svenska english

Also read

Thermocouple theory


Entire circuit used for measurement
Overview of thermocouples
Thermocouple Tolerance Classes (IEC 60584-2)
Wolfram-Rhenium types (ASTM E988)

Properties and sources of error
Material defects and ageing
Green-rot corrosion
Warning against use of alternative materials
Other material in the measuring junction
SRO hysteresis
Other ageing ailments

Faulty connections
Open circuit
Short circuit
Reversed polarity of entire measuring circuit
Reversed polarity of the sensor
Double reversed polarity of extension cable

Different construction methods
Wire sensors
Armoured, spring-loaded wire sensors
Metal sheathed thermocouples
Three different probe tips
Sheathed thermocouple wires in outer protective tubes
Signal connections for metal sheathed thermocouples

Signal connection cables
Extension leads and compensating cables
Error limits for extension leads and compensating cable (IEC 60584-3)
Colour codes for thermocouple materials (pdf)


The thermocouple has been with us for 150 years and is by far the most widespread type of temperature sensor in the industry. It is both relatively inexpensive and simple to use, and its shortcomings are well documented.

Used correctly, the thermocouple is an excellent thermometer; used incorrectly, it can transform measurements into the realms of pure guesswork.


Seebeck and his discoveries

The thermocouple was invented by T.J. Seebeck in 1821. He discovered that if two dissimilar metal or alloy wires are joined at both ends to form a circuit, an electromotive force will be produced when there is a temperature difference between the junctions. The greater the difference, the higher the electromotive force.

The two junctions are called the measuring junction and the reference junction - the latter being the point to which a measuring instrument is connected.


New theories

We now know that Seebeck did not fully understand the physics behind the thermocouple. Voltmeters in his day applied a load to the circuit, creating currents that resulted in voltage drops and other phenomena.

Science has subsequently adopted theories on pure Seebeck voltages, which makes it easier to understand the flaws in a measuring circuit - a subject well- documented at Pentronic and also central to our training courses. Modern instruments incorporate input amplifiers having a very high input resistance, which means that the circuits can now be regarded as load-free.

If a metal wire is placed in a temperature gradient, a Seebeck voltage will occur over the entire length of the gradient. The steeper the gradient, the higher the voltage at that point. 

3_1_w 200_sv

At the top of the diagram is a single metal conductor and, below that, a complete thermocouple. Both are exposed to the same temperature gradient, T1 to T2. The respective output signals are E A and EAB.
The imaginary Seebeck voltage in the single metal wire will be: EA = SA (T1 - T2) where SA is  the supposedly constant Seebeck coeffi-cient for the material. It follows that
EB = SB (T2 - T1)

The following expression is obtained for the complete thermocouple: EAB = EA - EB =
(SA - SB) (T1 - T2)
which normally is written EAB = SAB (T1 - T2).

EAB should mathematically be expressed as an integrated sum of products along all of the thermocouple.


Entire circuit used for measurement

A thermocouple consists of two metallic wires having different Seebeck coefficients. These generate different voltages, resulting in a measurable voltage that varies with the temperature difference in the gradient.

All the components of the measuring circuit that are inside the temperature gradient, including the compensating lead and connectors, contribute a voltage.

If a non-calibrated thermoelectric material is placed in a temperature gradient, the measuring uncertainty has to be deduced from the standard manufacturing tolerances, viz.: ± 2.5°C for types K, N and J as per IEC 60584, Class 2.

Instrument influence

Half of the temperature-sensing function of a thermocouple takes place in the indicator instrument. This is the reference-junction temperature, which is measured at the transition between the thermoelectric material and the extension lead (copper conductor). In the past, this was known as compensation for the cold junction, which is normally optimized for room temperature (20–25°C).

This means that the measured value will be incorrect the moment that the reference temperature deviates from room temperature. An alternative way to obviate this error in the laboratory is to place the reference junction in a bath of ice water.

Overview of thermocouples

There are a dozen or so standard types of thermocouple, each of which complements the others in terms of the measuring signal, temperature range and tolerance to different environments. Types S, B and R contain platinum and are therefore more expensive than the others.

