How to choose samples for Gage R&R - Micrometer to check tube thickness

Hello groups,

I am having doubts how to choose samples for GRR.
for e.g I am using a micrometer to check for tube thickness. So i cut random samples from to one particular lot of tube, asked operators to measure thickness at same spot in all tubes. The data i am getting is very same accross all operators. I mean there is not enough part variation. Also for information my gage resolution is 0.001 and the parts tolerance is from 0.025-0.035?

Similarly for GRR for a counter which measures length of tube. I made samples of same lengths, so when i asked operators for GRR I get same data from all.

I am really confused.

Overall should the samples used for GRR be same size or can I choose samples from different size tubes with varying thickness.

thanks for your help,
Robin

Gauge R & R

The samples should be representative of what your process puts out.

Your part tolerance could be to big compared to your actual measurements, which would cause you to flat line.

RGohil said:

Hello groups,

I am having doubts how to choose samples for GRR.
for e.g I am using a micrometer to check for tube thickness. So i cut random samples from to one particular lot of tube, asked operators to measure thickness at same spot in all tubes. The data i am getting is very same accross all operators. I mean there is not enough part variation. Also for information my gage resolution is 0.001 and the parts tolerance is from 0.025-0.035?

Similarly for GRR for a counter which measures length of tube. I made samples of same lengths, so when i asked operators for GRR I get same data from all.

I am really confused.

Overall should the samples used for GRR be same size or can I choose samples from different size tubes with varying thickness.

thanks for your help,
Robin

Click to expand...

Stick with the same nominal size of tube or you will obtain erroneous GRR results.

You may be in a situation where your gage has the required resolution for product inspection, but does not have enough resolution to see the process variation.

Can you attach the actual data?

Please analyse my GRR data and give your comments

Hello Miner and group,

Quite late in my reply, but here is the attached information for the tape GRR I have performed. Please note tolerance range is 0.034 inches. For % variation I am under 30% value i.e. acceptable to the customer but I have replacement measuring device on order that has better GRR. But for % tolerance I really exceed the 30% limit. The number of distinct categories is greater than 5 so that should be good.

Overall can I say the gage is acceptable for the time being till the alternative gage arrives.
But as I explained previously I have delibrately selected the samples with slight variation from one another even a few out of specs. If I had not done so I would have got all values the same i.e 90 same readings

Please give you comments on the GRR analysis.

thanks a lot
Robin

micrometer?

look like your gauge can only read to 0.005"

Yes it can read only 0.005, do you think this the reason for the poor %Tol value.

RGohil said:

Yes it can read only 0.005, do you think this the reason for the poor %Tol value.

Click to expand...

You made a very fundamental error in the structure of your study. You stated that you deliberately introduced product variation or you would have had the same results on all parts. This has a major impact on the results of the %GRR evaluation, artificially reducing %GRR and inflating the ndc. You should always select a subgroup of parts that reflect the same variation as the process itself.

After reviewing your data and factoring in your comments, your gage does not have adequate resolution to detect process variation. It really is not adequate for use as an inspection to tolerance gage.

Obtain the correct subgroup of parts and repeat the study with this gage and again with your new gage to obtain accurate results of the suitability of both gages.

Does the new gage have increased resolution?