In part 1, we covered the basics behind what thermocouples are and how they work. In part 2, we’ll cover how thermocouples can be made and know when to select the right thermocouple for your project.
Thermocouples can be made of any two dissimilar metal wires, and their emf voltage depends on the composition of the chosen metals. However, what makes thermocouples so popular is that the materials used to construct them are restricted and their output emf’s have been standardized. Certain materials and combinations are better than the others, and some have basically become the standard for given temperature ranges. Table 1 lists some of the available thermocouples in the U.S. market.
Thermocouple Type | Material Composition | Temperature Range | Uncertainity | Color Code |
T | Cu vs. Constantan | -250 to 350°C | Greater of 1°C or 0.75% | Blue-Red |
K | Chromel vs. Alumel | -200 to 1250°C | Greater of 2.2°C or 0.75% | Yellow-Red |
J | Iron vs Constantan | 0 to 750°C | Greater of 2.2°C or 0.75% | White-Red |
R | Platinum vs. Platinum-13% Rodium | 0 to 1450°C | Greater of 1.5°C or 0.25% | None Established |
S | Platinum vs. Platinum-10% Rodium | 0 to 1450°C | Greater of 1.5°C or 0.25% | None Established |
C | Tungsten 5% Rhenium vs. Tungsten 26% Rhenium | 0 to 2320°C | Greater of 4.5 °C to 425°C, 1% to 2320°C | None Established |
E | Chromel vs. Constantan | -200 to 900°C | Greater of 1.7°C or 0.5% | Purple-Red |
Table 1: Different types of thermocouples
Selecting the right type of thermocouple for an application depends on many factors. These include sensitivity, temperature range, corrosion resistance, linearity of output voltage, and cost. For example, types R and S are relatively expensive and are not sensitive. However, they perform well at high temperatures up to 1768°C and are resistant to a number of corrosives. Type C thermocouples are suitable for higher temperature applications, but they are relatively expensive and corrode easily in an oxidizing environment. A Type T thermocouple is inexpensive and very sensitive, but will corrode at temperatures above 400°C. Type K is very popular for general use, relatively inexpensive, reasonably corrosion-resistant, and can be used at high temperatures, up to 1372°C. K-type thermocouples also have provide relatively linear output as compared to the other types [2].
The actual magnitude of the thermocouple emf is very small, and is in the order of few millivolts. At a given temperature, Type E has the highest output emf among common types, but this voltage is still measured in millivolts. The sensitivity of thermocouples is also relatively low. For instance, the voltage change per degree Fahrenheit from 38 to 93°C is only 36 microvolts. As a result, thermocouples require accurate and sensitive measuring devices and cannot be used for temperature changes of less than about 0.1°C. Traditionally, expensive voltage balancing potentiometers were used to measure emf. Today, a high quality digital voltmeter is sufficient [3].
The National Institute of Standards and Technology (NIST) has developed standard calibration curves for determining temperature based on the measured emf voltage. These data represent the output emf of thermocouples when an ice cold junction is used, and are incorporated in the memory of most DAQ systems. Â Unfortunately, the temperature-voltage relationship of thermocouples is nonlinear and curve-fitted using polynomial functions. Obviously, the higher the order of the polynomial function, the higher the accuracy of temperature reading. The polynomial function should only be used inside the temperature range of the thermocouple type and should not be extrapolated. To save computational time, a lower order polynomial fit can be used for a smaller temperature range.
Thermocouple wires come in variety of sizes. Usually, the higher the temperature, the heavier should be the wire. As the size increases, however, the time response to temperature change increases. Therefore, some compromise between response and life may be required.
Thermocouples can be connected electrically in series or in parallel. When connected in series, the combination is usually called a thermopile, (whereas there is no particular name for thermocouples connected in parallel). A wiring schematic of a thermopile combination is shown in Figure 3.
Figure 3 – Series-connected thermocouples forming a thermopile
The total output from n thermocouples will be equal to sum of the individual emf’s. The main purpose of using a thermopile rather than a single thermocouple is to obtain a more sensitive element. Parallel connection of thermocouples is used for averaging.
References:
- Omega Temperature Handbook, Omega Engineering, Inc. 2nd edition.
- Wheeler, A. and Ganji, A., Introduction to Engineering Experimentation, Prentice Hall Inc., pp. 240-246, 1996.
- Beckwith, T., Marangoni, R., and Lienhard, J., Mechanical Measurements, Fifth Edition, Addison Wesley Longman, pp. 676-685, 1993.