Uncertainty of Measurement in Simple Terms
If a micrometer reads 10.002 mm, the true size is not exactly 10.002 mm. A statement such as 10.002 mm plus or minus 0.003 mm at 95 percent confidence tells you how far the true value could reasonably be from the reading.
Uncertainty is different from error. Error is the difference from the true value, which you cannot know exactly. Uncertainty describes how much doubt there is.
Main Sources of Uncertainty
- The reference standard and its calibration certificate.
- Instrument resolution and bias.
- Repeatability and reproducibility, as found in a Gauge R&R study.
- Temperature and other environmental effects.
- The part itself, such as form error or surface finish.
Each source is estimated as a standard uncertainty, combined, then multiplied by a coverage factor (often k = 2 for about 95 percent confidence) to give the expanded uncertainty.
Example
A bore has a tolerance of 25.000 to 25.021 mm and the measuring setup has an expanded uncertainty of 0.002 mm. A reading of 25.020 mm is inside the limits, but the true size could be up to 25.022 mm. If the plant applies guard bands, it will only accept readings up to 25.019 mm.
Why It Matters
If uncertainty is large compared with the tolerance, good parts get rejected and bad parts get accepted. Calibration laboratories must state uncertainty on certificates, and manufacturers should check that it suits the tolerances they measure. Calibration records and certificates can be kept in OperaHive calibration management software.