2 types of sensors are mostly used for temperature measurement: thermocouples and Pt100 RTD sensors. This page gives more detailed information on:
- history of temperature measurement and the first good thermometer by Fahrenheit
- thermocouples (voltage measurement):
- operating principle of thermcouples
- correct connection of thermocouples (thermocouple cable, compensation cable, thermocouple connectors)
- common mistakes with thermocouples
- Pt100 resistance thermometer (RTD) and SPRT (Standard Platinum Resistance Thermometers for calibration purposes)
- measurement principles and correct connection of Pt100 and SPRT sensors (2-, 3- or 4-wire)
- differences between thermocouples and Pt100/SPRT sensors
- accuracy classes of thermocouples and Pt100/SPRT sensors
History of temperature measurement
subjective temperature measurement
Within a reasonable temperature range we can give a rough estimate of how warm something is by touching it. But this can be unreliable: if you put one hand in cold water and one in warm water, and puth both hands in lukewarm water after that, one hand will give you a too hot feeling, while the other one will feel like it is too cold.
For things that are too hot to touch we can often get an impression of how hot it is by moving closer and feeling the radiated heat.
When the temperature rises enough things begin to glow, we can then see that it is hot!
The problem with these subjective temperature experiences is that they are possibly not the same for everyone. When even our hands cannot decide if something is hot or cold, how big is then the chance that a group of people can define an uniform standard?
We have to develop an apparatus that reacts to temperature in a simple and measurable way – we need a thermometer.
See temperature differences: Philo of Byzantium (Alexandria, classical antiquity)
One of the first documented experiments with temperature measurement is the thermoscope of the Greek engineer Philo of Byzantium (approx. 280-220 before Christ)
He took a hollow ball made of lead that was connected to the end of a pipe with an hermetic seal. The other end of the pipe was under water in another vessel.
To cite Philo:
- When you expose the ball to the Sun, a part of the air that is trapped in the tube will come out when the ball becomes hot. This can be seen by the fact that the air from the tube sinks in the water, brings it in motion and produces bubbles.
- When you now put the ball back in the shade again, in a place where the sunrays cannot reach it, the water will rise and flow through the tube.
- It doesn’t matter how often you repeat this action: the same will happen.
- Even more so, when you heat the ball with fire, or pour hot over over it, the result will be the same.
Take note that Philo did what a real scientific researcher should do: he checked if the experiment was reproduceable and determined that the expansion of air was a reaction to the heat applied to the ball and that it was independent of the source of the heat.
The first thermometer: Galileo Galileï (Italy, renaissance)
The thermometer was invented no later than 1612 in north-east Italy by Santorio Santorre or Galileo Galilei in Padua or Venice.
From that year, there are letters between Sagredo and Galileo discussing thermometers. Sagredo initially credits Santorio, but Galileo corrects him and claims to have invented them himself. Later Italian scientists consider Galileo the inventor of the thermometer.
That same year, Santorio published a book entitled Commentaries on Galenus, which describes a thermometer and its use in medicine. Santorio cites Hero as his inspiration for this device.
The Italian story is further complicated by a book Telioux wrote in Rome in 1611. His description of how it works is complete nonsense, and the picture he paints could not be reconstructed. Although this is earlier, he clearly did not invent it himself.
Since Telioux wrote from Rome, thermometers must have already spread throughout northern Italy. W.E. Knowles Middleton, author of A HISTORY OF THE THERMOMETER AND ITS USE IN METEOROLOGY (1966), believes the thermometer was most likely invented in the 1590s or 1600s, when Santorio and Galileo were studying together at the University of Padua. However, because Telioux’s diagram differs significantly from the Paduan thermometers, it is possible that it is a crude representation of another, independent invention from central Italy.
(left) thermometer by Santorio:
- The top sphere is full of air
- The bottom part of the tube contains colored water
- The wider part, at the bottom, is connected to the water in the tube and is in contact with the air
- The thermometer has no built-in scale, so strings were used to indicate the different values.
(right) Telioux’ drawing of a thermometer:
- Consists of two long-necked flasks nested inside each other. Note that you cannot connect these two flasks this way, as both have a flared tip.
- The top flask contains air
- The bottom flask contains water.
- There is clearly a scale to measure the water level between the two flasks
Thermometer experiment by Galileo
In a letter to Ferdinando Cesarini (ca. 1606-1646) dated September 20, 1638, Benedetto Castelli (1577/8-1643) described the use of the thermoscope: “I remember an experiment that our master Galileo showed me more than thirty-five years ago”:
- He took a glass bottle the size of a small hen’s egg, with a neck as long as two palms and as thin as a barley stalk.
- He warmed the bottle with palms of his hands and then turned its opening over into a vessel placed underneath, which contained a little water.
