Thermocouples

Article Index
Thermocouples
Thermocouple Types
Page 1 of 2
RTDs are completely passive sensing elements, requiring the application of an
externally-sourced electric current in order to function as temperature sensors.
Thermocouples, however, generate their own electric potential. In some ways,
this makes thermocouple systems simpler because the device receiving the
thermocouple’s signal does not have to supply electric power to the
thermocouple. The self-powering nature of thermocouples also means they do not
suffer from the same “self-heating” effect as RTDs. In other ways, thermocouple
circuits are more complex than RTD circuits because the generation of voltage
actually occurs in two different locations within the circuit, not simply at the
sensing point. This means the receiving circuit must “compensate” for
temperature in another location in order to accurately measure temperature in
the desired location.

Though typically not as accurate as RTDs, thermocouples are more rugged, have
greater temperature measurement spans, and are easier to manufacture in
different physical forms.

Dissimilar metal junctions

When two dissimilar metal wires are joined together at one end, a voltage is
produced at the other end that is approximately proportional to temperature.
That is to say, the junction of two different metals behaves like a temperature-
sensitive battery. This form of electrical temperature sensor is called a
thermocouple:

This phenomenon provides us with a simple way to electrically infer temperature:

simply measure the voltage produced by the junction, and you can tell the
temperature of that junction. And it would be that simple, if it were not for an
unavoidable consequence of electric circuits: when we connect any kind of
electrical instrument the iron and copper wires, we inevitably produce another
junction of dissimilar metals. The following schematic shows this fact:
Junction J1 is a junction of iron and copper – two dissimilar metals – which will
generate a voltage related to temperature. Note that junction J2, which is
necessary for the simple fact that we must somehow connect our copper-wired
voltmeter to the iron wire, is also a dissimilar-metal junction which will
generate a voltage related to temperature. Note also how the polarity of
junction J2 stands opposed to the polarity of junction J1 (iron = positive ;
copper = negative). A third junction (J3) also exists between wires, but it is
of no consequence because it is a junction of two identical metals which does
not generate a temperature-dependent voltage at all.

The presence of this second voltage-generating junction (J2) helps explain why
the voltmeter registers 0 volts when the entire system is at room temperature:
any voltage generated by the ironcopper junctions will be equal in magnitude and
opposite in polarity, resulting in a net (series-total) voltage of zero. It is
only when the two junctions J1 and J2 are at different temperatures that the
voltmeter registers any voltage at all.

We may express this relationship mathematically as follows:

With the measurement (J1) and reference (J2) junction voltages opposed to each
other, the voltmeter only “sees” the difference between these two voltages.

Thus, thermocouple systems are fundamentally differential temperature sensors.

That is, they provide an electrical output proportional to the difference in
temperature between two different points. For this reason, the wire junction we
use to measure the temperature of interest is called the measurement junction
while the other junction (which we cannot eliminate from the circuit) is called
the reference junction (or the cold junction, because it is typically at a
cooler temperature than the process measurement junction).

Thermocouples - Page 2
Article Index
Thermocouples
Thermocouple Types
Page 2 of 2

Thermocouple types

Thermocouples exist in many different types, each with its own color codes for
the dissimilar-metal wires. Here is a table showing the more common thermocouple
types and their standardized colors1, along with some distinguishing
characteristics of the metal types to aid in polarity identification when the
wire colors are not clearly visible:

Positive wire Negative wire

Type Plug Temp. range
characteristic characteristic
Copper (blue) Constantan (red)
T Blue -300 to 700 oF
yellow colored silver colored
 Iron (white) Constantan (red)
J Black 32 to 1400 oF
magnetic, rusty? non-magnetic
Constantan (red)
E Chromel (violet) shiny finish Violet 32 to 1600 oF
dull finish
Chromel (yellow) Alumel (red)
K Yellow 32 to 2300 oF
non-magnetic magnetic
N Nicrosil (orange) Nisil (red) Orange 32 to 2300 oF
S Pt90% - Rh10% (black) Platinum (red) Green 32 to 2700 oF

B Pt70% - Rh30% (grey) Pt94% - Rh6% (red) Grey 32 to 3380 oF

Note how the negative (−) wire of every thermocouple type is color-coded red.
While this may seem backward to those familiar with modern electronics (where
red and black usually represent the positive and negative poles of a DC power
supply, respectively), bear in mind that thermocouple color codes actually pre-
date electronic power supply wire coloring!

