Hi,
I'm looking for some feedback. I am only indirectly involved in Gage R & R's, CPK studies and the like through other departments in our Company. I have what should be, I think, a relatively easy question to answer.
We have a Customer who has a tolerance 'band' 0f 0.03 MM's. Is this tolerance simply too tight to expect to get a good Gage R & R? We are using a Vision System, which I'm told has a resolution of 0.004 MM's. Aren't we already out of luck when considering the 10:1 rule for tolerance value to measurement precision?
THanks,
Scott

Sorry for the speeling errors....not enough sleep lately! :)

Maybe, maybe not. The 10:1 "rule" is not really a rule, but a guideline that originated before MSAs. The only way to tell for sure is to actually perform the MSA study. Your gage resolution is in a range that could go either way. Another factor is how the gage will be used (for inspection or process control). Normally, it is more diffficult to meet the requirements for process control than for inspection.

Maybe, maybe not. The 10:1 "rule" is not really a rule, but a guideline that originated before MSAs. The only way to tell for sure is to actually perform the MSA study. Your gage resolution is in a range that could go either way. Another factor is how the gage will be used (for inspection or process control). Normally, it is more diffficult to meet the requirements for process control than for inspection.

Hello Miner,

Once the MSA is completed, would the ndc (number of distinct categories) provide the answer? Just wondering.....

Stijloor.

Guys,
Thank you for the quick response. We've got a heck of a Gordian's Knot that we're dealing with....
As I said, I'm really one of the 'unwashed' in this area. I DO know that from a plastic injection moldmaking perspective, which is my area of toil, the tolerances requested on plastic products are getting tighter and tighter - which of course translates into the steel. Although, this question IS dealing with the plastic product measurement itself. Where does it end and how do you tell a Customer, look, you're just asking too much from plastic to be 'capable' over a 5-7 year production run.
Is this where we apply the dreaded 'Ring' Method? (RNG - Random Number Generator...)
Great, now you're probably going to kick me off of the Forum!
:)
Scott

Guys,
Thank you for the quick response. We've got a heck of a Gordian's Knot that we're dealing with....
As I said, I'm really one of the 'unwashed' in this area. I DO know that from a plastic injection moldmaking perspective, which is my area of toil, the tolerances requested on plastic products are getting tighter and tighter - which of course translates into the steel. Although, this question IS dealing with the plastic product measurement itself. Where does it end and how do you tell a Customer, look, you're just asking too much from plastic to be 'capable' over a 5-7 year production run.
Is this where we apply the dreaded 'Ring' Method? (RNG - Random Number Generator...)
Great, now you're probably going to kick me off of the Forum!
:)
Scott

Hello Scott,

No, you will not be "kicked" off the Forum! People at The Cove love challenges! If you care to hang in there, you'll see what I mean. I'm sure that other MSA experienced Covers will join soon. Weekends are kinda quiet here, but wait 'till Monday!

Stijloor.

Hello Miner,

Once the MSA is completed, would the ndc (number of distinct categories) provide the answer? Just wondering.....

Stijloor.

The Range chart is the best assessment of this. The rule for the Range chart is that there must be at least 5 possible ranges, including zero, that lie beneath the Upper Control Limit. There is also a maximum % of zero ranges, but I'll have to get back to you on that because I don't have my reference material with me.

ndc does not directly answer the question about gage resolution. The reason for this is that poor resolution can work two ways on an MSA. If it results in too many zero ranges, it may make the MSA look good in the numbers. The Range chart will show the truth with too many zeroes and several out-of-control Ranges that must not be ignored. The poor resolution may also inflate the Ranges making the numbers look bad. In this case, the Range chart will show the reason for this.