Phoenix Contact, Americas R.B.U.
Thermocouples aren’t exactly the latest fad in automation and control. They’ve been in use for a long time, work well, and seem pretty simple in operation. More than 60% of all temperature measurements in the U.S. use thermocouples.
There are good reasons to use thermocouples over other temperature sensors. Your facility may already use them. The application may need a sensor that withstands a lot of physical stress or one that is physically small. The expected high, low, or span of temperature may exceed the limitations of other sensor types. Finally, it might be difficult to justify a higher price for more sophisticated devices.
Perhaps the most common misconception about thermocouples is that the signal is created by the thermocouple junction, like some sort of miniature battery. A thermocouple voltage is actually generated along an entire length of wire that has a temperature gradient. A useful concept to keep in mind concerns the relationship between heat and electrical energy within electrical conductors including thermocouple wires. As a demonstration, hold a single piece of wire on one end and then heat the opposite end. You’ll quickly discover that the heat moves up the wire. It does so because the hot atoms impart some of their kinetic energy to their colder neighbors.
The more kinetic energy an atom has, the faster it vibrates. Atoms with the highest kinetic energy are at the hot end of the wire, so they vibrate the most. While the nucleus of the atom remains trapped within the solid structure of the wire, some outer electrons are free to move. The vibrating atoms force free electrons toward the cold end and spread the kinetic energy in the form of heat conduction.
Because electrons have a negative charge, their forced migration creates a positive potential at the hot end of the wire and a negative potential at the cold end. The magnitude of the voltage depends on the composition of the wire and the temperature difference between the hot and cold ends. The length or gage size of the wire has no effect on the voltage generated.
Although there is electron movement in any single conductor, it is virtually impossible to get a voltage measurement from it. Readers will recall voltage is the difference in electrical potential between two points. Trying to take a voltage reading from a single point on the wire just can’t be done.
To get around this limitation, a second wire made from a different metal or alloy attaches to the first wire at the high-temperature point. The connection is called the “hot junction.” Different metals produce different quantities of electron motion when the wires are heated to the same temperature. The imbalance between the number of electrons creates a potential difference (voltage) between the wires.
Any other point those wires connect, including terminal blocks and measuring instruments, is called a “cold junction.” One cold junction at the place thermocouples connect to the measuring instrument is unavoidable. This is often called the reference junction. Temperature readings made at the reference junction correct any thermocouple voltage generated there in a process called cold junction compensation. Thermocouple circuits can have more than one cold junction, but that is strongly discouraged and easily avoided.
The raw voltage signal that represents temperature is the sum of all the hot and cold junctions. The temperature of the hot junction is determined by subtracting the offsets generated by all of the cold junctions. Thus the output voltage of the thermocouple depends only on the temperature of the hot junction.
The materials used to create the hot junction determines the type of thermocouple. The material type to use for a given application depends on temperature range and environmental conditions. Four types of thermocouples, identified by the letters K, J, T and E, account for more than 99% of all thermocouple applications.
The voltage thermocouples generate is small often less than 30 μV/°F. Such low potential is easily corrupted by noise. Typically, the microvolt signal feeds a transmitter that converts it into a robust process signal, such as a 4-to-20-mA instrumentation current loop. Transmitters have become economical with some priced in the $150 range. The low cost lets them serve in noncritical applications.
It’s fairly common to use transmitters that fit in a thermowell probe when connecting thermocouples with runs from several hundreds to thousands of feet. Thermowells are closed metal tubes that extend into the region where temperature measurements are made. The hockey-puck shaped transmitter fits in the exposed end of the well while the the rmocouple extends into the other end.
Hockey-puck style transmitters work well for hazard-classified areas. The transmitters mount in metal probes with sealed covers and connect securely to metal conduit that meets the various encapsulation or explosion-proof requirements. Intrinsically safe circuits are used to power the device while conduit fittings protect the wires mechanically. Hockey-puck transmitters are generally output loop-powered, meaning they are powered from a supply back in the control cabinet.