- As he released the vessel from the warmth of his hands, the water suddenly began to climb up the neck, rising to more than a palm’s height above the water level in the vase.
Galileo used this effect to create an instrument that could estimate the degree of heat and cold.”
Scale
Lateron, Galileo or his colleague Santorio Santorio placed a paper scale next to the tube to read changes in the water level. This certainly made it a thermometer, but who first came up with this idea is unclear (they both disagreed on that).
Actually, this type of thermometer is an air thermometer.
Problem: the air-thermometer is also a barometer
Aside from the fact that you have to add water every now and then, people wondered why this air thermometer would not be good for recording daily temperature changes.
The answer to that question is: because it is also a barometer! However, Galileo didn’t know back then that atmospheric pressure existed.
Torricelli, one of Galileo’s students, was the first to realize, shortly after Galileo’s death, that the real driving force behind the suction was external atmospheric pressure, a satisfactory mechanical explanation in contrast to the philosophical theory that “nature has an aversion for vacuum.”
In the 1640s, Pascal pointed out that the variability of atmospheric pressure made the air thermometer unreliable.
The solution: the first sealed thermometer with liquid in glass: Florentine thermometer
Thereafter the first experiments began with thermometers using a liquid in glass to exclude sensitivity to air pressure.
The first sealed thermometer with liquid in glass was produced in 1654 by a private academy of scientists, many of whom were former students of Galileo, under the sponsorship and active support of the Grand Duke of Tuscany, Ferdinand II (1610-1670). This was called the Florentine thermometer.
Florentine thermometers were sealed glass tubes filled with alcohol (wine liquor). They were the first thermometers that were not also barometers and therefore unaffected by air pressure.
Although an important development, the Florentine thermometer was inaccurate and had no standardized scale (scales varied between thermometers):
- the lowest point was the coldest day in Florence during the year a particular instrument was made
- the highest point was the hottest day of that year
Because of this measurement results were different between Florentine thermometers.
From this time onwards the first meteorological data were kept.
Uniformity in temperature measurement: Newton
The first systematic account of a series of different temperatures, “Degrees of Heat,” or the “Newton Scale,” was written by Newton but published anonymously in 1701. Presumably, he did not consider this project as important as some of his other achievements.
By setting the freezing point of water at zero, Newton found that the temperature of boiling water was almost three times that of the human body, and that the melting point of lead was eight times higher (actually 327°C, while 8 x 37 = 296°C, so that’s pretty good!). But for higher temperatures, such as those of a wood fire, he significantly underestimated the values.
He used liquid linseed oil in a glass thermometer heated to the melting point of tin (232°C). (Linseed oil only boils at 343°C, but that’s also its auto-ignition temperature!)
Newton attempted to estimate higher temperatures indirectly. He heated a piece of iron in a fire and then allowed it to cool in a steady breeze. He discovered that, at least at the lower temperatures he could monitor with his thermometer, the temperature decreased in a geometric progression. That is, if it took five minutes to drop from 80°C above the air temperature to 40°C above the air temperature, it took another five minutes to drop to 20°C above the air temperature, another five minutes to drop to 10°C above the air temperature, and so on.
He then assumed that this same pattern of temperature drop also applied at high temperatures beyond the range of his thermometer, and thus estimated the temperature of the fire and the glowing iron. This was not very accurate: he (under)estimated the temperature of the fire at about 600°C.
The first good thermometer and temperature scale: Daniel Gabriel Fahrenheit
The first really good thermometer, with mercury expanding from a bulb in a capillary, was made in Amsterdam in the early 1720s by the German physicist, inventor and instrument maker Daniel Gabriel Fahrenheit, who was born in Gdansk, Poland.
The life of Daniel Gabriel Fahrenheit (Gdansk, May 24, 1686 – The Hague, September 16, 1736) is so remarkable, we will therefore go into it in more detail.
A troublesome childhood
Young Fahrenheit had a turbulent, delinquent childhood. He had lost his parents at a young age, who had died in 1701 from eating poisonous mushrooms. His guardians, appointed by the government in Gdansk, forced him to follow in his very wealthy father’s footsteps and also enter the commerce trade.
Daniel was an excellent student with a particular interest in science; he had absolutely no desire to enter commerce. In 1702, he was apprenticed against his will to Herman van Beuningen, a member of a prominent Dutch merchant family, and sent to Amsterdam to learn bookkeeping and trade.
Fahrenheit's first contact with Florentine thermometers
During his studies, the young Fahrenheit first encountered Florentine thermometers, which were highly sought after in Europe.
The major problem with Florentine thermometers, however, was their widely varying scales, which meant that measurements between thermometers could not be compared properly.