Aside from having different usable temperature ranges, these thermocouple types
also differ in terms of the atmospheres they may withstand at elevated
temperatures. Type J thermocouples, for instance, by virtue of the fact that one
of the wire types is iron, will rapidly corrode in any oxidizing2 atmosphere.
Type K thermocouples are attacked by reducing3 atmospheres as well as sulfur and
cyanide. Type T thermocouples are limited in upper temperature by the oxidation
of copper (a very reactive metal when hot), but stand up to both oxidizing and
reducing atmospheres quite well at lower temperatures, even when wet.

Connector and tip styles

In its simplest form, a thermocouple is nothing more than a pair of dissimilar-

metal wires joined together. However, in industrial practice, we often need to
package thermocouples in a way that optimizes their ruggedness and reliability.
For instance, most industrial thermocouples are manufactured in such a way that
the dissimilar-metal wires are protected from physical damage by a stainless
steel or ceramic sheath, and they are often equipped with molded-plastic plugs
for quick connection to and disconnection from a thermocouple-based instrument.

A photograph of a type K industrial thermocouple (approximately 20 inches in

length) reveals this “sheathed” and “connectorized” construction:
The stainless steel sheath of this particular thermocouple shows signs of
discoloration from previous service in a hot process. Note the different
diameters of the plug terminals. This “polarized” design makes it difficult4 to
insert backward into a matching socket.

A miniature version of this same plug (designed to attach to thermocouple wire

by screw terminals, rather than be molded onto the end of a sheathed assembly)
is shown here, situated next to a ballpoint pen for size comparison:

Industrial-grade thermocouples are available with this miniature style of molded

plug end as an alternative to the larger (standard) plug. Miniature plug-ends
are often the preferred choice for laboratory applications, while standard-sized
plugs are often the preferred choice for field applications.

Some industrial thermocouples have no molded plug at all, but terminate simply
in a pair of open wire ends. The following photograph shows a type J
thermocouple of this construction:

If the electronic measuring instrument (e.g. temperature transmitter) is located

near enough for the thermocouple’s wires to reach the connection terminals, no
plug or socket is needed at all in the circuit. If, however, the distance
between the thermocouple and measuring instrument is too far to span with the
thermocouple’s own wires, a common termination technique is to attach a special
terminal block and connection “head” to the top of the thermocouple allowing a
pair of thermocouple extension wires to join and carry the millivoltage signal
to the measuring instrument.

This next photograph shows a close-up view of such a thermocouple “head”:

As you can see from this photograph, the screws directly press against the solid
metal wires to make a firm connection between each wire and the brass terminal
block. Since the “head” attaches directly to one end of the thermocouple, the
thermocouple’s wires will be trimmed just long enough to engage with the
terminal screws inside the head. A threaded cover provides easy access to these
connection points for installation and maintenance, while ensuring the
connections are covered and protected from ambient weather conditions the rest
of the time.

At the other end of the thermocouple, we have a choice of tip styles. For
maximum sensitivity and fastest response, the dissimilar-metal junction may be
unsheathed (bare). This design, however, makes the thermocouple more fragile.
Sheathed tips are typical for industrial applications, available in either
grounded or ungrounded forms:

Grounded-tip thermocouples exhibit faster response times and greater sensitivity

than ungrounded-tip thermocouples, but they are vulnerable to ground loops:
circuitous paths for electric current between the conductive sheath of the
thermocouple and some other point in the thermocouple circuit. In order to avoid
this potentially troublesome effect, most industrial thermocouples are of the
ungrounded design.
Manually interpreting thermocouple voltages

Recall that the amount of voltage indicated by a voltmeter connected to a

thermocouple is the difference between the voltage produced by the measurement
junction (the point where the two dissimilar metals join at the location we
desire to sense temperature at) and the voltage produced by the reference
junction (the point where the thermocouple wires join to the voltmeter wires):

This makes thermocouples inherently differential sensing devices: they generate

a measurable voltage in proportion to the difference in temperature between two
locations. This inescapable fact of thermocouple circuits complicates the task
of interpreting any voltage measurement obtained from a thermocouple.