The following are the primary standardized thermocouples by the IEC 60584:2013:

Type Temp. range Environment
T -200 – 370 °C Good in low temperatures.
Protection tube above 240 °C.
J -200 – 760 °C Not oxidizing environment or acids.
E -200 – 900 °C Good in oxidizing environment.
K -200 – 1260 °C Good in oxidizing environment. Not suitable in reductive environment, such as sulfur, cyanide, carbon and hydrogen.
N -200 – 1300 °C As type K, but better above 200 °C.
S, R 0 – 1480 °C Ceramic protection tube necessary in all environments.
B 0 – 1700 °C Ceramic protection tube necessary in all environments.
425 - 2315 °C Ceramic protection tube necessary in all environments. Mainly used in vacuum.
100 - 2500 °C  Ceramic protection tube necessary in all environments. Mainly used in vacuum.

Different output signals

The output signal, which is determined by the relative Seebeck coefficient of each material, constitutes an important difference between thermocouples.

Manufacturing tolerances for thermocouples

The standards of the International Electrotechnical Commission (IEC) are internationally recognized and should be followed whenever possible. In time, all national standards throughout the world will be harmonized with the IEC requirements.

IEC 60584 (2013)

IEC 60584, last revised in 2013, contains reference tables and calculation polynomials for the output signals of standardized thermocouples as a function of temperature.

Tables of polynomials, together with the output signals (Seebeck voltages) and Seebeck coefficients, for each ten-degree division are given using the link of the Info pane above. More-detailed tables are available from Pentronic on request.

The table of tolerances below shows the manufacturing tolerances for the standardized thermocouples as per IEC 60584.

The tolerance should not be confused with measuring uncertainty. The properties of each individual sensor are determined by calibration.

Thermocouple Tolerance Classes as per IEC 60584:2013

(Reference point at 0°C)

Termocouple type

Tolerance class 1
Tolerance class 2 (°C) Tolerance class 3
Temperature range  -40 < T < 350  -40 < T < 350 -200 < T <40
Greatest of: ±0.5
±0.004 · lTl

±0.0075 · lTl

-200 < T < 40
±0.005 · lTl
Temperature range  -40 < T < 750   -40 < T < 750 -
Greatest of: ±1.5
±0.004 · lTl

±0.0075 · lTl 

Temperature range -40 < T < 800 -40 < T < 900 -200 < T <40
Greatest of: ±1.5
±0.004 · lTl

±0.0075 · lTl 

-200 < T < 40
±2.5 · lTl
K & N      
Temperature range -40 < T < 1000 -40 < T < 1200  -200 < T <40
Greatest of: ±1.5
±0.004 · lTl
±0.0075 · lTl  
±0.0075 · lTl   
S & R      
Temperature range 0 < T < 1600  0 < T < 1600   -
Greatest of:

±1 for T<1100
[±1+0.003 · (T-1100)]
for T>1100

±0.0025 · T   

Temperature range - 600 < T < 1700  600 < T < 1700 
Greatest of: - ±1.5
±0.0025 · T 
±0.005 · T 
Temperature range - 426 < T < 2315  -
Tolerance  - ±0.01 · T  -
Temperature range - 1000 < T < 2500 -
Tolerance  - ±0.01 · T  -

The thermocouples type C and type A are new in the 2013 issue of the IEC standard. They contain different proportions of Tungsten and Rhenium as follows:

Type C: W-5%Re / W-26%Re
Type A: W-5%Re / W-20%Re

The type D: W-3%Re / W-25%Re also exists but has not been adopted by the IEC.

Note that the given tolerance figures apply to unused thermocouples. Even after very little use, changes must be assumed. Changes will be accelerated by higher temperatures and exposure to harsh environments.

Calibration of individual thermocouples or production batches at lower temperatures and in neutral environments can reduce the measuring error.



Properties and sources of error

Material defects and ageing

All types of thermocouple are subjected to varying degrees of wear and ageing depending on the environments in which they are used. It is therefore essential that all types of sensor should be regularly inspected and calibrated. The type-K sensor is the most widely used and therefore the best documented.