During the last years of his studies Fahrenheit began experimenting with ways to create a better thermometer.
A better thermometer
Fahrenheit’s goal was to create a thermometer with the following properties:
- easier to manufacture
- easy to calibrate
- more reliable
- more accurate
- high reproducibility
Fahrenheit's first thermometers: 1705-1706
His first thermometers, like all other thermometers of that time, used alcohol distillates of wine.
In 1706 he began producing and shipping thermometers equipped with a Florentine temperature scale.
Cruquius and Fahrenheit: 1705
Hydraulic engineer Nicolaus Cruquius was the first to use the Fahrenheit thermometer for meteorological observations. His series of measurements began in Delft at the end of 1705, giving the Netherlands three centuries of temperature measurements.
Arrest warrant and request for deportation to the East Indies: 1707
According to a letter his guardians sent to the mayor of Gdansk in 1706, Daniel got into trouble as soon as he set foot in Amsterdam.
According to his guardians, he stole money and repeatedly ran away, not to mention his behavior, which was so repulsive he couldn’t even describe it. According to later biographers, Fahrenheit still managed to complete his four years of study, but in his final years he had already been making and selling meteorological instruments. The money his guardians claimed he stole, as it turned out, was actually borrowed to finance his thermometer experiments.
When he was unable to repay the debt, his guardians were forced to pay it from his inheritance.
Fahrenheit’s guardians grew tired of dealing with him and decided to send him abroad with the Dutch East India Company. In 1706, they wrote to the mayor of Gdansk, requesting the city council to issue a warrant for Daniel Fahrenheit’s arrest and deportation to the East Indies. In January 1707, the mayor complied with the request.
On the run and visit to Danish astronomer Ole Roemer: 1708
Now 20 years old and with a warrant out for his arrest, Daniel traveled constantly through Germany, Sweden, and Denmark, far from the reach of the authorities. In 1708, he visited the Danish astronomer Ole Roemer in Copenhagen.
Roemer was not only the first astronomer to determine the speed of light, he also worked on thermometers and invented his own temperature scale.
Since Florentine thermometers used widely varying scales (the lowest point being the coldest day in Florence and the highest point the hottest day of the year of manufacture), Roemer decided to introduce a fixed scale of 60 degrees:
- 0: freezing brine
- 60: boiling water
- and because water froze at 1/8 from the bottom of the scale (0) towards the top (60), he divided it into eight parts of 7.5 degrees each.
It was Roemer who told Fahrenheit that there was a great demand for accurate thermometers. It was this visit with Roemer that inspired Fahrenheit to create -even better- thermometers.
Arrest warrant revoked
Shortly after Fahrenheit’s visit to Roemer, the warrant against his arrest was withdrawn.
It is possible that Roemer had a hand in this, as he was not only an important figure in the scientific community, but also the chief of the Copenhagen police.
Traveling and solving technical problems with thermometers 1709-1712
In 1709, Fahrenheit returned to Gdansk and made observations with his barometers and thermometers. In 1710, he traveled further, and in 1711, he returned to Gdansk to settle his parents’ estate. After another trip to Königsberg and Mitau in 1711, he returned to Gdansk in 1712 and stayed there for two years. During this period, he worked on solving technical problems with his thermometers.
First experiments with mercury thermometers and learning glassblowing techniques 1713-1714
All of Fahrenheit’s research and experiments led him to doubt the reliability of the freezing and boiling points of water (which are dependent on atmospheric pressure, but that was not known back then).
Fahrenheit therefore began experimenting with mercury thermometers in 1713. At that time, Fahrenheit used a modified version of Roemer’s scale for his thermometers, which would later evolve into his own Fahrenheit scale.
Creating a capillary with a uniform diameter is not easy. Therefore, in 1714, Fahrenheit left Gdansk for Berlin and Dresden to work closely with the glassblowers there. Fahrenheit learned glassblowing so he could make his own thermometer tubes.
Ultimately, it was Fahrenheit who was the first to succeed in manufacturing thermometers with completely identical thermometer tubes.
In that year, 1714, philosopher Christian Wolff wrote about Fahrenheit’s thermometers in a journal after receiving some of his alcohol-based instruments, which helped cement Fahrenheit’s reputation in the scientific community.
Boerhaave and Fahrenheit: ever-improving thermometers
Fahrenheit had a lot of contact with the physician and professor Herman Boerhaave, who saw the human body as a kind of mechanical machine with a pump and valves rather than a divine miracle.
Boerhaave realised better than anyone that the medical use of body temperature thermometers was of course an important tool for diagnosis and medical research.
Boerhaave gave Fahrenheit a lot of good advice and as a result his thermometers became better and better.
