Important Notice: Phoenix Contact South Africa does not supply or specialise in thermocouples or temperature sensors as implied by this article, which was sourced from Phoenix Contact USA.
Some of the basic technical issues that engineers should consider when applying thermocouples.
More than 60% of all industrial temperature measurement applications in the USA use thermocouples. Despite their widespread use, there are many misconceptions about thermocouples.
Why use thermocouples?
Some common reasons for using thermocouples are:
Although the thermocouple is a simple device, its small voltage signal is easily corrupted and wiring one requires care.
How thermocouples work
Heat and electrical energy are related when it comes to electrical conductors, including thermocouple wires. A common misconception is that the thermocouple junction creates the voltage signal, similar to a battery. This is not true; rather, the signal is generated along the entire length of wire where there is a temperature gradient.
If you’ve ever held one piece of a wire and then heated the other end, you discovered that the heat moves up the wire. Heat moves toward the cold end because the higher kinetic energy of hot atoms imparts some of their energy to their colder neighbours, making them vibrate. The atoms vibrate faster toward the hot end. They don’t move, however, because the solid structure holds the nucleus in place (see Fig. 1).
Fig. 1: When one end of a wire is heated, the kinetic energy
(heat) travels the length of the wire.
Some of the outer electrons in metals are free to actually move, not just vibrate. As the electrons become hot, they crowd toward the cold end, helping to spread kinetic energy (heat) more effectively than other materials – an example of heat conduction. Electrons have a negative charge, so their movement also creates an electrical property resulting in positive potential at the hot end and a negative potential at the cold end. The magnitude depends on the composition of the wire and the temperature difference between the hot and cold ends, not length or gauge size of the wire.
Thermocouples are made of two dissimilar wires connected at one end, called a hot junction, where the temperature measurement will take place.
Anywhere else those wires are connected – including terminal blocks and measuring instruments – is called a cold junction. A thermocouple circuit (see Fig. 2) can have more than one cold junction, but this is strongly discouraged and, as we will see, easily avoided. One cold junction (the place where the thermocouples connect to the measuring instrument) is unavoidable. A process called cold junction compensation takes place at the instrument. By using another temperature sensor which monitors the temperature of the cold junction, it is possible to cancel out the voltage created there mathematically.
Since electrons move at different rates in the two different metals, it is possible to measure the voltage potential (or difference) between them. The raw voltage signal that represents temperature is a sum of all the hot and cold junctions. We can determine the temperature of the hot junction by subtracting the effect caused by the cold junctions.
One chooses which type of thermocouple to use based on temperature range and environmental conditions. Four types of thermocouples account for more than 99% of all applications: K, J, T and E (see Fig. 3).
Achieving accurate thermocouple measurements
Thermocouples create a very small voltage signal (often less than 30 μV/°F). Noise can easily corrupt this signal. Use a transmitter to convert the low-level microvolt signal into robust process signal such as the 4 – 20 mA signal. These transmitters have become very cost-effective, and this low price even makes them practical in non-critical applications.
In process thermocouple applications, which have runs of many hundreds of feet, it is common to use a transmitter that fits into a thermowell probe, which is basically a closed metal tube. The “hockey puck” style transmitter is well suited for hazardous areas (such as Class I, Div. 2) common in process applications. The transmitters are mounted in metal probes with metal covers and connect securely to metal conduit that meets the various encapsulation or explosion-proof requirements.
Fig. 2: Example of a thermocouple circuit.
The conduit fittings are convenient to protect the wires mechanically, even where intrinsically safe circuits are used. The hockey pucks are generally output loop-powered, meaning that they get their energy from the 4 – 20 mA signal sourced by a power supply back in the control cabinet (see Fig. 4).
DIN rail-mounted transmitters are most popular in automation applications such as sealing operations on packaging machines. It is reasonable to run the wires back to a control cabinet or junction box for conversion because the length of wire needed is often just a fraction of that found in process applications.
Embedded thermocouples are often used inside machine components, often with tape or bolt-on connections.
It is very important to consider electrical isolation when deciding on a transmitter. On loop-powered devices, isolation between the thermocouple input and the output circuit prevents ground loops from degrading the temperature signal. Typical isolation values range from about 1000 – 2000 V. The transmitter will need additional isolation if an external DC power supply is the power source.
Three-way isolation will provide protection between input/output, input/power supply and output/power supply. If the thermocouple touches a high-voltage source in the control cabinet, this will provide additional protection for the measuring device.
Best practices: thermocouples in the field
If it is not possible to use a transmitter in the field or on the machine, another option might be to extend the length of the thermocouples by using special terminals mounted in the thermowell or surface-mounted on the machine. The area where they are connected is away from the extreme process temperatures and should be similar to the temperature at the controller.
Using a single length of thermocouple wire can be impractical, so thermocouples are wired to these terminals. They are specially constructed to minimise any temperature difference between their two ends and between each individual wire, and are sometimes referred to as isotherm blocks.
Isotherm blocks allow the Law of Intermediate Metals to operate. This law states that the addition of a third metal in both wires of the circuit won’t degrade the signal if all of the connection points on the isotherm blocks are of equal temperature. Connected to the other side of those terminals is the “extension” or “compensation” wire.