In order to translate a voltage measurement produced by a voltmeter connected to

a thermocouple, we must add the voltage produced by the measurement junction
(VJ2) to the voltage indicated by the voltmeter to find the voltage being
produced by the measurement junction (VJ1). In other words, we manipulate the
previous equation into the following form:

We may ascertain the reference junction voltage by placing a thermometer near

that junction (where the thermocouple wire attaches to the voltmeter test leads)
and referencing a table of thermocouple voltages for that thermocouple type.
Then, we may take the voltage sum for VJ1 and re-reference that same table,
finding the temperature value corresponding to the calculated measurement
junction voltage.

To illustrate, suppose we connected a voltmeter to a type K thermocouple and

measured 14.30 millivolts. A thermometer situated near the thermocouple wire /
voltmeter junction point shows an ambient temperature of 73 degrees Fahrenheit.
Referencing a table of voltages for type K thermocouples (in this case, the NIST
“ITS-90” reference standard), we see that a type K junction at 73 degrees
Fahrenheit corresponds to 0.910 millivolts. Adding this figure to our meter
measurement of 14.30 millivolts, we arrive at a sum of 15.21 millivolts for the
measurement (“hot”) junction. Going back to the same table of values, we see
15.21 millivolts falls between 701 and 702 degrees Fahrenheit. Linearly
interpolating between the table values (15.203 mV at 701 oF and 15.226 mV at 702
oF), we may more precisely determine the measurement junction to be at 701.3
degrees Fahrenheit.

The process of manually taking voltage measurements, referencing a table of

millivoltage values, performing addition, then re-referencing the same table is
rather tedious. Compensation for the reference junction’s inevitable presence in
the thermocouple circuit is something we must do, but it is not something that
must always be done by a human being. The next subsection discusses ways to
automatically compensate for the effect of the reference junction, which is the
only practical alternative for continuous thermocouple-based temperature
instruments.

Reference junction compensation

Multiple techniques exist to deal with the influence of the reference junction’s
temperature5. One technique is to physically fix the temperature of that
junction at some constant value so it is always stable. This way, any changes in
measured voltage must be due to changes in temperature at the measurement
junction, since the reference junction has been rendered incapable of changing
temperature. This may be accomplished by immersing the reference junction in a
bath of ice and water:

In fact, this is how thermocouple temperature/voltage tables are referenced:

describing the amount of voltage produced for given temperatures at the
measurement junction with the reference junction held at the freezing point of
water (0 oC = 32 oF).

However, this is not a very practical solution for dealing with the reference
junction’s voltage. Instead, we could apply an additional electrical circuit to
counter-act the voltage produced by the reference junction. This is called a
reference junction compensation or cold junction compensation circuit:
Please note that “cold junction” is just a synonymous label for “reference
junction.” In fact the “cold” reference junction may very well be at a warmer
temperature than the so-called “hot” measurement junction! Nothing prevents
anyone from using a thermocouple to measure temperatures below freezing.

This compensating voltage source (Vrjc in the above schematic) uses some other
temperature-sensing

device such as a thermistor or RTD to sense the local temperature at the

terminal block where junction J2 is formed, and produce a counter-voltage that
is precisely equal and opposite to J2’s voltage (Vrjc = VJ2). Having canceled the
effect of the reference junction, the voltmeter now only registers the voltage
produced by the measurement junction J1:

Some instrument manufacturers sell electronic ice point modules designed to

provide reference junction compensation for un-compensated instruments such as
standard voltmeters. The “ice point” circuit performs the function shown by Vrjc
in the previous diagram: it inserts a counter-acting voltage to cancel the
voltage generated by the reference junction, so that the voltmeter only “sees”
the measurement junction’s voltage. This compensating voltage is maintained at
the proper value according to the terminal temperature where the thermocouple
wires connect to the ice point module, sensed by a thermistor or RTD.