The first two thermometers with identical results: 1714
In 1714, Fahrenheit demonstrated his success by producing two thermometers that produced virtually identical results across the entire temperature range, despite their different operating ranges. This was a major breakthrough.
That same year he had another breakthrough when he replaced alcohol with mercury.
The first mercury thermometer: 1714
- When constructing thermometers and defining temperature scales, one must take into account that not all liquids expand uniformly when heated.
- For example, water first contracts when heated from its freezing point and then begins to expand around forty degrees Fahrenheit.
- Fahrenheit got the idea to use mercury from a colleague’s comment that one had to correct the barometer reading to account for the variation in the density of water with temperature.
- Fahrenheit completed a series of investigations based on the work of Danish astronomer Ole Roemer. In these investigations, he determined the boiling point of water and other liquids and studied the expansion (increase in volume) of mercury.
- These experiments led to the discovery that the boiling point of water varies with changes in atmospheric pressure.
- Fahrenheit also discovered the method for supercooling water—that is, cooling water below its normal freezing point without it turning into ice. He did this by adding sea salt to the ice water.
- During Fahrenheit’s research, it was found that mercury (quicksilver, as it was then called) contracts and expands more evenly than alcohol.
- If the expansion and contraction of mercury on the temperature scale is assumed to be uniform, it is found to agree very closely with the true temperature scale.
- Mercury is also more accurate over a wider range, allowing measurement of temperatures below the freezing point of brine and above the boiling point of water.
- This allowed the Fahrenheit mercury thermometer to accurately record lower and higher temperatures, with much finer gradations in between.
Other projects and financial problems: 1715
Besides his interest in meteorological instruments, Fahrenheit also worked on ideas for a mercury clock, a perpetuum mobile, and a heliostat around 1715.
He began a correspondence with Leibnitz about some of these projects. The correspondence reveals that Fahrenheit ran out of money while working on his projects and asked Leibnitz for help in obtaining a permanent position so he could continue his work.
Fahrenheit's workshop in Amsterdam: 1717 or 1718
- In 1717 or 1718, Fahrenheit returned to Amsterdam and started a workshop there. From 1718 onward, he also taught chemistry (alchemy at the time).
- He began selling barometers, hydrometers, and his mercury and alcohol thermometers commercially.
- Fahrenheit’s thermometers were so successful because of his method of purifying mercury.
- He also introduced the use of cylindrical shapes at the bottom of his thermometers instead of the usual spherical ones.
By 1721, Fahrenheit had perfected the process of manufacturing and standardizing his thermometers.
However, he did not make his detailed process for making thermometers public until many years later, as he wanted to keep his methods secret.
The creation of the Fahrenheit scale in 1724
- All of Fahrenheit’s experiments and research led him to doubt the reliability of the freezing and boiling points of water.
- Fahrenheit was inspired by Roemer’s scale to create one of his own that was less awkwardly laid out and had three fixed reference points:
- 0°F: ice-cold brine (a mixture of ice, water, and sea salt)
- 32°F: ice cold water (melting ice in water)
- 96°F: (later corrected to 98.6°F): “blood heat” or body temperature of a healthy human.
- This last reference point was later dropped when the boiling point of water was placed at approximately 212°F (an arbitrarily chosen point).
- He chose to use a factor of eight instead of Roemer’s 7.5 and then divided each degree by four.
- Multiply eight by four and you get 32.
- That coincidental set of circumstances caused the freezing point on the Fahrenheit scale is 32 degrees.
The superiority of his mercury thermometers over alcohol-based thermometers made them very popular, leading to the widespread adoption of his Fahrenheit scale.
Member of the Royal Society and controversy
Fahrenheit spent the rest of his life in Amsterdam.
In 1724 he visited England and was elected a fellow of the Royal Society on the 5th of May. That year he published no fewer than five articles in Latin for the Royal Society’s scientific journal, Philosophical Transactions, on various subjects.
In his second article, “Experimenta et observationes de congelatione aquæ in vacuo factæ” he describes his thermometers and the reference points he used to calibrate them. For two centuries, this document remained the only description of Fahrenheit’s thermometer-making process.
According to this 2nd article of Fahrenheit from 1724, his scale is based on three reference points:
- The zero point was obtained by placing the thermometer in a mixture of ice, water and ammoniumchloride (a salt), which has a temperature of −17.8 °C. This was a common method of obtaining a low temperature at the time, and Fahrenheit was able to use it during his travels.
- The second reference point, 32°F, he found by placing the thermometer in pure water on which ice was forming (the temperature of the water when ice is floating on it at the constant freezing point).
- The third and final reference point was determined by the body temperature of a healthy person, measured in the mouth or under the armpit. The temperature difference between this third point and the second point is about twice as great as between the zero point and the second point, or about 64°F. He set this third point at 96°F (later corrected to 98.6°F).