Fig. 3: Recommended maximum ranges for 14 AWG thermocouples.
Technically, extension wires have the same chemical composition as their thermocouples, while compensation wires have a different composition. In practice, though, it seems that the term “extension cable” is used most often, regardless of the wire composition. Extension cables are often half the price of the thermocouples themselves and are available in sizes up to 14 AWG.
The large wire size reduces the loop resistance in long runs. They work because, at temperatures between 0 and 200°F, they exhibit the same electrical properties as the thermocouples themselves and are electrically indistinguishable from the thermocouples. The outer insulation jacket on the extension wires is different to those on the thermocouples to prevent extension wire from inadvertently being used as thermocouple wire. The individual wire insulation colour matches those of the thermocouples.
Convention for both thermocouples and extension wires calls for a P to be used to indicate the positive leg and an N for the negative. Extension wires are also designated by X. So, if JP is written, it refers to the positive leg of a Type J thermocouple, and KNX refers to the negative leg of a Type K extension cable (see Table 1).
Best practices: thermocouples in the cabinet
Back at the control cabinet, extension wires are run into some sort of distributed control system, PLC or analyser. Ideally, the user would run one continuous piece of wire from the hot junction to the terminals on the measuring instrument. But we are discussing practical considerations and this is often not practical.
Cabinets built off-site and by third-party companies can’t be wired directly. Therefore, terminal blocks are used to facilitate easier wiring. The panel shop wires from the controller to the terminal block, leaving the other side open for the field connection of extension wire.
Unfortunately, many engineers believe they can simply connect thermocouples to standard terminal blocks without degrading the signal, using the Law of Intermediate Metals as the rationale. The contact area and current bar in most terminal blocks is made of steel, or in the best quality ones, nickel-plated copper. In both cases, this material doesn’t match the chemical composition of the thermocouple, so a cold junction is formed.
As mentioned earlier, you could theoretically use terminal blocks without making a cold-junction correction, but only if the temperatures on both sides of the blocks remain the same or change at exactly the same rate.
| Type | Composition | Conductor insulation | Outer jacket | Temperature range |
| KPX | Nickel chromium | Yellow | Yellow | 0 to 200ºC (32 to 392°F) |
| KNX | Nickel aluminum silicon | Red | ||
| JPX | Iron | White | Black | 0 to 200ºC (32 to 392°F) |
| JNX | Copper nickel | Red | ||
| TPX | Copper | Blue | Blue | 0 to 100ºC (32 to 212°F) |
| TNX | Copper nickel | Red | ||
| EPX | Nickel chromium | Purple | Purple | 0 to 200ºC (32 to 392°F) |
| ENX | Copper nickel | Red |
This isn’t likely in control cabinets because they are packed with heat-producing equipment, fans and other devices that block airflow and heat the interior unevenly. Standard terminal blocks are not designed to minimise temperature differentials like isotherm terminals do. This introduces unpredictability to the measurement whenever a heat-generating device is switched or the cabinet is opened.
The solution is to use thermocouple blocks whose metal parts match the composition of the thermocouple extension wire. There is no cold junction because the materials match. It looks like a continuous piece of extension wire to the thermocouple circuit. The errors won’t be introduced if the temperatures change. While thermocouple blocks cost more than a standard terminal block, they eliminate the potentially tedious source of errors, making them worth the extra cost.
Simple temperature control
Resources in the main controller, whether PLC, PC or DCS, may not be needed to perform temperature control. In some cases it may be preferable to “outsource” or distribute that control to another device.
Today, even low-cost products provide excellent solutions. For example, the DIN rail-mounted temperature transmitters mentioned earlier convert the low-level thermocouple signal to a 4 – 20 mA one. It also isolates the circuit from ground loops electrically. Some allow for on/off control with programmable dead bands using a transistor or relay output and cost $300 or less.
Sophisticated control and bus systems
Other devices measure several thermocouples at once and have sophisticated control algorithms that allow them to exercise more advanced control than simple on/off. Proportional, integral derivative (PID) control has been used for decades in process industries and is rapidly finding applications in the discrete world. Temperatures must be controlled extremely rapidly and precisely in applications such as heat-seal packaging and extrusion processes. In these applications, heat has to be added and removed in fractions of a second.
Some PID controllers accept thermocouple inputs. They accept up to eight thermocouple inputs, and the control program resides within its plastic body. There are also outputs that can be used to trigger a heating or cooling control action. In this case, the temperature data isn’t converted into a high-level analogue signal – it goes directly to a digital value.
Fig. 4: Example of a thermocouple application.
There are three very useful things about bus-enabled devices:
Common problems and solutions
There are a few problems that show up regularly in thermocouple applications, including:
Conclusion
Despite the prevalence of thermocouples, engineers still encounter many problems when using them. In the past, temperature measurement might have been too expensive or complicated. In other cases, the transmission distance of the application caused difficulties. However, engineers can simplify and enhance their thermocouple design by understanding the basics of thermocouple theory and taking advantage of lower priced accessories.
Contact Sheree Britz, Phoenix Contact, Tel 011 801-8200, sbritz@phoenixcontact.co.za
Important notice: Phoenix Contact South Africa does not supply or specialise in thermocouples or temperature sensors as implied by this article, which was sourced from Phoenix Contact USA.
