2026-11-16
Every working gauge in a factory is a copy of a copy. Its size came from a setting master, which came from a gauge block set, which came from a national standard. Each transfer in that chain introduces uncertainty, and the discipline of a gauge room is mostly about making sure the uncertainty stays small enough that the gauge is still useful at the bottom of the chain.
A part measured with the working gauge carries the uncertainty of every link above it. That is why the answer to "how accurate is this gauge" is never a single number.
The working rule is that the calibration uncertainty of a gauge should be no more than one quarter of the tolerance it is used to inspect. On a part with a 0.02 mm band, the gauge's uncertainty contribution should be within 0.005 mm. On a 0.05 mm band, 0.0125 mm.
The rule exists because uncertainties combine. If the gauge contributes a quarter of the band and the measurement process contributes another quarter, the inspection is still meaningful. If the gauge contributes half the band, a part at the middle of tolerance has a genuine chance of being called good or bad on the gauge alone, and the resulting argument cannot be settled with more inspections — only with a better chain.
| Part tolerance | Gauge uncertainty budget | Chain that fits |
| 0.20 mm | 0.050 mm | Grade 2 blocks, shop-floor gauge |
| 0.05 mm | 0.0125 mm | Grade 1 blocks, calibrated setting master |
| 0.01 mm | 0.0025 mm | Grade 0 blocks, comparator, temperature control |
Steel expands by about 11.5 micrometres per metre per kelvin. On a 100 mm gauge, a 5 K difference from the reference temperature moves the reading by roughly 5.75 micrometres — more than half the uncertainty budget of the 0.01 mm row in the table above. The reference temperature is 20 degrees C, and the rule that follows is that the gauge, the master and the workpiece must all be at 20 degrees C, not merely "in the room".
A 100 mm steel gauge and a 100 mm aluminium workpiece at the same air temperature are not the same size. Aluminium expands at roughly twice the rate of steel, so a workpiece that soaked at 25 degrees C and a gauge that soaked at 20 degrees C differ by about 11 micrometres on 100 mm before either is measured. Soak time matters more than air temperature: a 100 mm steel block needs roughly an hour to settle after a 10 K change, and a large casting needs far longer.
Gauge blocks are wrung together to build a size. A properly wrung stack has an interface film of the order of 0.01 to 0.02 micrometres, which is negligible on most work but should be in the budget when the stack is long. Two practical rules:
A calibration certificate that says "within tolerance" is not traceability. The record needs the measured value, the uncertainty, the reference temperature, the standard used and the identification of the master. Without the measured value, a drift trend cannot be built, and drift is what predicts the next failure.
Two habits make the chain auditable:
For the artefacts themselves, the gauge and test equipment range includes setting masters and wear-check accessories; the site index links the calibration accessories used with them.
Contact: Eason Wang
Phone: +86-13751010017
Tel: +86-755-33168386
Add: 1F Junfeng Building, Gongle, Xixiang, Baoan District, Shenzhen, Guangdong, China