The following are the chief hazards:

Green-rot corrosion

Green rot is a type of corrosion that occurs on the positive wire of a type-K thermocouple (and type E) if two conditions are met:

The problem sets in slowly as the wire degrades. If it is not discovered in time, measuring errors of tens of degrees can occur. In extreme cases, Pentronic has found measuring errors of 50°C or more.

Sheathed thermocouples are available which incorporate sacrificial titanium wires to delay the reaction.

An alternative solution is to change to the use of a type N or S thermo-couple, although the type S requires a ceramic sheath.

Warning against use of alternative materials

Some thermocouples are expensive and it can therefore be tempting to use other materials as extension leads - don’t!

As explained earlier, the Seebeck effect is created throughout the measuring circuit, so the use of other materials having a different Seebeck coefficient can give rise to a faulty output signal.

There are materials, known as compensating leads, that have the same electrical properties as the thermocouple within limited temperature ranges. However, it is better to use thermoelectric material throughout the circuit.

If this is not possible, make sure that you select the right compensating leads and terminals for the thermocouple you are using. Compensating connectors have pins that prevent reversed polarity.

Other materials in the measuring junction   Other materials are sometimes used for the measuring junction. A typical case is when the thermocouple wires are soldered to a tab, which is secured by screws at the required measuring point.

This works if the tab has good electrical conductance and the temperature is the same at both ends of the wire. The wires must be as close together as possible so that they assume the same temperature—which constitutes the measured value.

SRO hysteresis

SRO (Short-range order) is a hysteretic phenomenon that affects type-K thermocouples at measuring temperatures above 200 °C. During manufacture, the thermocouple is annealed, which causes local disorder in the metal lattice. On cooling, the alloy atoms remain in the positions they happen to be in at the time.

Nature strives for a low level of energy and, at temperatures above about 200 °C, the atoms migrate towards more-efficient forms. This process produces an increase in the Seebeck effect and, thus, associated measuring errors. Errors can be of the magnitude of 5 °C.

Other thermocouples are not affected to the same extent by SRO. The error induced in a type-N thermocouple, for instance, is only a degree or so.

The process is reversible: on annealing, the thermocouple will resume its original properties. If the sensor is not going to be used at temperatures above 600 °C, the phenomenon can be utilized for ageing of the sensor before use.

Other ageing ailments

The ravages of time make their mark on all types of thermocouple. One common cause of a change in the emf is that material migrates from one wire to the other. A clear example of this can be seen in a type-S thermo-couple, which has wires of platinum/rhodium and pure platinum. At high temperatures, the rhodium vaporizes and drifts across to the pure platinum wires, which results in a gradual fall in the output signal at a given temperature.

Similar effects occur between the thermocouple wires and the sheath, insulation and tubes. The process is accelerated by high temperatures, vacuum and various atmospheres.

Mechanical shocks can also alter the properties of the thermocouple. Regular inspection and calibration is the only effective way to safeguard against measuring errors.


Faulty connections

Note that the colour code of the sketches refers to ANSI MC 96.1, which is American standard. Hence, for type K yellow colour marks the positive and red marks the negative wire. See pdf with different colour codes of the world.

Circuit break (open circuit)

A sensor wire has fractured, come adrift or is making poor contact with the instrument. Modern instruments trigger an alarm, eg, by “Open” appearing on the display.


3_2_w 200
Short circuit

If the insulation has chafed and a short circuit occurs, a measuring junction is created. The instrument will then display the temperature at the short-circuit point, instead of at the tip of the probe.


3_3_w 200
Reversed polarity of entire measuring circuit

If the polarity has been reversed, the instrument will also operate in reverse, ie, a temperature increase will be recorded as a temperature decrease.


3_4_w 200
Reversed polarity within the measuring circuit

The extension lead must have the same polarity as the thermocouple wires. If the polarity of the thermocouple element is reversed, opposing voltages occur. The reading obtained will then be twice the temperature in the terminal head minus the temperature at the measuring junction.


3_5_w 200
Double reversed polarity

If the polarity of the extension lead has been reversed at both ends, the temperature at the ends will affect the output signal. The reading will be the temperature at the measuring junction less twice the temperature difference between the terminal head and the reference junction.