At first it may seem pointless to go through the trouble of building a reference

junction compensation (ice point) circuit, when doing so requires the use of
some other temperature-sensing element such as a thermistor or RTD. After all,
why bother to do this just to be able to use a thermocouple to accurately
measure temperature, when we could simply use this “other” device to directly
measure the process temperature? In other words, isn’t the usefulness of a
thermocouple invalidated if we have to go through the trouble of integrating
another type of electrical temperature sensor in the circuit just to compensate
for an idiosyncrasy of thermocouples?

The answer to this very good question is that thermocouples enjoy certain
advantages over these other sensor types. Thermocouples are extremely rugged and
have far greater temperature measurement ranges than thermistors, RTDs, and
other primary sensing elements. However, if the application does not demand
extreme ruggedness or large measurement ranges, a thermistor or RTD would likely
be the better choice.

Law of Intermediate Metals

It is critical to realize that the phenomenon of a “reference junction” is an

inevitable effect of having to close the electric circuit loop in a circuit made
of dissimilar metals. This is true regardless of the number of metals involved.
In the last example, only two metals were involved: iron and copper. This formed
one iron-copper junction (J1) at the measurement end and one iron-copper
junction (J2) at the indicator end. Recall that the copper-copper junction J3 was
of no consequence because its identical metallic composition generates no
thermal voltage:

The same thing happens when we form a thermocouple out of two metals, neither
one being copper. Take for instance this example of a type J thermocouple:

Here we have three voltage-generating junctions: J1 of iron and constantan, J2 of

iron and copper, and J3 of copper and constantan which just happens to be the
metallic combination for a type T thermocouple. Upon first inspection it would
seem we have a much more complex situation than we did with just two metals
(iron and copper), but the situation is actually just as simple as it was
before.

A principle of thermo-electric circuits called the Law of Intermediate Metals

helps us see this clearly. According to this law, intermediate metals in a
series of junctions are of no consequence to the overall (net) voltage so long
as those intermediate junctions are all at the same temperature. Representing
this pictorially, the net effect of having four different metals (A, B, C, and
D) joined together in series is the same as just having the first and last metal
in that series (A and D) joined with one junction, if all intermediate junctions
are at the same temperature:
In our Type J thermocouple circuit where iron and constantan both join to
copper, we see copper as an intermediate metal so long as junctions J2 and J3 are
at the same temperature. Since those two junctions are located next to each
other on the indicating instrument, identical temperature is a reasonable
assumption, and we may treat junctions J2 and J3 as a single iron-constantan
reference junction. In other words, the Law of Intermediate Metals tells us we
can treat these two circuits identically:

The practical importance of this Law is that we can always treat the reference
junction(s) as a single junction made from the same two metal types as the
measurement junction, so long as all dissimilar metal junctions at the reference
location are at the same temperature.

This fact is extremely important in the age of semiconductor circuitry, where

the connection of a thermocouple to an electronic amplifier involves many
different junctions, from the thermocouple wires to the amplifier’s silicon.
Here we see a multitude of reference junctions, inevitably formed by the
necessary connections from thermocouple wire to the silicon substrate inside the
amplifier chip:

It should be obvious that each complementary junction pair cancels if each pair
is at the same temperature (e.g. gold-silicon junction J12 cancels with silicon-
gold junction J13 because they generate the exact same amount of voltage with
opposing polarities). The Law of Intermediate Metals goes one step further by
telling us junctions J2 through J13 taken together in series are of the same
effect as a single reference junction of iron and constantan. Automatic
reference junction compensation is as simple as counter-acting the voltage
produced by this equivalent iron-constantan junction at whatever temperature
junctions J2 through J13 happen to be at.
Software compensation

Previously, it was suggested this automatic compensation could be accomplished

by intentionally inserting a temperature-dependent voltage source in series with
the circuit, oriented in such a way as to oppose the reference junction’s
voltage:

If the series voltage source Vrjc is exactly equal in magnitude to the reference
junction’s voltage (VJ2), those two terms cancel out of the equation and lead to
the voltmeter measuring only the voltage of the measurement junction J1:

This technique is known as hardware compensation. A stand-alone circuit designed

to do this is sometimes called an ice point, because it electrically
accomplishes the same thing as physically placing the reference junction(s) in a
bath of ice-water.