In this way, the difference between the melting point of ice (32°F) and body temperature (96°F) was 64 degrees, which allowed him to apply his scale to his instruments by dividing the interval six times in half. Water boils at (approximately) 212°F.
In the 20th century, Dutch chemist Ernst Cohen discovered correspondence between Fahrenheit and physician and professor Herman Boerhaave. This correspondence raised significant doubts about the originality of Fahrenheit’s article explaining the reference points for his scale. Furthermore, there were doubts about the fact that Fahrenheit had not invented his scale himself, but had largely derived it from the scale of Danish astronomer Roemer. This was likely due to the instrument maker’s desire to keep his methods secret so they could not be copied.
Fahrenheit's Death and Will: 1736
From August 1736, Fahrenheit stayed in the house of Johannes Frisleven on the Plein in The Hague in connection with a patent application from the States of Holland, for a machine to pump water from polders.
In early September he fell ill and on the 7th his health had deteriorated to such an extent that he called notary Willem Ruijsbroek to draw up his will.
On the 11th, the notary came by again to make some changes. For those interested, here’s the will, with Fahrenheit’s shaky signature underneath, which clearly shows he was seriously ill.
Five days later, Fahrenheit died at the age of fifty. He received a fourth-class funeral, meaning he was destitute, in the Kloosterkerk in The Hague four days later. Fahrenheit had never been married.
A memorial plaque was unveiled in his honor in this church in 2002. A large Fahrenheit thermometer hangs outside his grave (pictured right). We assume it works and provides the temperature of the ambient air, not the Fahrenheit temperature of his remains.
Known examples of remaining Fahrenheit thermometers
Due to the materials used, virtually no original Fahrenheit thermometers have survived the test of time. Only three Fahrenheit thermometers are known to exist worldwide.
The two thermometers from 1718 (middle) and 1727 (left) are currently in the Boerhaave Museum, Leiden, the Netherlands.
In 2012, an original Fahrenheit thermometer, with his signature on the back and an unknown date of manufacture, was auctioned at Christie’s (pictured far right). A picture of this thermometer can still be found on the auction website, and it’s fully zoomable. Note that when the photo was taken, the temperature was 78°F, so it was a warm auction house. The thermometer sold for £67,250 ($107,802), a respectable sum. We expect it attracted enthusiastic bidding.
Where is the Fahrenheit scale still used?
Temperature measurement in degrees Fahrenheit is now only used in the United States of America and the older generation of the United Kingdom, Australia, South Africa, Canada, Ireland, India and Jamaica because the Fahrenheit temperature scale was also previously in use in these countries.
For this reason, temperature controllers are still switchable between degrees Celsius and Fahrenheit.
The heat manager for heat treatments can also be switched to Fahrenheit for use in the United States and Canada.
The Réaumur scale & the Lyon thermometer
In 1730, the French entomologist Réaumur proposed a temperature scale based on the freezing (0 °R) and boiling (80 °R) points of water. This scale is called the Réaumur temperature scale.
Reaumur chose alcohol (ethanol) because it expands more and thus the difference in temperature was more visible.
Because of this, his thermometers were always very large and long.
This scale has been in use in Europe for a long time, especially in France, Germany and Russia.
The Celsius scale
In 1742, a Swedish scientist named Anders Celsius (1701-1744) developed a thermometer scale that divided the freezing and boiling points of water into 100 degrees.
Celsius chose 0 degrees for the boiling point of water and 100 degrees for the freezing point.
In 1743, French physicist Jean-Pierre Christin had independently developed a scale where 0 represented the melting point of ice and 100 the boiling point.
In 1744, the Swedish physician and biologist Carl Linnaeus proposed inverting Anders Celsius’s temperature scale, so that 0 represented the freezing point of water and 100 the boiling point.
In 1948, Cristin’s modified Celsius scale was internationally agreed upon and is still used today.
Absolute zero temperature: the Kelvin scale
The 19th-century British scientist Lord Kelvin developed and first proposed the (Kelvin) scale in 1848. The increments per degree correspond to the increments of the Celsius scale. In the early 20th century, it was therefore often called the “absolute Celsius” scale.
In 1859, the Scottish engineer and physicist Rankine proposed a similar Rankine scale, but in the step size of the Fahrenheit scale.
The kelvin was formally added to the International System of Units in 1954, with 273.16 K being defined as the triple point of water. Subsequently, the Celsius, Fahrenheit, and Rankine scales were redefined in terms of the Kelvin scale using this definition.
Thermocouples
History
Around 1821, the German physicist Seebeck experimented with magnetism.
