Low pressure differential and gauge calibrations require a calibration setup that is different as compared to its higher pressure counterparts.
- Difference #1: Limitations on generating really low pressures by the primary standard
- Difference #2: Minor changes in the pressure within the calibration setup can cause a significant offset at low pressure points
- Difference #3: Minor leaks in the calibration setup can play a significant role in calibration results
Deadweight testers are often the designated primary standards for calibrations. The lowest pressure generated by a deadweight tester is limited by the mass of the piston and its area. To perform gauge and/ or differential pressure calibrations, two deadweight testers floating at the nominal pressure are required, which adds the uncertainty on the pressure as well as reference side of the differential sensor.
CPB5800 Deadweight Tester
Standard calibration setups allow the pressure side of the system to come to equilibrium and the reference side of the system to come to its own equilibrium. This can cause two different areas of the system to read different pressures, which is especially significant at low pressures like zero point adjustment.
At atmospheric or high pressure calibrations, minor leaks can be ignored and are often compensated by the calibration setup; however at low pressures, these minor leaks can cause pressure gradients.
Solutions for performing stable low pressure calibrations
Robert Clayton, director of research at Mensor, performed detailed experiments to highlight these problems and recommends changes to the calibration setup with detailed uncertainty analysis in his white paper. For more detailed guidance, read "Low Pressure Measurements," which Clayton wrote following these experiments. The paper reviews the typical calibration setup for low pressure gauge and differential calibrations, provides experimental data using a 10 in. H20 pressure sensor calibration with the traditional and improved setup, and provides recommendations and conclusions from the results.
Related Reading:
