Measurement & calibration
Torque-wrench results: accuracy needs an uncertainty budget
A small worked example separates calibration uncertainty, resolution and repeatability.
This is a measurement example, not a calibration certificate or a record of a TTC tool.
Known and assumed
Assume a 100 N·m indication, certificate expanded uncertainty 1.0 N·m at k = 2, resolution 0.1 N·m and repeatability standard deviation 0.20 N·m for one reading. Assume independent contributions and neglect other effects only for this example.
Calculate
The certificate standard uncertainty is 1.0/2 = 0.50 N·m. Uniform rounding over ±0.05 N·m gives resolution uncertainty 0.1/√12 = 0.0289 N·m. Root-sum-square combination gives uc = √(0.50² + 0.0289² + 0.20²) = 0.5393 N·m. With k = 2, expanded uncertainty is approximately 1.08 N·m.
Independent check
The combined standard uncertainty must be at least the largest independent contribution. Here 0.5393 exceeds 0.50. The relative expanded uncertainty is approximately 1.08% of 100 N·m.
Still unresolved
Bias correction, loading direction, adapter geometry, temperature, reproducibility and the decision rule can matter. An uncertainty interval is not itself a pass/fail tolerance. Use the applicable calibration method and certificate to interpret a real tool.
Last reviewed: Not recorded