Home > Applications

THERMOCOUPLES

INTRODUCTION

The basic theory of thermocouples dates back to 1821 when T.J. Seebeck discovered that a current is induced into a closed circuit of two dissimilar metals by heating one of the two junctions. And, as long as the temperature differences exists between the two junctions, current will continue flowing through the circuit.

While the theory is nearly 150 years old, incorrect application of thermocouples still affects today's sophisticated industrial processes. In any temperature control system, the heart of that system is the temperature sensing device -- in this case, thermocouple. Without proper application or understanding of basic thermocouple circuits, even the most complicated system cannot function.

In his discovery, Seebeck also concluded that any two metals can be used. However, the magnitude and direction of the generated current are functions of the magnitude of the temperature difference between the junctions and the thermal properties of the metals used in the circuit. Therefore, not every combination of metals is acceptable for thermocouple usage.

THERMOELECTRIC CHARACTERISTICS

A thermocouple should have thermoelectric characteristics such that the electromotive force (emf) produced per degree of temperature change is sufficient to be detected by standard measuring instruments. The device must also be capable of withstanding temperature extremes for prolonged periods, rapid temperature changes, and corrosive atmospheres while exhibiting reproducibility and a high degree of accuracy.

The Instrument Society of America (ISA) has established a type of code and limits-of-error specifications for thermocouple wire, shown below:

Six thermocouples are covered by this system:

• Copper vs. Constantan (ISA Type T) (Nickel Alloy)
  Superior for use at sub zero temperatures. Withstands corrosion well and is recommended for temperatures within the range of -300 deg F to +700 deg F.

• Iron vs. Constantan (ISA Type J)
  Above 1000 deg F, the rate of oxidation of the iron wire increases rapidly and the thermocouple should be enclosed in a protection tube of suitable material. Protected Iron vs. Constantan thermocouples are recommended for temperatures up to 1600 deg F.

• Chromel vs. Alumel (ISA Type K) (Nickel Chromium Alloy) or (Nickel Manganese, Aluminum, Silicon)
  These are trade names of Hoskins Manufacturing Company. Chromel vs. Alumel thermocouples have excellent characteristics when supplied with protection tubes up to 2200 deg F.

• Chromel vs. Constantan (ISA Type E)
  Although only in limited use in industrial applications, this type thermocouple has the highest emf output of any standardized metallic type. They may be used in oxidizing, inert or reducing atmospheres to 1600 deg F and at sub-zero temperatures they are not subject to corrosion. Indications are that the future will see more consideration given to this combination.

• Platinum vs. Platinum-Rhodium
  Platinum vs. 90% Platinum + 10% Rhodium (ISA Type S) and Platinum vs. 87% Platinum + 13% Rhodium (ISA Type R) are used for temperatures up to approximately 3100 deg F, depending upon the atmosphere. Both types should always be provided with a high temperature ceramic protection tube.

Various other combinations of materials are used for thermocouples throughout industry but with far less frequency than the six basic types. Combinations of platinum and rhodium with various percentages of each have been regularly available for some years. Iridium vs. Iridium with 40, 50 or 60% Rhodium thermocouples have found acceptance in some high temperature applications up to 3600 deg F.

Proper thermocouple selection is primarily determined by the temperature range in which its use is intended. Other factors, such as atmosphere, abrasion, vibration, and location will determine the type, size, and configuration of the complete assembly which includes protection tubes and mounting facilities.

LOCATION

Proper location of the thermocouple is probably the most important factor in obtaining accurate temperature control. Thermocouples should be in a position to have a definite temperature relationship to the heat source and workload. A good 'rule of thumb' in locating thermocouples is to place them between the workload and heat source. The thermocouple should be located 1/3 the distance from the heat source and 2/3 the distance to the workload.

If a thermocouple is located too close to the heaters, a long warm-up time will result. The thermocouple will sense the heat before it reaches the workload, and this means rapid on/off action of the controller. In effect, the controller is controlling the heater and not the workload. In rare cases, voltage will be induced into the thermocouple circuit at high temperatures when located too near the heaters.