Bear in mind that if a temperature controller having a set-point value of 1000 °C is connected, the power will be stepped up, thus giving a true value approximately 150 °C higher than the set-point value. Yet the instrument will still display a reading of 1000°C.


3_6_w 200

Different construction methods

Wire sensors

In its simplest form, a thermocouple element consists of a pair of wires that are connected together at a measuring junction. The junction must be electrically conducting and can be formed by soldering, crimping or twisting the wires together, depending on the strength requirement.

Wire sensors are mainly used to measure low temperatures in favourable environments. The limitations are imposed by the insulation material, eg, PVC, nylon, Kapton, PTFE or some form of glass-fibre material. Another limitation is that the measuring junction is exposed to the atmosphere.

PVC insulation can withstand temperatures of up to about 100 °C, whereas certain ceramic, fibrous materials are tolerant of temperatures up to 1000°C or more.


3_7_w 200
Armoured, spring-loaded wire sensors

Wire sensors used in industrial applications need to be protected against pinching and splashing with hot fluids. Protection can be achieved using a stainless-steel flexible armoured sheathing.

The armour has external spiralling and includes a spring-loaded bayonet fitting for the sensor, which can be screwed along the sheath for free adjustment of the insertion length.

The probe tip on the measuring junction is usually grounded but can be insulated using PTFE or sheathing material. This will make it tolerant to higher temperatures.


6203000-iso -w 200
Sheathed thermocouples

In sheathed thermocouples, the wires are insulated by densely packed magnesium oxide enclosed in a metal alloy suitable for the given thermocouple element.

The sheathing is available in different diameters and can be supplied on reels from the drawing mill. This means, for instance, that we can produce a continuous 100-m length of 3-mm-diameter sheathed thermocouple.

  • Sheathed thermocouples have several advantages:
  • The metal sheathing is hermetically sealed
  • Thermocouple tolerant to higher temperatures than wire sensor of corresponding size
  • Sheathing can be bent to required shape
  • Vibration resistant
3_9_w 200_sv

Three different probe tips

Sheathed thermocouples are available with three measuring-junction options. Pentronic usually supplies sensors with insulated measuring junctions but the other variants are available to special order.

Insulated measuring junction

This is to be preferred in the majority of applications. For instance, the sensor can readily be used for differential measurements, it avoids the problems associated with ground currents and has greater mechanical strength.

3_10_w 132
Grounded measuring junction.

The wires are welded to the bottom. This gives a somewhat faster response. Because the sheath, insulation and thermocouple wires have different coefficients of thermal expansion, rapid and high temperature fluctuations can cause the wires to come adrift from the probe.

3_11_w 132
Exposed measuring junction.

The measuring junction protrudes from the sheath and the tube is sealed, eg, by glass. The advantage is the shortest possible response time. However, several advantages of the fully enclosed sheathed thermo-couple are lost - tolerance to high temperatures being one. We would only recommend this type of probe when fast response is the most important criterion.

3_12_w 132

Method of manufacture critical

During manufacture, the wires are exposed at both ends. At the sensing end, the wires are welded together to form the measuring junction and the sheath is filled with magnesium oxide, which is packed tightly to increase the thermal conductance. The sheath is then sealed. The other end of the wires terminates in a connector, an extension lead or a terminal block.

Cleanliness is paramount during manufacture of the measuring junction. Any traces of oil, dirt or other contaminants can destroy the insulation and thermocouple.

The material must also be kept sealed to prevent the ingress of air, as the moisture in the air can break down the insulation and give rise to measuring errors. It is not possible to remove enough of the moisture by heating.

Sheathed thermocouple wires in outer protective tubes

Sheathed thermocouple wires supersede an earlier type of thermocouple detector insert encapsuled in a protective tube. In the past, thick, uninsulated thermocouple wires were used, separated from each other and the protective tube by ceramic insulators. The construction was fragile, and it was difficult to position the measuring junction at the probe-end of the tube.

