Big picture: why temperature measurement matters 🌡️

Temperature is one of the most common and important measurements in process industries. Many industrial processes depend on knowing temperature accurately because temperature affects reaction rates, pressure, viscosity, material stability, safety, and product quality.

The chapter says that from about −250°C to +650°C, temperature can usually be measured using either:

Above about 650°C, thermocouples and radiation pyrometers become more common.

The chapter mainly covers:

1. Thermocouple principles
2. Thermocouple types
3. Thermocouple installation
4. RTD principles
5. RTD construction
6. RTD installation
7. Thermowells
8. Practical comments on installation

16.1 Thermocouple principles ⚡🌡️

A thermocouple is made from two different metals joined together. When one junction is hotter than the other, a small voltage is generated.

This is called the Seebeck effect.

The Seebeck effect

Imagine two different metals joined at two ends, forming a loop:

• One junction is hot.
• The other junction is cold.
• A small electromotive force, or emf, is produced.
• This emf is measured in millivolts, so it is very small.

The important point is:

A thermocouple does not directly measure absolute temperature. It measures a voltage caused by a temperature difference between two junctions.

So the output depends on:

1. The temperature of the hot junction
2. The temperature of the cold/reference junction
3. The metal pair used

Hot junction and reference junction 🔥❄️

The hot junction is the measuring point. This is placed where you want to measure temperature.

The cold junction, also called the reference junction, is the known comparison point.

In laboratory conditions, the cold junction can be kept at 0°C using melting ice. But in industrial plants, this is inconvenient. Instead, the reference junction is usually kept at ambient temperature, and that ambient temperature is measured separately, often using an RTD.

This is called cold junction compensation.

Why high input impedance is needed

The emf from a thermocouple is tiny. If the measuring instrument draws current from the thermocouple, it can disturb the voltage and give a wrong reading.

So the thermocouple should be measured almost like an “open circuit” using a device with very high input impedance.

In simple terms:

The measuring instrument should “look but not touch” electrically. It should sense the voltage without significantly drawing current.

16.2 Thermocouple types 🔠

Although any two different metals can theoretically form a thermocouple, only certain combinations are commonly used. These combinations are chosen because they have:

• Good sensitivity
• Useful temperature range
• Chemical stability
• Predictable voltage-temperature behavior

The chapter lists common thermocouple types: E, J, K, R, and T.

Type E thermocouple

Metals: Chromel / Constantan Continuous range: 0 to 1100°C Short-term range: −270 to 1300°C

Type E is described as the most sensitive thermocouple type. That means it gives a relatively large voltage change per degree Celsius.

Advantages:

• High sensitivity
• Resistant to oxidation
• Resistant to moisture corrosion

Important installation note:

• Must use compensating cable

Good for situations where you need a strong signal and the environment is not too chemically aggressive.

Type J thermocouple

Metals: Iron / Constantan Continuous range: 20 to 700°C Short-term range: −180 to 750°C Accuracy: about 1% or 3°C

Type J is useful in:

• Vacuum atmospheres
• Reducing atmospheres
• Inert atmospheres

But there is a major limitation:

The iron part oxidizes above about 540°C.

It is also unsuitable for:

• Moist conditions
• Sulphurous conditions

So Type J is not ideal where corrosion or oxidation is a concern.

Type K thermocouple

Metals: Chromel / Alumel Continuous range: 0 to 1100°C Short-term range: −180 to 1350°C Accuracy: about 0.75% or 3°C

Type K is one of the most widely used industrial thermocouples.

Advantages:

• Good high-temperature range
• Resistant to oxidation, especially above 500°C

Limitations:

• Unsuitable for reducing atmospheres
• Unsuitable for sulphurous conditions

Type K is often a “general-purpose” thermocouple, but the environment still matters.

Type R thermocouple

Metals: Platinum / Platinum-Rhodium Continuous range: 0 to 1600°C Short-term range: −50 to 1750°C Accuracy: about 0.15% or 1°C

Type R is for high temperatures and more demanding applications.

Advantages:

• Very high temperature capability
• Suitable for oxidising and inert conditions

Limitations:

• Rapidly poisoned by reducing atmospheres
• Contaminated by metal vapours
• Needs non-metallic sheaths, such as alumina

Because it uses platinum and rhodium, it is more expensive, but it performs well at high temperatures.