He discovered that the needle of a compass would begin to rotate if he held it near a circuit made up of several metals and only if he heated one of the metals.
Later, it was discovered that a very small current flowed when one of the metals was heated. This effect was called the Seebeck-effect
Thermocouple measuring principle
The image makes this Seebeck effect a little clearer:
- A rotary coil voltmeter has 2 copper wires connected to wires of 2 different metals.
- the two metals are brought into contact with each other: the thermo-couple, and this point is heated
- a very small current starts to flow and that makes the meter deflect
Cold-junction
Note: Only the temperature difference is being measured this way. To determine the actual absolute temperature, the ambient temperature at the point where metals A and B transition into copper wires must be added to the measured temperature.
The point where the thermocouple metals change into copper must be measured, this is the cold-junction temperature.
Here’s a practical example with an RKC Instrument thermometer:
- On the left in the image a tube thermometer for the cold junction temperature
- Read the rotary coil voltmeter: for example 100°C, add the ambient temperature value from the tube thermometer: for example 21°C.
- So the actual absolute temperature of the thermocouple is 100+21=121°C
Voltage dependent on temperature difference
International standard tables are available to convert the measured voltage into a temperature difference.
- Read the measured voltage in mV on the meter. For example: 6.1 mV
- Find the box in the table that most closely matches 6.1mV. That is 50°C on the x-axis and 100°C on the y-axis.
- So the temperature difference is: 150°C
- Add to this the ambient temperature where the thermocouple transitions into copper (often: connection point on the controller): suppose it is 25°C there, the actual absolute temperature of the thermocouple is then 175°C.
Temperature measurement with a standard multimeter in the millivolt mV measuring range
- In case of emergency, you can use your multimeter in millivolt (mV) mode to measure the temperature difference with a thermocouple.
- Use the appropriate table (there is a table for each type of thermocouple) to read the measured temperature difference.
- Note that for the actual absolute temperature, you must also add the ambient temperature of the thermocouple’s connection point to the multimeter: this is the cold-junction compensation!
See the appendix at the back of our general catalogue, starting on page 143, for measurement conversion tables.
Automatic cold-junction compensation
Modern electronic equipment, such as this RKC Instrument FB series temperature controller, automatically performs this cold junction compensation correction. An additional temperature sensor is often located at the terminals to measure the ambient temperature. The electronics automatically adds this temperature to the measured voltage at the thermocouple terminals.
Modern controllers also come with the correct millivolt to temperature tables and will automatically convert voltage to temperature for you.
Correct connection of thermocouples
You should carefully consider certain things when connecting thermocouples. Mistakes are often made when connecting thermocouples, such as:
- terminal block used to extend a thermocouple
- wrong type of thermocouple compensation cable used
- polarity connected incorrectly
- wrong type of thermocouple measuring input set on the controller
Extending thermocouple wire: do not introduce other metal types!!!
- The operating principle of a thermocouple is based on the generation of a small current caused by a temperature difference across two different metals.
- If other metals are now introduced into the wires, for example, by extending the thermocouple wires with the notorious terminal block, two additional metal junctions (thermocouple wire-copper) are introduced. This junction also generates a temperature-dependent current.
- This means that not only the desired temperature of the actual thermocouple connection point is measured, but additional temperature deviations are also introduced due to the additional metal transitions in the terminal block.
- This problem also occurs if you connect a thermocouple with plain copper wire.
- For maximum accuracy, connect thermocouple wires to the controller without using ferrules. Even ferrules (which are made of a different metal) can introduce a slight deviation.
Thermocouple compensation cable: use the correct type!
- To connect thermocouples correctly, the correct type of thermocouple compensation cable must be used.
- Each type of thermocouple has its own type of thermocouple compensation cable.
- The color of the cable and the wire colors indicate for which type of thermocouple the cable is suitable.
- There are several international standards in use, make sure you use the correct standard!
Thermocouple polarity
- A thermocouple generates a small direct current depending on the temperature difference.
- This current only flows in one direction.
- Therefore, it is important to observe the correct polarity of the thermocouple when connecting a thermocouple to a controller.
- If the polarity is connected incorrectly, the measurement on the controller will often go downwards while the temperature across the thermocouple goes up.
- This is easily checked by heating the thermocouple with a cup of hot water or a lighter. The temperature on the reading should then increase.
- If you are unsure about the polarity, you can use the above mentioned measuring method using a multimeter in millivolt (mV) mode with warm water or a cigarette lighter to determine the correct polarity.
- If you carefully feel the roughness of the thermocouple wires with your fingers, you’ll notice a difference. One wire feels often rougher than the other. This can also be used to determine the polarity of the thermocouple.
For different types of thermocouple wire and compensation cable, see the document below from page 81.
