Help with NDC (Number of Distinct Categories) Please

[breitman]

Hi Everyone,

I have to perform gage R&R studies. For some reason, i can not get the ndc where it should be, let alone over 1. I understand that parts needed to be taken over a period of time, which has been done, but our numbers do not differentiat. This has become quite frustrating. Any advice would be greatly appreciated.

 

[bobdoering]

Re: Help with NDC Please!!

ndc assumes that you presented to the study the entire range of variation you expect that characteristic to see over the life of the process. Then, it compares the resolution the gage is capable of (with error factored in, not just readability), and determines if there is enough statistically significant resolution. Low ndc usually means there is not enough resolution for the variation presented, or too much error. Not sure which you have.

 

[Barbara B]

I understand that parts needed to be taken over a period of time, which has been done, but our numbers do not differentiat.

This is more a requirement for process capability. To get a clear picture whether your measurement system/process is able to distinguish different parts in the whole area of your characteristic you have to choose different parts. If all parts are similar, ndc will remain 1 - regardless of the measurement uncertainty.