A more modern technique for reference junction compensation is called software

compensation. This is applicable only where the indicating device is
microprocessor-based, and where an additional analog input channel exists:
Instead of canceling the effect of the reference junction electrically, we can
cancel the effect mathematically inside the microprocessor. In other words, we
let the meter see the difference in voltage between the measurement and
reference junctions (Vmeter = VJ1 − VJ2). After digitizing this voltage
measurement, the microprocessor adds the equivalent voltage value corresponding
to the ambient temperature sensed by the RTD or thermistor (Vrjc):

Since we know the calculated value of Vrjc should be equal to the real reference
junction voltage (VJ2), the result of this digital addition should be a
compensated total equal only to the measurement junction voltage VJ1:

Perhaps the greatest advantage of software compensation is flexibility. Being

able to re-program the compensation function means this instrument may easily
interpret the voltage output of different thermocouple types with no
modifications to the hardware. So long as the microprocessor memory is
programmed with look-up tables relating voltage values to temperature values, it
may accurately measure (and compensate for the reference junction of) any
thermocouple type. With hardware-based compensation (an “ice point” circuit),
re-wiring or replacement is necessary to accommodate different thermocouple
types.

Extension wire

In every thermocouple circuit there must be both a measurement junction and a

reference junction: this is an inevitable consequence of forming a complete
circuit (loop) using dissimilar-metal wires. As we already know, the voltage
received by the measuring instrument from a thermocouple will be the difference
between the voltages produced by the measurement and reference junctions. Since
the purpose of most temperature instruments is to accurately measure temperature
at a specific location, the effects of the reference junction’s voltage must be
“compensated” for by some means, either a special circuit designed to add an
additional canceling voltage or by a software algorithm to digitally cancel the
reference junction’s effect.

In order for reference junction compensation to be effective, the compensation

mechanism must “know” the temperature of the reference junction. This fact is so
obvious, it hardly requires statement. However, what is not so obvious is how
easily this compensation may be unintentionally defeated simply by installing a
different type of wire in a thermocouple circuit.

To illustrate, let us examine a simple type K thermocouple installation, where

the thermocouple connects directly to a panel-mounted temperature indicator:

Like all modern thermocouple instruments, the panel-mounted indicator contains

its own reference junction compensation, so that it is able to compensate for
the temperature of the reference junction formed at its connection terminals,
where the internal (copper) wires of the indicator join to the chromel and
alumel wires of the thermocouple. The indicator senses this junction temperature
using a small thermistor thermally bonded to the connection terminals.

Now let us consider the same thermocouple installation with a length of copper
cable (two wires) joining the field-mounted thermocouple to the panel-mounted
indicator:
Even though nothing has changed in the thermocouple circuit except for the type
of wires joining the thermocouple to the indicator, the reference junction has
completely shifted position. What used to be a reference junction (at the
indicator’s terminals) is no longer, because now we have copper wires joining to
copper wires. Where there is no dissimilarity of metals, there can be no
thermoelectric potential. At the thermocouple’s connection “head,” however we now
have a joining of chromel and alumel wires to copper wires, thus forming a
reference junction in a novel location. What is worse, this new location is
likely to be at a different temperature than the panel-mounted indicator, which
means the indicator’s reference junction compensation will be compensating for
the wrong temperature.

The only practical way to avoid this problem is to keep the reference junction
where it belongs: at the terminals of the panel-mounted instrument where the
ambient temperature is measured and the reference junction’s effects accurately
compensated. If we must install “extension” wire to join a thermocouple to a
remotely-located instrument, that wire must be of a type that does not form
another dissimilar-metal junction at the thermocouple head, but will form one at
the receiving instrument.

