Thermal Dissipation
Metrological parameters that describe the temperature rise of a sensor per unit of electrical power dissipated within it determine the measurement uncertainty in temperature sensors. The self-heating coefficient characterizes this thermal behavior, usually expressed in milliwatts per degree Celsius. It indicates the power required to raise the internal temperature of the sensor.
Measurement Drift
When an excitation current is applied to a resistance temperature detector or a thermistor, the power dissipated as heat causes the sensor temperature to rise above that of the medium being measured. This temperature offset appears as a positive error in the measured temperature. In precision applications, this error can be larger than the specified accuracy of the sensor itself, which makes it necessary to minimize the excitation current or calculate the offset and correct for it.
Calibration Compensation
The effect of self-heating can be minimized by utilizing a pulsed excitation current or a very low continuous current, typically less than one milliampere. Calibration laboratories measure this coefficient by recording the sensor resistance at two different current levels under identical environmental conditions. The calculated self-heating coefficient can then be utilized to correct the measurement results or to establish the maximum safe operating current for the sensor in the field.
Operational Boundary
The value of this coefficient is not constant and depends heavily on the thermal conductivity of the surrounding medium and the flow rate of the fluid. A sensor placed in a moving liquid will have a much higher self-heating coefficient than the same sensor placed in still air, because the liquid is much more effective at dissipating the heat. This variation means that the coefficient must be determined under the specific installation and operating conditions where the sensor will be utilized.
If the sensor is mounted in a stagnant pocket or a thermowell with poor thermal contact, the actual heating will be much higher than predicted by the datasheet, leading to unexpected measurement offsets.