Type T thermocouple

Metals: Copper / Constantan Continuous range: −185 to 300°C Short-term range: −250 to 400°C Accuracy: about 1% or 1°C

Type T is especially useful for low-temperature measurement.

Advantages:

• Works in vacuum
• Works in oxidising, reducing, moist, and inert atmospheres
• Has defined error limits below 0°C
• Can use copper extension leads

Limitation:

• Not resistant to acid fumes

Type T is often used when low temperatures need reliable measurement.

16.3 Thermocouple installation 🧰

Thermocouple installation matters a lot. Even if the thermocouple itself is correct, poor wiring or placement can produce bad measurements.

Third metal rule

The chapter explains that a third metal can be introduced at a junction without affecting the emf, as long as the added junctions are at the same temperature.

For example, the hot junction may be brazed or soldered. Copper can also be introduced at the reference junction to connect the thermocouple to a measuring instrument.

But there is a catch:

If the reference junction is split into two points, those two points must be kept at the same temperature.

If not, an extra unwanted voltage can be produced.

Reference junction placement

The reference junction should be kept away from heat sources.

In industrial installations with many thermocouples, the reference junctions are often gathered in a shared input cabinet. One RTD may be used to measure the cabinet temperature for reference compensation.

This makes the system easier to manage.

Compensating cable 🔌

Thermocouple wires are often thin and fragile. In a plant, the sensor may be far away from the control cabinet, so thicker extension leads are used.

Ideally, the extension leads would be made of the same metals as the thermocouple. But this can be expensive.

So industrial systems often use compensating cable.

Compensating cable has similar thermoelectric behavior to the thermocouple wire. This helps prevent distortion of the measured emf.

It also reduces corrosion effects caused by electrochemical cells between dissimilar metals.

Simple version:

You cannot just extend thermocouple wires with any random cable and expect accurate readings.

The cable type must match the thermocouple type.

Thermocouple sheaths 🛡️

Thermocouples are usually protected inside a sheath.

The sheath protects against:

• Chemical attack
• Physical damage
• Moisture
• Mechanical stress

A typical sheath is:

• Stainless steel
• Up to about 6 mm diameter
• Filled with mineral insulation, often magnesium oxide, MgO

MgO is useful because it provides:

• Electrical insulation
• Good heat transfer

Bonded vs insulated hot junctions

The hot junction may be:

1. Bonded to the sheath
2. Insulated from the sheath

A bonded junction gives faster response because heat transfers more directly.

But it can be harder to manufacture and may become detached.

An insulated junction is slower but can be more electrically isolated and robust.

Thermocouples for average temperature

If you want the average of three temperatures, there are two options:

1. Measure each thermocouple separately and calculate the average.
2. Wire three thermocouples in series and scale the resulting emf by 3.

The chapter explains that wiring them in series can produce an average temperature “at source,” which may be more accurate.

This is useful when you care about an average process temperature rather than one single point.

Thermocouples for temperature difference

Thermocouples can also be wired “back to back” to measure temperature difference directly.

Instead of measuring two absolute temperatures and subtracting them later, the wiring arrangement can produce a voltage proportional to the difference between two points.

This can be useful in heat exchangers, reactors, or flow systems where the temperature difference is more important than absolute temperature.

16.4 Resistance Temperature Devices: RTDs 🧪

An RTD, or resistance temperature device, measures temperature using the fact that electrical resistance changes with temperature.

The most important RTD in process automation is the platinum resistance thermometer.

Why platinum?

Platinum is used because it is:

• Chemically stable
• Predictable
• Accurate
• Repeatable
• Suitable over a wide temperature range

The standard industrial platinum RTD is the Pt100.

What does Pt100 mean?

Pt100 means:

• “Pt” = platinum
• “100” = 100 ohms at 0°C

So at 0°C, the sensor resistance is exactly around 100 Ω.

At 100°C, the chapter states that a Pt100 has a resistance of about 138.5 Ω.

So as temperature increases, resistance increases.

RTD equations

Above 0°C, the relationship is approximately:

R = R_0(1 + aT - bT^2)

Below 0°C, an extra correction term is added:

R = R_0(1 + aT - bT^2 - c(T - 100)T^3)

Where:

• (R) = resistance at temperature (T)
• (R_0) = resistance at 0°C
• (T) = temperature in °C
• (a), (b), and (c) are standard coefficients

You do not need to memorize the equations at first. The main concept is:

RTDs convert temperature into resistance, and for platinum this relationship is very well standardized.