When a thermocouple is located too close to the workload, there is a substantial delay in sensing the proper control point and the result is overshooting the temperature. In most cases, it is better to be too close to the heaters than the workload as once a temperature point is passed, it becomes difficult to cool the workload unless a forced cooling system is used. Two thermocouples connected in parallel could be used, one located near the heaters and the other near the workload. Both will balance these two factors and provide closer control.

Another consideration in location is when locating a thermocouple in a thermocouple well. If it is not bottomed correctly, located at the bottom of the well, the thermocouple will be reading the air temperature around it and not the temperature of the workload.

COMPENSATION

The compensation method used by all millivoltmeter manufacturers is to attach a bi-metallic spiral to the top hairspring of the coil suspension system. This spiral is selected according to the range of the instrument and will deflect the indicating pointer correspondingly with changes in ambient temperature. Once ambient is set mechanically, using a zero adjust screw, it is not necessary to change the setting during the operation of the instrument. In solid state instruments, the compensation is achieved electronically by placing a temperature sensor, such as a thermistor or RTD, at the cold junction to monitor its temperature. The signal from this sensor is used to compensate for variations in cold junction temperature.

The automatic compensation for ambient temperature is sufficient in most industrial applications. However, in laboratory experiments or critical control situations, when maximum accuracy is desired, one of two cold junction compensation methods are used. One method is to place the cold junction in an agitated ice bath, shown below. The instrument will then be set at 32 deg F (0 deg C), which is the temperature of the ice bath.

In the other method, the cold junction is situated in a precisely controlled temperature above ambient, as shown below. In this case, ambient compensation is not necessary. The mechanical zero adjustment is set at the cold junction temperature being maintained. The normal temperature being maintained is 150 to 200 degrees F at the cold junction.

In normal applications, if the cold junction is located too close to the heat source, conduction and radiation heating will cause inaccurate readings. Errors will also occur when using copper wire or the wrong thermocouple lead wire. When copper wire is used, the cold junction in effect remains at the thermocouple connector block instead of the instrument. This will cause the instrument to read low in most cases unless the cold junction and instrument are known to have the same ambient temperature.

There is one application where cold junction compensation is not a factor. When two thermocouples are connected in series opposing, as shown below, a millivoltage is produced which is the difference in millivolts between the temperature at both thermocouples. As the difference in degrees between the two thermocouples is being measured, cold junction compensation is not necessary.

Each millivolt measuring instrument is calibrated for both the type of thermocouple being used and the length and gauge of the lead wire. The thermocouple lead wire is in effect in series with the thermocouple wire and the meter movement. Using wrong thermocouple lead wire can be avoided by simply following the color-coding used by all manufacturers (Table 2, below). A solid state controller can be used with up to 100 ohms of external resistance without having to be recalibrated.

When a millivoltmeter is calibrated, a series resistance (commonly called a calibrating spool) is used between the moving element coil of the instrument and the thermocouple tip. The resistance of the wire must be determined and used in the calibration of the instrument. If the resistance of the thermocouple wire and extension wire is higher than the instrument is calibrated for, the temperature readings will be low and if the resistance is lower, the temperature readings will be high.

Where the thermocouple and extension wire are a significant portion of the circuit, then we must also consider the resistance change of the thermocouple wire at elevated temperatures. It may be necessary to calibrate instruments at the operating temperature. As an example: 5 feet of .020 dia. platinum vs. platinum 10% Rhodium thermocouple wire, would have a resistance of 2.3 ohms. At 2500 deg F, the resistance would be 2.3 x 3.5 ohms, or 8.5 ohms. A millivoltmeter with a sensitivity of 10 ohms per volt would have an error of approximately 4% at 2500 deg F. The effects of temperature on the thermocouple and thermocouple extension wire are shown in Table 3.