Thermocouple connectors
- The correct connection of thermocouples must also be done using the correct thermocouple connectors and chassis parts.
- Each type of thermocouple has its own type of connectors and chassis parts.
- The color of the connector or chassis part corresponds to the type of thermocouple.
- The conductors in the thermocouple connectors are of the same material as the compensation cable therefore the measured temperature is not influenced.
- You can use thermocouple connectors to extend thermocouple wires in a proper way.
Wrong type of thermocouple measuring input set on controller
- It’s also crucial, of course, that you select the correct type of thermocouple input for the controller’s measurement input. Otherwise, the wrong millivolt/temperature difference conversion table will be used, resulting in measurement deviation and potentially even a dangerous situation (actual temperature higher than measured!!!).
- For this, see parameter group F21. As an example, here is an instruction for an RKC Instrument RB controller.
Resistance thermometers: Pt100, (S)PRT reference standards
Resistance thermometer classifications:
- Ultra Precise Platinum Resistance Thermometers (UPRTs):
- wide measuring range: −200 °C to 1000 °C
- Accuracy +/- 0.001 °C over the entire measuring range.
- UPRTs are developed for laboratory use
- Standard Platinum Resistance Thermometers (Standard PRTs or SPRTs) laboratory standards:
- measuring range: −200 °C to 670 °C
- Accuracy +/- 0.03 °C over the entire measuring range.
- (SPRTs are built like UPRTs, but with cheaper materials.
- Industrial Platinum Resistance Thermometers (PRTs):
- Designed to withstand industrial environments.
- Can be almost as durable as a thermocouple.
History
Simultaneously with the Seebeck-effect , Humphry Davy discovered in 1821 that the resistivity of metals is dependent on temperature.
The practical application of this property in electrical conductors was first described by Wilhelm Siemens during the 1871 Bakerian Lecture to the Royal Society in Great Britain, where Siemens suggested platinum as a suitable element. Siemens also presented a three-term interpolation formula.
However, the Siemens resistance thermometer quickly fell out of favor due to the instability of the temperature measurement.
Hugh Longbourne Callendar developed the first commercially successful platinum resistance thermometer (RTD) in 1885–1886. This resistance thermometer could measure accurately from -190°C to +660°C (he later improved it to -200°C to +1000°C).
Temperature dependent resistance
The following table shows resistance values for a Pt100 sensor from 0°C up to 800°C.
Measuring principles: 2, 3 or 4 wire
An important aspect that is very often overlooked is the measuring principle.
A Pt100 measurement is often chosen precisely for its accuracy. However, if it’s connected in a 2-wire configuration, this higher accuracy is completely negated!
In this case we therefore recommend a 3- or 4-wire measuring configuration.
2-wire: 1 ohm extra resistance: approximately 3°C deviation!!!
- The simplest configuration with a resistance thermometer uses two wires.
- This measuring principle is only used when high accuracy is not required.
- the resistance of the connecting wires is added to that of the sensor, leading to large measurement errors.
- Note that at 200°C an extra resistance of 1 ohm is already about 3°C!!!
3-wire: measuring lines of the same length compensated
- To minimize the effects of line resistance, a three-wire configuration is used.
- If the three connecting wires to the Pt100 are of the same thickness and length, this connection principle is more than sufficient for good accuracy: the 4-wire measuring principle is then not necessary.
- The suggested setting for the configuration shown is with R1 = R2 and R3 around the middle of the range of the Pt100/RTD.
- Looking at the Wheatstone-bridge circuit shown, the voltage drop at the bottom left is V_rtd + V_lead and at the bottom right is V_R3 + V_lead. The bridge voltage (V_b) is therefore the difference: V_rtd − V_R3. The voltage drop due to the resistance of the leads is compensated for. This always applies when R1 = R2 and R1, R2 >> RTD, R3.
- R1 and R2 can be used to limit the current through the RTD. For a Pt100, for example, limiting it to 1 mA and 5 V would suggest a limiting resistor of approximately R1 = R2 = 5/0.001 = 5000 ohms.
4-wire: compensation of different measuring lines
- The four-wire measurement configuration increases the accuracy of the resistance measurement when there are differences in the measurement leads.
- eliminates voltage drop through the test leads.
- To further enhance accuracy, any residual thermoelectric voltages generated by different wire types or screw connections are eliminated by reversing the direction of the 1 mA current and the test leads to the DVM (digital voltmeter). The thermoelectric voltages are generated in only one direction. By averaging the reversed measurements, the thermoelectric error voltages are eliminated.
Always choose at least a 3-wire Pt100
When in doubt, always choose a 3-wire Pt100 sensor. This offers an excellent compromise between cost and accuracy. If the sensor’s three wires are identical, a 4-wire sensor is unnecessary; only choose 4-wire if the highest precision is required.