Inserts consisting of sheathing material are easier to work with and have a longer service life - despite the thinner wires used. In addition, the terminal block can be spring-loaded, to ensure that the thermocouple element makes contact with the bottom of the tube.

Sheathed thermocouples are often used as the detector for measuring temperatures from 500–600°C upwards, ie, above the normal temperature range of the Pt100.

3_17_w 200

Terminal block mounted on sheath for insertion in protective tube with terminal head.

Signal connections of metal sheathed thermocouples

There are a variety of connection types for sheathed thermocouples. The most common of these are shown below.

The sheathed thermocouple is the modern replacement for thick thermocouple wires isolated by ceramic insulators (former DIN 43763, type A). 3_13_w 200
Compensating connector moulded onto or fitted straight to the sheath. TC-8102000-iso -w 200
Non-terminated wires. TC-8101000-w 200
Sheathed thermocouple with transition fitting. TC-8105000-w 200


Signal connections

Extension leads (KX) and compensating  cables (KCA)

Thermocouple types are designated with a letter code as per IEC 60584-2.

Wire type Properties
K Thermocouple material of type K with a composition
that meets the tolerance requirements of IEC 584-2.
KX Extension lead of the same material as K but with limitedtemperature range. Probably downgraded K that did not fulfill the whole temperature range. See table below.
KCA Compensating cable, ie, wires made of a completely different material but having the same thermoelectric
properties as the thermocouple within a limited
temperature range - usually < 200°C. The letter A in the code designates a specific material. See table below.


These days, it is really only the type S (or type R 13%) platinum 10% rhodium versus platinum that requires a compensating lead of a different, much less-expensive material.

Manufacturing methods have improved over the years and IEC 584-2 describes two classes, 1 and 2, above 0 °C. To distinguish class 1 having tighter tolerances than class 2, "standard", the manufacturer usually uses a double-letter thermocouple code, as in the following example:

K is class 2 and KK is used to designate class 1.

IEC 60584-3 (1989). Error limits for extension leads and compensating cables.

  Estimation of error 
Designation Class of tolerance
Environment temperature Temperature of measurement
Extension wire (X)
min/max (°C)  (°C) 1
TX 30 60 -25/100 300 0.5 1.0
JX 85 140 -25/200 500 1.5 2.5
EX 120 200 -25/200 500 1.5 2.5
KX 60 100 -25/200 900 1.5 2.5
NX 60 100 -25/200 900 1.5 2.5
Compensating cable (C)
KCA - 100 0/150 900 - 2.5
KCB - 100 0/100 900 - 2.5
NC - 100 0/150 900 - 2.5
SCA - 30 0/100 1000 - 2.5
SCB - 60 0/200 1000 - 2.5
RCA - 30 0/100 1000 - 2.5
RCB - 60 0/200 1000 - 5.0


Colour coding

Cables and connectors are colour-coded. Unfortunately, several disparate national standard codes are in use, although IEC 60584-3 includes a uniform colour-coding standard for cables, and most of the manufacturers now stock cables conforming to the new standard colourcoding.

The IEC has not yet completed the colour-coding standard for connectors, but the Swedish standard gives the connectors the same colour code as that for the cable sheath (see below).

IEC 60584-3 (1989)

IEC 60584-3 (latest revision, 1989) deals with error limits for cables and the associated colour markings. The relevant standard in Sweden at present is SS 4853018 (1994) which, in addition to the provisions of IEC 60584-3 (1989) includes colour coding for type-B and type-N thermo-couples together with connectors.

The additional coding contained in the Swedish standard is expected to be included in the next revision of IEC 60584-3.


W-Re lead wires error limits per ASTM E988 (1990)

Lead wires for thermocouples containing tungsten and rhenium have not been standardized by the IEC. However, there is an American standard, ASTM E988-90, which is shown below:

Extension wires Temperature
range (°C)
Error limits
Reference at 0°C
Error  at low/high limit (°C)
For W 3%Re / W 25%Re:    
300P/300N 0 – 330 ±125µV 7/13
203P/225N 0 – 260 ±110µV 6.5/11.5
For W 5%Re / W 26%Re:    
405P/426N 0 – 871 ±110µV 6/9