A widely used rule of thumb sorts %GRR into three zones. Under 10 percent, the measurement system is acceptable for the job. Between 10 and 30 percent, it is marginal: it may be used, depending on how important the characteristic is and what it would cost to improve, but improvement should be planned. Over 30 percent, it is not acceptable and should not be used to make decisions about the process until it is fixed. These are guidelines, and the criticality of the feature can move the line.
When the result is marginal or unacceptable, look at which component dominates. If EV is the larger of the two, the problem is repeatability: check gage resolution against the tolerance, look for jaw wear, loose parts, dirt, burrs on the hole edge, and whether the part can be held the same way every time. A gage whose smallest reading step is more than about a tenth of the tolerance will struggle regardless of the operator. If AV is the larger, the problem is reproducibility: write a one-page method, agree on jaw pressure and measuring position, train all appraisers together, and repeat the study.
Rerun the study after any fix. A study is a snapshot of the gage, the method, and the people on that day. Treat the result as a reason to act, not as a label the gage carries forever.
Standard Procedure
- 1Place %GRR in a zone: under 10 acceptable, 10-30 marginal, over 30 unacceptable.
- 2Compare EV with AV to see which source dominates.
- 3For EV problems, examine gage resolution, wear, fixturing, and part condition.
- 4For AV problems, standardize the method and train all appraisers together.
- 5Rerun the study after the fix and record the new %GRR before releasing the gage.
Lesson Summary
- ✓The zone tells you whether to act; the EV-versus-AV comparison tells you what to act on.
- ✓Every fix is confirmed by rerunning the study, never by assumption.