An obvious approach is to simply use thermocouple wire of the same type as the
installed thermocouple to join the thermocouple to the indicator. For our
hypothetical type K thermocouple, this means a type K cable installed between
the thermocouple head and the panel-mounted indicator:
With chromel joining to chromel and alumel joining to alumel at the head, no
dissimilar-metal junctions are created at the thermocouple. However, with
chromel and alumel joining to copper at the indicator (again), the reference
junction has been re-located to its rightful place. This means the thermocouple
head’s temperature will have no effect on the performance of this measurement
system, and the indicator will be able to properly compensate for any ambient
temperature changes at the panel as it was designed to do. The only problem with
this approach is the potential expense of thermocouple-grade cable. This is
especially true with some types of thermocouples, where the metals used are
somewhat exotic.

A more economical alternative, however, is to use something called extension-

grade wire to make the connection between the thermocouple and the receiving
instrument. “Extension-grade” thermocouple wire is made less expensive than full
“thermocouple-grade” wire by choosing metal alloys similar in thermo-electrical
characteristics to the real thermocouple wires within modest temperature ranges.
So long as the temperatures at the thermocouple head and receiving instrument
terminals never exceed a modest range, the extension wire metals joining to the
thermocouple wires and joining to the instrument’s copper wires need not be
precisely identical to the true thermocouple wire alloys. This allows for a
wider selection of metal types, some of which may be substantially less
expensive than the measurement-grade thermocouple alloys. Also, extension-grade
wire may use insulation with a narrower temperature rating than thermocouple-
grade wire, reducing cost even further.

An interesting historical reference to the use of extension-grade wire appears

in Charles Robert Darling’s 1911 text Pyrometry – A Practical Treatise on the
Measurement of High Temperatures. Darling describes “compensating leads”
marketed under the name Peake designed to be used with platinum-alloy
thermocouples. These “compensating” wires were made of two different copper-
nickel alloys, each copper-nickel alloy matched with the respective thermocouple
metal (in this case, pure platinum and a 90%-10% platinum-iridium alloy) to
generate an equal and opposite millivoltage at any reasonable temperature found
at the thermocouple head. Thus, the only reference junction in the thermocouple
circuit is where these copper-nickel extension wires joined with the indicating
instrument, rather than being located at the thermocouple head as it would be if
simple copper extension wires were employed.
Extension-grade cable is denoted by a letter “X” following the thermocouple
letter. For our hypothetical type K thermocouple system, this would mean type
“KX” extension cable:

Thermocouple extension cable also differs from thermocouple-grade (measurement)

cable in the coloring of its outer jacket. Whereas thermocouple-grade cable is
typically brown in exterior color, extension-grade cable is usually colored to
match the thermocouple plug (yellow for type K, black for type J, blue for type
T, etc.).
Side-effects of reference junction compensation

Reference junction compensation is a necessary part of any precision

thermocouple circuit, due to the inescapable fact of the reference junction’s
existence. When you form a complete circuit of dissimilar metals, you will form
both a measurement junction and a reference junction, with those two junctions’
polarities opposed to one another. This is why reference junction compensation –
whether it takes the form of a hardware circuit or an algorithm in software –
must exist within every precision thermocouple instrument.

The presence of reference junction compensation in every precision thermocouple

instrument results in an interesting phenomenon: if you directly short-circuit
the thermocouple input terminals of such an instrument, it will always register
ambient temperature, regardless of the thermocouple type the instrument is built
or configured for. This behavior may be illustrated by example, first showing a
normal operating temperature measurement system and then with that same system
short-circuited. Here we see a temperature indicator receiving a 4-20 mA current
signal from a temperature transmitter, which is receiving a millivoltage signal
from a type “K” thermocouple sensing a process temperature of 780 degrees
Fahrenheit:
The transmitter’s internal reference junction compensation feature compensates
for the ambient temperature of 68 degrees Fahrenheit. If the ambient temperature
rises or falls, the compensation will automatically adjust for the change in
reference junction potential, such that the output will still register the
process (measurement junction) temperature of 780 degrees F. This is what the
reference junction compensation is designed to do.

