Attribute MSA with limited test samples


I have multiple parts that we use pin gage on the ID in a go/no-go manner. We are in the process of validating the test method. Unfortunately we don't have many parts on hand and making known BAD parts is limited as well.

Would designating a set of gage pins as BAD and another as GOOD to measure the same part or similar parts (assuming passing parts) a known strategy in this case? So for instance, on the go-inspection the operator should determine that a pin gage that measures within the spec of the part should go in, and a pin gage that measures bigger than that part spec should not go in, thus the inspection is determined passing. While in that same case the BAD gage pins would be opposite. Then for each inspection the good and bad pins would be swapped with the inspector not knowing.

I've only changed the samples to create the variance and never the gage itself. Any thoughts?

I've suggested maybe finding a worst case part and just biting the cost to have proper samples using those. As what I am historically used to doing.
Top Bottom