Every number written on an inspection sheet contains two things mixed together: the true size of the part and the error of the measuring process. When a stamped bracket's hole reads 8.52 mm, some of that reading is the hole and some is the caliper, the person holding it, and how the jaws sat in the hole. We call the spread caused by the measuring process measurement variation, and the spread caused by real differences between parts part variation.
The trouble is that the inspection sheet does not tell you which is which. If readings bounce around, you might blame the press when the caliper is the culprit, or trust a stable-looking process that is really being smoothed by a sloppy gage. Measurement system analysis, or MSA, is the set of methods used to separate the two. Gage R&R is the most common MSA study for variable gages, meaning gages that give a number rather than a pass or fail.
A quick sanity check before any formal study is to measure one part several times and look at the spread. If that spread is a large fraction of the tolerance, the gage is already in trouble. Tolerance here means the allowed range from the lower specification limit to the upper specification limit.
Standard Procedure
- 1Identify the characteristic, the gage, and the tolerance width.
- 2Measure one part repeatedly, at least five times, without looking at previous readings.
- 3Compute the range: largest reading minus smallest.
- 4Compare the range to the tolerance width as a first warning sign.
Lesson Summary
- ✓Readings always include measurement error mixed with real part differences.
- ✓A quick repeat-measurement check on one part reveals whether a formal study is urgent.