DIN-rail-mounted transmitters are most widely used in automation applications, such as those in sealing operations on packaging machines. The length of wire needed is often a fraction of that found in process applications. So it is reasonable to run the thermocouple wires back to a control cabinet or junction box for conversion. Also, thermowells found in the process industries may not be well suited for physical installation on a machine. Embedded thermocouples with tape or bolt-on connections are often used inside machine components.
Electrical isolation is important to consider when deciding on a transmitter. Isolation between the thermocouple input and output circuit on loop-powered devices prevents ground loops from degrading the temperature signal. Typical isolation values are in the 1 to 2-kV range. Transmitters that get their operating power from an external dc power supply need additional isolation. Often called three-way isolation, it provides similar levels of protection between input/output, input/power supply, and output/power supply. Isolation also provides a degree of protection for the measuring devices in the control cabinet in case the thermocouple somehow touches a high-voltage source.
If a transmitter is impractical, then it may be possible to extend the length of the thermocouple wires by using special terminals mounted in the thermowell or surface-mounted on the machine. The connection area should be away from extreme process temperatures. In fact, temperatures should be close to that of the controller. Thermocouples are wired to terminals specially constructed to minimize any temperature difference between their two ends and between each individual wire. These terminals are sometimes referred to as “isotherm” blocks.
Isotherm blocks allow the famous “Law of Intermediate Metals” to operate. This law states that if all connection points on the isotherm block are of equal temperature, adding a third metal in both wires of the circuit won’t degrade the signal.
With the thermocouple connected to one side of the isotherm block, the other side connects the special thermocouple extension or compensation wire. Technically, extension wires have the same chemical composition as their thermocouples, while compensation wires have a different composition. In practice, though, the extension cable term is most often used regardless of the wire composition.
Extension cables can cost up to 50% less than thermocouple wire. They also come in sizes up to 14 awg, which reduces loop resistance in long runs. Between 0 and 200°F, extension cables exhibit the same electrical properties as the thermocouples themselves and are electrically indistinguishable from thermocouple wire. To keep extension wire from inadvertently being used as thermocouple wire, the outer insulation jacket on the extension wires is color coded differently than the outer jacket on thermocouple wires. However, the individual wire insulation color matches that of thermocouple wire.
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 the letter “X.” So a “JP” designation refers to the positive leg of a Type J thermocouple. “KNX” refers to the negative leg of a Type K extension cable.
Extension wires typically run into distributed control systems, PLCs, or temperature analyzers back at the control cabinet. For the very best results, a user would run one continuous piece of wire from the hot junction to the terminals on the measuring instrument. But often this isn’t practical. Cabinets built off-site and by third-parties can’t be wired directly. Therefore, terminal blocks are used to help make wiring easier. The panel shop wires from the controller to the terminal block, leaving the other side open for the field connection of the extension wire.
There is an unfortunate widespread belief that thermocouples can simply connect 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, though high-quality blocks use nickel-plated-copper alloy. Either material forms a cold junction because it doesn’t match the chemical composition of the thermocouple wire. As mentioned in the last section, it’s possible to use terminal blocks without making a cold-junction correction. But that works only if the temperature on both sides of the block remains the same or changes at exactly the same rate. That isn’t likely in control cabinets packed with heat-producing equipment, fans, and other devices that heat and cool the interior unevenly. Standard terminal blocks are not designed to minimize temperature differentials like isotherm terminals. They make the measurement unpredictable whenever a heat-generating device is switched on or off or the cabinet is opened.
The solution is to use thermocouple blocks whose metal parts match the composition of the thermocouple extension wire. Because the materials match, there is no cold junction. To the thermocouple circuit, it looks like a continuous piece of extension wire. Changing temperature won’t introduce errors. Thermocouple blocks are typically more expensive than standard terminal blocks, but eliminate a potentially tedious source of errors.
Main system controllers, whether PLC, PC, or DCS, no longer need to perform temperature control. It may be preferable to “outsource” or distribute that control to another device.
For example, DIN-rail-mounted temperature transmitters convert low-level thermocouple signals to 4-to-20 mA and electrically isolate the circuit from ground loops. Some allow simple ON/OFF control with programmable dead bands using a transistor or relay output while others may also incorporate displays.