Now, we disconnect the thermocouple from the temperature transmitter and short-
circuit the transmitter’s input:

With the input short-circuited, the transmitter “sees” no voltage at all from
the thermocouple circuit. There is no measurement junction nor a reference
junction to compensate for, just a piece of wire making both input terminals
electrically common. This means the reference junction compensation inside the
transmitter no longer performs a useful function. However, the transmitter does
not “know” it is no longer connected to the thermocouple, so the compensation
keeps on working even though it has nothing to compensate for. Recall the
voltage equation relating measurement, reference, and compensation voltages in a
hardware-compensated thermocouple instrument:
Disconnecting the thermocouple wire and connecting a shorting jumper to the
instrument eliminates the VJ1 and VJ2 terms, leaving only the compensation voltage
to be read by the meter6:

This is why the instrument registers the equivalent temperature created by the
reference junction compensation feature: this is the only signal it “sees” with
its input short-circuited. This phenomenon is true regardless of which
thermocouple type the instrument is configured for, which makes it a convenient
“quick test” of instrument function in the field. If a technician short-circuits
the input terminals of any thermocouple instrument, it should respond as though
it is sensing ambient temperature.

While this interesting trait is a somewhat useful side-effect of reference

junction compensation in thermocouple instruments, there are other effects that
are not quite so useful. The presence of reference junction compensation becomes
quite troublesome, for example, if one tries to simulate a thermocouple using a
precision millivoltage source. Simply setting the millivoltage source to the
value corresponding to the desired (simulation) temperature given in a
thermocouple table will yield an incorrect result for any ambient temperature
other than the freezing point of water!

Suppose, for example, a technician wished to simulate a type K thermocouple at

300 degrees Fahrenheit by setting a millivolt source to 6.094 millivolts (the
voltage corresponding to 300o F for type K thermocouples according to the ITS-90
standard). Connecting the millivolt source to the instrument will not result in
an instrument response appropriate for 300 degrees F:

Instead, the instrument registers 339 degrees because its internal reference
junction compensation feature is still active, compensating for a reference
junction voltage that no longer exists. The millivolt source’s output of 6.094
mV gets added to the compensation voltage (inside the transmitter) of 0.865 mV –
the necessary millivolt value to compensate for a type K reference junction at
71 oF – with the result being a larger millivoltage (6.959 mV) interpreted by the
transmitter as a temperature of 339 oF.

The only way to properly use a millivoltage source to simulate a desired

temperature is for the instrument technician to “out-think” the transmitter’s
compensation feature by specifying a millivolt signal that is offset by the
amount of equivalent voltage generated by the transmitter’s compensation. In
other words, instead of setting the millivolt source to a value of 6.094 mV, the
technician should set the source to only 5.229 mV so the transmitter will add
0.865 mV to this value to arrive at 6.094 mV and register as 300 degrees
Fahrenheit:

Years ago, the only suitable piece of test equipment available for generating
the precise millivoltage signals necessary to calibrate thermocouple instruments
was a device called a precision potentiometer. These “potentiometers” used a
stable mercury cell battery (sometimes called a standard cell ) as a voltage
reference and a potentiometer with a calibrated knob to output low voltage
signals. Photographs of two vintage precision potentiometers are shown here:

Of course, modern thermocouple calibrators also provide direct entry of

temperature and automatic compensation to “un-compensate” the transmitter such
that any desired temperature may be easily simulated:
In this example, when the technician sets the calibrator for 300 o F (type K), it
measures the ambient temperature and automatically subtracts 0.865 mV from the
output signal, so only 5.229 mV is sent to the transmitter terminals instead of
the full 6.094 mV. The transmitter’s internal reference junction compensation
adds the 0.865 mV offset value (thinking it must compensate for a reference
junction that in reality is not there) and “sees” a total signal voltage of
6.094 mV, interpreting this properly as 300 degrees Fahrenheit.