To illustrate the effects of incorrect lead length calibration on the millivoltmeters, we have charted the errors that can result for various ranges and thermocouples by deviating in resistance from the calibrated lead length. Table 4 is based upon 10 ohms per millivolt sensitivity instrument. Instruments with less sensitivity would show greater errors. A meter with a 5 ohm per millivolt sensitivity would have errors twice as great.

Table 4: Deviation in Ohms from calibrated lead length

1 -- 0-2000 deg F C/A (4.02% at 20 Ohms) 2 -- 0-1200 deg F I/C, 0-1600 deg F C/A (4.97% at 20 Ohms) 3 -- 0-800 deg F I/C, 0-600 deg F C/A (7.65% at 20 Ohms) 4 -- 0-500 deg F CU/C, 0-2200 deg F PLT/PLT + 13% RH 0-2400 deg F PLT/PLT + 10% RH (14.32% at 20 Ohms) 5 -- 0-300 deg F I/C, o-350 deg F CU/C (22.15% at 20 Ohms)

THERMOCOUPLE CONNECTION

There are two common errors in connection thermocouple circuits. One is to connect the extension lead wire completely reversed. In this case, you would receive a low reading because the reversal causes the emf generated at the connection of the thermocouple and extension lead wire to be subtracted from the emf generated by the thermocouple. A more obvious error is to completely reverse the thermocouple. The instrument in this case will read downscale with an increase in temperature.

Some control instruments feature 'thermocouple break protection' which means that in the event of an open or broken thermocouple, a small voltage is applied to the instrument which will cause it to read full scale and turn off the external circuit. Thus, in the event the thermocouple breaks because of a mechanical shock or vibration or if it is over-exposed to extremely high temperature and deterioration sets in, an unattended process will not overheat because of the loss of control.

Another consideration in the thermocouple use is that the leads wires should never run in the same conduit with electrical lines. This may induce currents in the thermocouple wire, resulting in instrument errors and poor control. However, if this cannot be avoided, or if the induced currents are being picked up at the thermocouple itself, then one side of the thermocouple lead wire should be grounded through a 1.0 microfared paper capacitor at one of the thermocouple terminals in the instrument. In emergencies, a direct ground will sometimes work as well.

Occasionally, because of atmospheric conditions, corroding may occur on connections which cause a loss of the millivolt signal. Or, a poor connection between the lead wire and thermocouple could cause loss of signal.

The gauge size of the wire used in thermocouples is again dependent upon the application. Usually, when longer life is required, for the higher temperature ranges, the larger size wires are chosen. When sensitivity is the prime concern, the smaller sizes should be used.

GLOSSARY OF TERMS

CALIBRATE - General: to determine the indication or output of a measuring device with respect to that of a standard.

CALIBRATE - Thermocouple: to determine the emf developed by a thermocouple with respect to temperature established by a standard.

CALIBRATION POINT - General: a specific value, established by a standard, at which the indication or output of a measuring device is determined.

CALIBRATION POINT - Thermocouple: a temperature, established by a standard, at which the emf developed by a thermocouple is determined.

CELSIUS - The designation of the degree on the International Practical Temperature Scale. Also used for the name of the Scale, as "Celsius temperature scale." Formerly (prior to 1948) called "centigrade."

CENTIGRADE - The designation of the degree on the International Temperature Scale prior to 1948. (See Celsius)

COAXIAL THERMOCOUPLE ELEMENT - A thermocouple element consisting of a thermoelement in wire form, within a thermoelement in tube form with the two thermoelements insulated form each other and from the tube except at the measuring junction.

CONNECTION HEAD - A housing enclosing a terminal block for an electrical temperature-sensing device and usually provided with threaded openings for attachment to a protecting tube and for attachment of conduit.

ELECTROMOTIVE FORCE - (emf) The electrical potential difference which produces or tends to produce an electric current.

EXTENSION WIRE - A pair of wires having such temperature-emf characteristics relative to the thermocouple with which the wires are intended to be used that, when properly connected to the thermocouple, the reference junction is transferred to the other end of the wires.