Other devices measure several thermocouples at once using sophisticated control algorithms to do more advanced control than simple ON/ OFF. Proportional, integral, and derivative (PID) control has been used for decades and is rapidly finding applications in the discrete world. In applications like heat seal packaging and extrusion processes, temperatures must be controlled rapidly and precisely. Heat must be added or removed in fractions of a second.
One of the latest devices for temperature monitoring uses a communication bus controlling and transmitting temperatures. Because they typically mount in a junction box on standard DIN-rail, they can reside close to the thermocouples, reducing possible sources of error. The input filtering, control logic, and process output all take place in the device. There is no need to stretch PLC resources trying to process the PID loops. And because the device is part of a bus network, the information can be sent out via standard industrial protocols such as Ethernet, Interbus, Profibus, or DeviceNet. Other I/O signals can be added for analog, digital, or serial devices.
A few problems show up regularly in thermocouple applications. Watch out for wires that are damaged or broken by rough installation, vibration, or other stress.
Take care not to reverse the polarity of the thermocouple loop. Remember that USA standard ASTM E 230 dictates the negative leg always has red insulation and its symbol is the thicker wire. The first named element of a thermocouple is always positive.
Uncompensated cold junctions cause many problems in areas where stable temperatures can’t be guaranteed, which is just about everywhere. Only use thermocouple or extension wire in the circuit. Be suspicious of any standard nonisotherm or nonthermocouple terminal blocks in the loop.
Make sure the loop resistance isn’t too high. The maximum value you should shoot for is usually 100 Ω. Remember that total length is twice the actual length because there is both a supply and return wire length. Use wire tables or test the cable to determine its resistance.
Keep thermocouple wires away from other current-carrying conductors, motors, and sources of radio frequency noise. If in doubt (that is, always), convert the thermocouple voltage to a more robust signal like a 4-to-20-mA instrumentation loop.
When using thermocouples from outside the U.S., verify the color coding is correct. Europe and Japan use different conventions than the standards that originated in the U.S. The varying conventions have been used for more than 60 years in their respective countries, so there is little chance of further “harmonization” in the future.
Many affordable products exist that can simplify and enhance thermocouple measurements. Transmitter, meter, and switch prices have become low enough to avoid many traditional difficulties using thermocouples. This also allows consideration of new temperature measurements that were either too expensive or too complicated in the past. Finally, conversion to a process control signal (4-to-20 mA) or a digital signal allows long-distance transmission without having to worry about the quirks of a thermocouple circuit.
Phoenix Contact USA, (800)888-7388, phoenixcon.com
Power supplies and thermocouple loops
Temperature loops that use thermocouple transmitters require steady and reliable 24-Vdc power for all components to work correctly. Most supplies are now switch-mode, giving users several advantages that include low purchase price, light weight, and long life. These modern supplies have efficiencies in the 90% range for more economical operation and lower expenditures for cabinet cooling equipment and operation.
From a purely technical standpoint, selecting a power supply means finding one that accepts a convenient input voltage and provides enough amperage to drive all the loads. Supplies can come in a wide-range of input voltages with output currents up to 40 A.
Supply location also plays an important role in the selection process. For example, supplies listed with UL1604 Class I, Div. 2 ratings limit the escape of electrical and heat energy for safe use in areas that may contain explosive vapors.
NEC Class 2 power supplies find use in or on machines because that class limits the amount of energy they can produce to 100 W or less, or about 4.2 A at 24 Vdc. Wires connected to that type of supply do not need conduit runs.
A recent innovation in control power is the 24-Vdc uninterruptible power supply, or UPS. The UPS system combines a power supply, battery, and a charging circuit in one box. In the event of power loss, the UPS maintains voltage long enough to ride through the outage or provide an orderly shutdown if the outage lasts too long.
Thermocouple transmitters typically convert weak millivolt signals from the thermocouple into robust 4-to-20-mA process control signals. Thermowell-mount circular (hockey puck) and DIN-rail cabinet mount transmitter styles perform identical functions.
Connectors and terminal blocks specifically designed for thermocouple installations have contacts made of the same material as the thermocouple to prevent the formation of cold junctions.