Getting Zero for GR&R Reproducibility?

SPC_Newbie · Mar 22, 2012

I'm in need of some help. I'm sure I'm doing something wrong as I am getting ZERO for my reproducibility results in a Crossed GR&R Study (ANOVA). I am attaching the raw data and the output text from Minitab. This has happened twice today and I am just not seeing what I'm doing wrong. Thanks for taking a look!

Miner · Mar 22, 2012

You did not do anything incorrectly. This appears to be an artifact of the apparent gauge resolution. There are a very large number of measurements at 0.00020". Try a dotplot by operator. There are two odd observationbs that are not consistent with the apparent gauge resolution.

Note: I used the term apparent resolution because this is the limit of how the measurements were recorded (with 2 exceptions). If the indicator is capable of higher resolution measurements, try the study again using the full resolution.

Xylaine · Jun 1, 2015

How about the attached? I also get zero for reproducibility...please advise. thanks.

Miner said:
You did not do anything incorrectly. This appears to be an artifact of the apparent gauge resolution. There are a very large number of measurements at 0.00020". Try a dotplot by operator. There are two odd observationbs that are not consistent with the apparent gauge resolution.

Note: I used the term apparent resolution because this is the limit of how the measurements were recorded (with 2 exceptions). If the indicator is capable of higher resolution measurements, try the study again using the full resolution.

Miner · Jun 2, 2015

Is this gage used for SPC or part inspection? How were the parts selected? They do not appear to be randomly selected from production.

Xylaine · Jun 2, 2015

The parts are purposely selected to cover the range of product, some parts are over the control spec to cover upper & lower spec limit.

Miner · Jun 3, 2015

Miner said:
Is this gage used for SPC or part inspection?

The part selection is acceptable if this gage is used only for inspection, but is not acceptable if the gage is used for SPC.

The zero repeatability is an artifact of the gage resolution. There should be a minimum of 5 measurable levels within the UCL of the Range chart, which would require a resolution of 0.0005. This would be important if this gage were to be used for SPC, but if it is used for inspection, it is probably unnecessary as your %Tolerance is pretty good.

Getting Zero for GR&R Reproducibility?

SPC_Newbie

Attachments

Miner

Forum Moderator

Xylaine

Attachments

Miner

Forum Moderator

Xylaine

Miner

Forum Moderator

Similar threads