The no-go gauge is the more sensitive of the pair, because it is the one that stops a bad lot, and it is also the one that produces the most false alarms. Before a lot is quarantined, three things should be ruled out in order: the gauge, the setup, and the part.
Verify that the gauge is the correct sheet and amendment, that it is within its calibration date, and that it has no damage on the functional surfaces. A no-go member with a flattened crest or a polished entry will enter a part that should fail. This takes a minute and resolves a surprising share of alarms.
Confirm the part is at room temperature, that the gauge is being applied without force, and that the operator is using the no-go member rather than the go member. Mixing the two is common when the pair is not clearly marked, and it produces a clean lot being rejected wholesale.
If the gauge and setup are sound, measure the failing feature with a comparator or a coordinate machine to see how far out it is. A small overshoot across the whole lot points to a tool that has drifted; a wide spread points to a process that is not capable; a handful of outliers points to handling damage after production. The three need different responses, and only the measurement tells you which one you have.
