By now, you should be aware of the concept of “ndc” or number of discriminate categories that a measurement gage is capable of discerning. It is a statistically significant discrimination, or the capability of telling the difference between items even with the gage's intrinsic error. That is good to know when analyzing how a process varies. But, it can all go to waste very quickly if you do not measure frequently enough with that gage to detect process variation. In fact, “ndc” should be paired with “nds” or number of discriminate samples! For example, it your process varies in a manner that if you do not adjust it every hour, it will go out of control – and you only check it once an hour – your sampling frequency if so woefully infrequent that you will have no idea what your process is doing. A rule of thumb is there should be between 5 and 7 sampling points between and adjustments. That way, just as your gage has adequate resolution to detect your process variation, your sampling frequency will also be suitable. The error caused by insufficient sampling is related to measurement error - or the wrong use of a perfectly good gage. Otherwise, you are wasting your time.
So, ponder adequate sampling, or “nds”, to assure you are doing all you need to control your process!
So, ponder adequate sampling, or “nds”, to assure you are doing all you need to control your process!
