / Equipment variation (approx. 60%) is too high: You could not seperate different parts due the measured value
• Reproducability / Appraiser variaion (approx. 20%) is high: Different operators will measure different results for the same part

R-Chart in the middle of the left side:

• Measurements of the second operator (Sanil) have a smaller variation than the measurements of the first and third operator (except for part 10)
• Measurements of the third operator (Sebastian) have a higher variation than the others

Boxplots in the middle of the right side:

• Variation for the three operators is different
• Sanil's measurements tend to be smaller and Sebastian's measurements tend to be higher than those of Shine babu.

All in all your measurements system is too uncertain to give you a clear picture of the process outcome since 70% is above 30% (required maximum Gage R&R value). But this statement is only valid, if the selected parts in your analysis represent the whole process spread! They should be chosen deliberately for a MSA.

Best regards,

Barbara

 

[Allattar]

Re: Intrepreting Minitab Guage R&R Results

Random part selection from the process is another alternative to help obtain a representative set of parts for a gage study.
%study variable as Barbara indicates is only a measure of how acceptable the gage is at measuring the samples in the study. If the samples do not represent the population/process then the %study variable tells you very little about the gages use in practice. Random selection or careful selection to get a representative set of samples helps.

You can also include Tolerances, and historical standard deviations with the Gage Study. They are a useful additions to the analysis. This can be performed from the options sub dialogue.

%Tolerance is then asking can you measure to specifications.
%Process is based off of the historical standard deviation from a process, and then asks how acceptable the gage is to historically obtained process information.

Both values can also be a good indicator of have you chosen a representative set of samples to test as well.

 

[velan26]

Thank you mr.barbara and allattar.


Now i got some insight..

 

[Barbara B]

Thank you mr.barbara and allattar.

You're welcome. And it is Ms Barbara