A 2-wire sensor can quickly exhibit a deviation of several degrees due to the wiring alone, which can also increase over time. The higher accuracy of a Pt100 sensor is quickly negated by the wiring alone. For such applications, always choose a thermocouple!
Advantages, disadvantages and differences between thermocouples and resistance thermometers
Thermocouple
- Measuring range: -180 to +2320°C
- Accuracy: Standard +/- 2°C. With reference standard also +/- 0.2°C possible.
- Stability: good
- Response time: < 1 second or even 0.03 seconds
- Dimensions: very small <1.6mm or even 50um
- Resistance to shock: very good
- Relatively cheap
Pt100/RTD resistance thermometer
- Measuring range: -200 to +500°C (difficult above 660°C due to possible contamination of the platinum)
- Accuracy: +/- 0.001°C (UPRT), +/- 0.0015°C (SPRT)
- Stability: very good, very low drift
- Response time: a few seconds
- Dimensions: 3-6mm minimum
- Resistance to shock: very poor
- Generally more expensive than thermocouples
Thermocouple accuracy classes
- The IEC 60584-1 standard defines three accuracy classes: 1, 2 and 3
- An accuracy class determines the maximum allowable error that a thermocouple may exhibit within its calibrated range.
The table below summarizes the IEC 60584-1 tolerances for three common thermocouple types:
- K (chrome-alumel)
- J (iron-constantan)
- T (copper-constantan)
Class 3: cryogenic applications
- Class 3 is primarily intended for cryogenic applications (approximately -200 °C to +40 °C).
- Base metal thermocouple materials may not meet Class 3 tolerances for temperatures below -40°C without specific selection processes. Therefore, a specific order and material selection are essential if Class 3 compliance is required.
- Due to the difficulty of producing Class 3 compliant thermocouples, they are often more expensive.
Choose the correct thermocouple class 1, 2 or 3
Class 1: very small tolerances
- Very small tolerances
- Suitable for applications with:
- a high reproducibility
- minimal drift
- accurate temperature measurements
- Typical applications:
- calibration laboratories
- aerospace testing
- precision metallurgy
- semiconductor production
Class 2: average tolerances
- Average tolerances: provides a balance between acceptable accuracy and lower costs compared to Class 1.
- Typical Accuracy: Generally tolerances of approximately ±2.5°C or ±0.0075|t|, sufficient for most common industrial needs.
- Typical applications:
- industrial process control
- temperature monitoring in the food industry
- HVAC systems
Class 3: cryogenic applications
- Limited application: Cryogenic temperature measurements where extremely low temperatures must be measured accurately.
- Higher costs due to challenges in material selection and strict quality requirements.
- Typical applications:
- Cryogenic storage
- liquefied gas processing
- specialized scientific research
Pt100 accuracy classes
Standards define a kind of ideal Pt100 sensor that manufacturers can strive for. If it were possible to build the ideal sensor, tolerance classes would be irrelevant.
Because Pt100 sensors cannot be adjusted to compensate for errors, you must purchase a sensor with an accuracy suitable for your application.
Sensor errors can be corrected in some measuring instruments using certain coefficients.
Pt100 sensors are available in various accuracy classes, the most common of which are AA, A, B and C, as defined in the IEC 60751 standard.
There are also so-called fraction tolerance classes, such as 1/3, 1/5, and 1/10. These were previously standardized classes in, for example, DIN 43760:1980-10, which was withdrawn in 1987, but they were not defined in the later standard IEC 60751.
In any case, the IEC 60751:2022 standard defines these tolerance classes in section 5.2.3.3 (Marking of thermometers). The tolerance of these sensors is based on the sensor with accuracy class B, but both the fixed part (0.3°C) and the relative part of the error are divided by the given number (3, 5, or 10). The tolerance classes of these sensors are therefore:
A sensor manufacturer can of course produce sensors with his own accuracy classes.
The formulas are difficult to compare. Therefore, the accuracy classes are calculated based on temperature (°C) in the table below:
Sources
A HISTORY OF THE THERMOMETER AND ITS USE IN METEOROLOGY by W. E. Knowles Middleton (1966)
Timeline of temperature and pressure measurement technology – wikipedia
“Thermoscoop” experiment Philo van Byzantium (280-220 voor Christus)
The life of Fahrenheit – Christie’s September 2012 Fahrenheit Thermometer Auction Description
Important wikipedia page on Fahrenheit
Thermocouples – wikipedia
Resistance thermometers – wikipedia
Pt100 accuracy classes – Beamex
Thermocouple accuracy classes – Thermo Electric
