Pressure gauge temperature error

Pressure gauge temperature error is the change in a gauge's indication error associated with a change from its specified temperature conditions, while the pressure comparison is otherwise controlled. It belongs to the complete Bourdon tube pressure gauge response, not merely the number of scale divisions. Distinguish an instrument-temperature effect from a real change in the pressure being measured.

A concrete manufacturer specification is WIKA's June2025 US data sheet, which gives an additional error bound of $\pm 0.4\%$ of span per $10\mathrm{K}$ departure from the $20{}^{\circ }\mathrm{C}$ reference.^[ref] This is a version-specific bound; it is not a measured signed correction for an individual gauge.

For an original gauge with span $P_{\mathrm{span}}=1\mathrm{MPa}$ at $40{}^{\circ }\mathrm{C}$, the specified form givesAt a pressure reading of $100\mathrm{kPa}$, that bound is $8\%$ of the reading. The percentage is referenced to span, so the same absolute bound represents a larger relative contribution near the bottom of the range. Without the sign and an appropriate measurement calibration, subtracting $8\mathrm{kPa}$ would not be justified.

For an original controlled comparison, hold reference pressure at $400\mathrm{kPa}$ and allow the gauge assembly to equilibrate first at $20$ and then at $40{}^{\circ }\mathrm{C}$. Suppose indications are $401$ and $407\mathrm{kPa}$. Their errors relative to the reference are $1$ and $7\mathrm{kPa}$, so the observed temperature-associated change is $6\mathrm{kPa}$, or a local average $0.3\mathrm{kPa}{\mathrm{K}}^{-1}$. These are illustrative readings, not manufacturer test results. Repeat at other pressures and on return to the reference temperature to check reversibility and hysteresis.

Record gauge and reference conditions, equilibration and uncertainty in the pressure comparison. Heating a sealed gas can change its actual pressure through the ideal gas law; an uncontrolled change of that kind cannot be assigned entirely to instrument error. The bound or fitted correction remains one contribution to the final measurement uncertainty.