Gage R&R for Instant Measurement Machine

[qualityguy]

Hello,

I'm beginner to G R&R and working upon collecting the data for an Instant Measurement Machine. This machine captures 2D image for the part and measures the all type of linear values which is done using a optical comparator and hand tools. It also calculate GD&T values as well.

My main target is for evaluating the true position as it isn't being repeatable from time to time.

Would the attached template help me for performing Gage R&R for inspection machines. In this scenario, even though I evaluate the machine with two operators would the attached template be ideal.

Kindly advise!

 

[optomist1]

Please clarify, pics/specs etc always make the process easier...." Instant Measurement Machine... machine captures 2D image", help me out here please clarify....optomist1

 

[qualityguy]

Thanks for reverting! Please see attached image for the measuring machine. I hope this would answer your question

 

[UncleFester]

Are you trying to use an attribute study but using variable data? I'm not sure what you're trying to determine. If you're trying to determine reproducibility of a figure for true position, then you'll need to perform a full GR&R. If you're trying to determine pass/fail, then an attribute study will work fine, but first you would need to determine what constitutes a pass or a fail. I think that the table you propose is not adequate for the type of study you need to perform.

 

[optomist1]

qualityguy, thanks for the attachment....what you call an "IMM", appears to be a form of an Optical Comparator...below is a link to a prior Elsmar Cove post, hope this helps...

GR&R (Gage R&R) for an Optical Comparator or Vision System

 

[qualityguy]

Thanks for sharing the hyperlink to the discussion.

I have performed a variable study and see attached image for results. However, I'm not sure what should I consider for the tolerance because on the blue print they had mentioned as j|n.004m|A|B|C|

I had considered the tolerance as .004 and see attached image for results. Do suggest if I need to change the tolerance ad it's maximum material condition as well.

Thank you!

 

[Welshwizard]

Hello,

It would appear that you can measure variable data with this machine and the output is in the form of 2D Coordinates.

You are interested in true position which, as i'm sure you are aware is a computation based upon the variability of two inputs, the square root of the X and Y variables.

My advice would be to initially narrow down the characteristics of interest, otherwise you may be conducting lots of studies which will take a lot of your valuable time and the causes of any variability you see in your results may well be correlated.

Conduct consistency studies on the characteristics of interest by selecting one trained or familiar operator and repeatedly measuring the same part at least 20 or more times. Note down any factors such as a change of temperature during this process. Plot the results in time series order for each X and Y output on a Individual moving Range Charts.

If the Upper control limit on any of these charts breached you have an inconsistent measurement process which will require investigation and fixing after which the study should be repeated. You haven't mentioned the datuming process, this could also be a prime cause of excessive variation during these studies.

Once the process is consistent you can with confidence chart, in the same way, the true position and characterise the process for repeatability and any consumption of specification. If this is acceptable, then it will be worth performing a crossed Gauge R&R safe in the knowledge that its possible for the process to be consistent and adding operators should not in theory add knew unforseen issues.

Gauge R&R Studies can take a lot of time and planning, there is little point in this if the process does not behave and perform with one operator measuring one part multiple times. Debug in the simplest way and then if required add the multi level studies with confidence.

I hope this helps you.