Does the probe make proper contact?
4. Measurements in surfaces
Temperature of solid objects is often measured by attaching a sensor to the surface -complicated when the temperature of the object is very different from the temperature of the ambient environment.
Surface is the interface between the object and its environment (usually air) - if hot, there are large heat flows and temperature gradients.
Using a probe to measure the temperature of the interface is problematical because:
- the probe conducts heat from the surface and so cools it
- if it covers a significant area it will mask the surface and so change the temperature gradients
- the sensor in the probe is some distance from the tip, so is cooler than the tip
Minimise these errors by making the sensor temperature as close as it can be to the undisturbed surface temperature.
Special ‘surface probes’ are available aiming to increase the intimacy of the contact with the surface, but they inevitably entail compromises between accuracy and practicality. Errors are likely to be several tenths of a degree at best, rising to several degrees as the temperature increases.
Illustration of some of the options available
Option 1: apply a thin probe, to reduce the disturbance to the surface temperature. A thermocouple junction in the probe can be close to the tip, or even make direct contact with the surface if this is electrically acceptable. However, heat flows and temperature gradients (and hence errors) will be rather significant with this solution, and it is not recommended.
Option 2: for better contact the probe termination should be flat (for a flat surface) and have an enlarged area. The diameter of the probe stem should be as small as practicable, to reduce the heat it conducts away from the surface. The sensor should again be as close to the probe tip as possible, and in this case a thermocouple junction or a flat film-type Pt100 may be suitable. This solution entails greater disturbance to the surface temperature, and heat flow and temperature gradients are still significant problems.These two options may give results which are quite repeatable but of limited accuracy: a sensor will only measure its own temperature, and it is difficult to get it to the temperature of the surface in the presence of heat flows.
Some hotplates have been developed as calibration sources, though there is a difficulty in calibrating them (e.g. using an embedded sensor or an infrared thermometer). Investigations with such apparatus are needed if reliable offsets between surface and immersion calibrations are to be obtained.
Option 3: the problem of heat flow is much reduced if the probe makes contact for some distance along the surface, before breaking away. In a typical configuration, a thermocouple using fine wires or cable makes contact with the surface under an adhesive patch or another attachment The wires keep in contact with the surface for some distance before leading off to a remote indicator, so the region near the junction is close to the surface temperature.
This option is not always convenient, but for measurements of pipes up to about 100 °C, thermocouples in special metal bands, or even Velcro straps, can be wrapped around the pipe. In semiconductor processing, very fine thermocouple wires may be spot-welded to the wafer to ensure good contact with minimal conduction errors.
None of these options are a complete solution, and different compromises must be chosen for different applications. We are left with the basic fact that contact sensors are not well adapted to measure surface temperatures. Infrared thermometers are adapted to do this but have their own problems.
5 of 11