Design of a setup for reproducible calibration of air thermometers

A European interlaboratory comparison (ILC) for comparing methods of air thermometer calibrations was launched in 2019. The comparison was organized in three loops with 7 or 8 thermometers circulated within each loop. Justervesenet (JV) acted as the link laboratory, and developed a flexible, stable setup for calibration of sensors in air, where the most important design constraint was to ensure comparability between the three loops. The design consists of a purpose-built inner chamber that fits inside a commercial climate chamber, with a volume large enough to accommodate any of the circulating probes. The chamber consists of two concentric cylinders separated by a gap of around 15 mm. The inner cylinder, which encapsulates the probes under test, has a perforated bottom with meticulously placed holes to allow controlled air flow into the measuring volume. A stable flow is established quickly, from around 10 cm above the bottom. The air speed is controlled by a fan placed at the top, whose rotation speed is controlled with a motor placed outside the main climate chamber, coupled to the fan via a long shaft. This ensures that heat generated by motor is not interfering with the air temperature control. Temperature and humidity controlled air is fed into the gap between the cylinders. The setup homogenizes the air flow in the measuring volume and ensures consistent conditions for all probes. Furthermore, a rigid support helps to ensure a reproducible placement of both sensors under test and reference sensors. Under operation 7 reference probes were used in various locations in the measuring volume, and 2 sensors placed in the wall of the inner chamber. The walls are constructed from aluminium, which provides a good compromise between machineability, cost, weight and thermal properties. We present here the results of extensive characterization of the setup, including the temperature uniformity, flow conditions, and the effect of self-heating on the reference probes in various locations inside the chamber. We claim that the setup provided the necessary stability and performance to justify the loop linkage with JVs measurements.