The following photograph shows the display of a modern thermocouple calibration

device (a Fluke model 744 documenting process calibrator) being used to generate
a thermocouple signal. In this particular example, the thermocouple type is set
to type “S” (Platinum-Rhodium/Platinum) at a temperature of 2650 degrees
Fahrenheit:

The ITS-90 thermocouple standard declares a millivoltage signal value of 15.032

mV for a type S thermocouple junction at 2650 degrees F (with a reference
junction temperature of 32 degrees F). Note how the calibrator does not output
15.032 mV even though the simulated temperature has been set to 2650 degrees F.
Instead, it outputs 14.910 mV, which is 0.122 mV less than 15.032 mV. This
offset of 0.122 mV corresponds to the calibrator’s local temperature of 70.8
degrees F (according to the ITS-90 standard for type S thermocouple junctions).
When the calibrator’s 14.910 mV signal reaches the thermocouple instrument being
calibrated (be it an indicator, transmitter, or even a controller equipped with
a type S thermocouple input), the instrument’s own internal reference junction
compensation will add 0.122 mV to the received signal of 14.910 mV, “thinking”
it needs to compensate for a real reference junction. The result will be a
perceived measurement junction signal of 15.032 mV, which is exactly what we
want the instrument to “think” it sees if our goal is to simulate connection to
a real type S thermocouple at a temperature of 2650 degrees F.

Burnout detection

Another consideration for thermocouples is burnout detection. The most common

failure mode for thermocouples is to fail open, otherwise known as “burning
out.” An open thermocouple is problematic for any voltage-measuring instrument
with high input impedance because the lack of a complete circuit on the input
makes it possible for electrical noise from surrounding sources (power lines,
electric motors, variable-frequency motor drives) to be detected by the
instrument and falsely interpreted as a wildly varying temperature.

For this reason it is prudent to design into the thermocouple instrument some
provision for generating a consistent state in the absence of a complete
circuit. This is called the burnout mode of a thermocouple instrument.

The resistor in this circuit provides a path for current in the event of an open
thermocouple. It is sized in the mega-ohm range to minimize its effect during
normal operation when the thermocouple circuit is complete. Only when the
thermocouple fails open will the miniscule current through the resistor have any
substantial effect on the voltmeter’s indication. The SPDT switch provides a
selectable burnout mode: in the event of a burnt-out thermocouple, we can
configure the meter to either read high temperature (sourced by the instrument’s
internal milli-voltage source) or low temperature (grounded), depending on what
failure mode we deem safest for the application.

1The colors in this table apply only to the United States and Canada. A stunning diversity of colors has been “standardized” for each thermocouple
type depending on where else in the world you go. The British and Czechs use their own color code, as do the Dutch and Germans. France has its own
unique color code as well. Just for fun, an “international” color code also exists which does not match any of the others.

2By “oxidizing,” what is meant is any atmosphere containing sufficient oxygen molecules or molecules of a similar element such as chlorine or
fluorine.

3“Reducing” refers to atmospheres rich in elements that readily oxidize. Practically any fuel gas (hydrogen, methane, etc.) will create a reducing
atmosphere in sufficient concentration.

4Itshould be noted that no amount of engineering or design is able to completely prevent people from doing the wrong thing. I have seen this style of
thermocouple plug forcibly mated the wrong way to a socket. The amount of insertion force necessary to make the plug fit backward into the socket
was quite extraordinary, yet this apparently was not enough of a clue for this wayward individual to give them pause.
5Early texts on thermocouple use describe multiple techniques for automatic compensation of the reference (“cold”) junction. One design placed a
mercury bulb thermometer at the reference junction, with a loop of thin platinum wire dipped into the mercury. As junction temperature rose, the
mercury column would rise and short past a greater length of the platinum wire loop, causing its resistance to decrease which in turn would
electrically bias the measurement circuit to offset the effects of the reference junction’s voltage. Another design used a bi-metallic spring to offset the
pointer of the meter movement, so that changes in temperature at the indicating instrument (where the reference junction was located) would result in
the analog meter’s needle becoming offset from its normal “zero” point, thus compensating for the offset in voltage created by the reference junction.

6The effect will be exactly the same for an instrument with software compensation rather than hardware compensation. With software compensation,
there is no literal Vrjc voltage source, but the equivalent millivolt value is digitally added to the zero input measured at the thermocouple connection
terminals, resulting in the same effect of measuring ambient temperature.