Gage R&R (GRR) Study - Please comment on the results

[Shahzad67]

I just finished conducting GR&R Study. Please comment on the results.

Thank you

[Miner]

This raises a lot of questions.

1) Please verify the data entry on the following cells: B8, F4, F6, F9, F10, F11, F13. These data points are having a very large impact on the results.

2) You have large differences between operators, both on average and in variation. Study the methodology of Operator S (lowest variation), and replicate that method for the other two.

3) There is inconsistency between your stated gauge resolution and the actual data entered (0.003 versus 0.0001)

What is the purpose of this gauge (inspection or SPC)?

[Shahzad67]

Miner,

First of all thank you for taking time and providing me with feedback. I have attached a revised GR&R study data sheet. This study was conducted with a different part but by same operators.

The gage we used was the Vernier Caliper with accuracy of .001 inch.

The part had nominal dimension of 0.118" and +/_ 0.005". USL of 0.123" and LSL of 0.113".

The purpose of this guage is inspection. I was trying to understand operator variance in the inspection method.

S

[Miner]

Sorry, I have been very busy and have not been able to respond. Please refrain from posting the same question in multiple threads. If no one has responded after several days, it is acceptable to "bump" the same thread by asking again.

There are two basic issues: The first, which hides a lot of other issues, is the repeatability variation. What type of product/feature is being measured?

There is also a reproducibility issue. Operator S is definitely different from the other two. Again, additional information would help.

[Shahzad67]

Miner.

Thank you for your reply.

The product being measured is PVC extrusion profiles I have attached a drawing of a frame. As you can see there are quite few critical dimension marked on the profile.

Yes, operator S is different then others two operators. We have anywhere from 45-50 operators who are checking these profiles coming off from lines.

Do I need to do more tests? Or does this test tell me that I need to find a better method for measuring the products.

Thanks,
S

[Miner]

Your attachment is missing. Without seeing the thicknesses involved, will the part be deformable under pressure? Or, does it have location sensitive features (measurement size is highly dependent on location of measure)?

[Shahzad67]

I was unable to load the drawing as it have loaded it on my previous thread. Here is the link. The part is rigid PVC.

http://elsmar.com/Forums/attachment....7&d=1267497213

[Miner]

I could not find the dimensions on this drawing, so I don't know what the probable issues may be. Do you section the product in order to measure it, or measure the cut ends? Any chance of distortion from this?

There may also be within-part variation. I have experience with measuring rubber extrusions. Once we tried a Perceptron laser measurement device and were able to see regular variations in size that corresponded to the flights on the extruder screw and the speed of the screw.

In addition, rigidity is relative. Even PVC may flex some depending on the location of measurement and the force applied. Calipers can apply significant force, and are highly sensitive to perpendicularity.