FAHRENHEIT - The designation of the degree and the temperature scale used commonly in public life and engineering circles in English-speaking countries. Related to the International Practical Temperature Scale by means of the equation:

• Degrees Fahrenheit = (Degrees Celsius x 1.8) + 32
• Degrees Celsius = (Degrees Fahrenheit - 32) / 1.8

  FREEZING POINT - The fixed point between the solid and liquid phases of a material when approached from the liquid phase under a pressure of one standard atmosphere (101325 N/m squared). For a pure material, this is also the melting point.

  ICE POINT - The fixed point between ice and air-saturated water under a pressure of one standard atmosphere (101325 N/m squared). This temperature is 0 deg C on the International Practical Temperature Scale.

  KELVIN - The designation of the thermodynamic temperature scale and the degree on this scale. This kelvin scale was defined by the Tenth General Conference on Weights and Measures in 1954 by assigning the temperature of 273.16 degrees Kelvin to the triple point of water. Also the degree on the International Practical Kelvin Temperature Scale.

  MELTING POINT - The fixed point between the solid and liquid phases of a material when approached from the solid phase under a pressure of one standard atmosphere (101325 N/m squared). For a pure material, this is also the freezing point.

  PROTECTING TUBE - A tube designed to enclose a temperature-sensing device and protect it from the deleterious effects of the environment. It may provide for attachment to a connection head, but is not primarily designed for pressure-tight attachment to a vessel.

  RANGE - The region between the limits within which a quantity is measured. It is expressed by stating the lower and upper range-values.

  REFERENCE JUNCTION - That junction of a thermocouple which is at a known temperature.

  REFERENCE POINT - (liquid-in-glass thermometer) A temperature at which a thermometer is checked for changes in bulb volume.

  REFRACTORY METAL THERMOCOUPLE - A thermocouple whose thermoelements have melting points above that of 60 percent platinum, 40 percent rhodium, 1935 deg C (3515 deg F)

  RESISTANCE, INSULATION - (sheathed thermocouple wire) The measured resistance between wires or between wires and sheath multiplied by the length of the wire expressed in megohms (or ohms) per foot (or meter) of length. (NOTE: The resistance varies inversely with the length.)

  SEEBECK COEFFICIENT - the rate of change of thermal emf with temperature at a given temperature. Normally expressed as emf per unit of temperature. synonymous with thermoelectric power.

  SEEBECK EMF - The net emf set up in a thermocouple under condition of zero current. It represents the algebraic sum of the Peltier and Thompson emf. Synonymous with thermal emf.

  SHEATHED THERMOCOUPLE - A thermocouple having its thermoelements, and sometimes its measuring junction, embedded in ceramic insulation compacted within a metal protecting tube.

  SHEATHED THERMOCOUPLE MATERIAL - One or more pairs of thermoelements (without measuring junction(s)) embedded in ceramic insulation compacted within a metal protecting tube.

  THERMOCOUPLE - Two dissimilar thermoelements so joined as to produce a thermal emf when the junctions are at different temperatures.

  THERMOCOUPLE ASSEMBLY - An assembly consisting of a thermocouple element and one or more associated parts such as terminal block, connection head, and protecting tube.

  THERMOCOUPLE ELEMENT - A pair of bare or insulated thermoelements joined at one end to form a measuring junction and intended for use as a thermocouple or as part of a thermocouple assembly.

  THERMOCOUPLE (TYPE E, B, J, K, R, S, OR T) - A thermocouple having an emf-temperature relationship corresponding to the appropriate letter-designated table in ASTM Standard E 230, Temperature Electromotive Force (EMF) Tables for Thermocouples, within the limits of error specified in that Standard.

  THERMOPILE - A number of thermocouples connected in series, arranged so that alternate junctions are at the reference temperature and at the measured temperature, to increase the output for a given temperature difference between reference and measuring junctions.

  THERMOWELL - A closed end reentrant tube designed for the insertion of a temperature-sensing element, and provided with means for pressure-tight attachment to a vessel.

  WORKING STANDARD THERMOCOUPLE - A thermocouple that has had its temperature-emf relationship determined by reference to a secondary standard of temperature.