RTD accuracy classes

The chapter mentions two tolerance classes:

Class A has tighter tolerance, so it is better for high-accuracy measurement.

Pt100 vs Pt10

For normal measurements up to around 600°C, Pt100 is commonly used.

From 600°C to 850°C, the chapter says Pt10 may be used.

Pt10 has:

• 10 Ω resistance at 0°C
• Thicker wire
• Better reliability at higher temperatures

The lower resistance makes measurement more challenging, but the thicker construction can survive harsh high-temperature service better.

16.5 RTD construction 🧵

There are two main types of platinum RTD:

1. Film type
2. Wire wound type

Film RTD

A film RTD is made by depositing platinum on a ceramic substrate.

Advantages:

• Flat structure
• Good for surface temperature measurements
• Compact

Limitations:

• Less reliable than wire wound RTDs
• Not suitable for conventional sheathed industrial probes

Film RTDs are useful when you need to measure the temperature of a surface.

Wire wound RTD

The wire wound RTD is the standard choice for accurate industrial temperature measurement.

It consists of:

• Platinum wire
• Wound around a glass or ceramic former
• Sealed with glass or ceramic coating

For high temperatures, ceramic materials are preferred.

Like thermocouples, wire wound RTDs are usually placed inside mineral-packed stainless steel sheaths.

The platinum leads are insulated with silica tubing to prevent short circuits.

In simple terms:

A wire wound RTD is like a carefully protected platinum resistance coil whose resistance tells you the temperature.

16.6 RTD installation 🔌

RTDs are resistance-based devices, so the resistance of the connecting wires can affect the measurement. This is a major installation issue.

The chapter explains three main wiring methods:

1. Two-wire RTD
2. Three-wire RTD
3. Four-wire RTD

Two-wire RTD

In a two-wire setup, two leads connect the RTD to a bridge circuit.

This is simple and cheap.

But there is a problem:

The resistance of the wires is added to the resistance of the RTD.

If the leads are short, this may not matter much.

But in process plants, the sensor may be far from the measurement electronics. Long cables add resistance, and that resistance can be falsely interpreted as extra temperature.

So two-wire RTDs are usually only acceptable for short lead lengths or low-accuracy applications.

Three-wire RTD

A three-wire RTD is the common industrial solution.

The three-wire arrangement places lead resistances on adjacent sides of the bridge circuit. If the lead resistances are equal, they cancel out.

This greatly reduces error from long cable runs.

Important limitation:

If the lead resistances are not equal, some error remains.

Still, for most industrial process measurements, three-wire RTDs are a good practical balance between accuracy and cost.

Four-wire RTD

Four-wire RTDs are used for high-accuracy measurements.

There are two methods described:

1. Null balance method

This method uses different wiring configurations and bridge balancing to calculate the true RTD resistance.

It reduces lead resistance errors, but small errors can still come from switching circuits and contacts.

2. Constant current method

A small constant current is passed through the RTD using two wires.

The voltage across the RTD is measured using two separate wires connected to a high-impedance instrument.

Because the voltage measurement draws almost no current, the resistance of the voltage leads has almost no effect.

Using Ohm’s law:

R = \frac{V}{I}

If current (I) is known and voltage (V) is measured, RTD resistance can be calculated accurately.

Potential issue:

The current can slightly heat the RTD itself. This is called self-heating.

Self-heating makes the RTD warmer than the process temperature, causing measurement error.

RTDs for temperature difference

RTDs can also be placed in adjacent arms of a bridge circuit to measure temperature difference directly.

This is similar in purpose to back-to-back thermocouples.

It is useful when the important value is:

\Delta T = T_1 - T_2

rather than either temperature individually.

16.7 Thermowells 🧱

A thermowell is a protective pocket inserted into the process equipment. The temperature sensor goes inside it.

The chapter describes a complete temperature probe assembly as having:

1. Thermowell
2. Insert
3. Head cap

What is the thermowell?

The thermowell is part of the plant installation. It is in direct contact with the process medium.

It must survive:

• Pressure
• Temperature
• Corrosion
• Flow forces
• Chemical attack
• Mechanical stress

Why use a thermowell?

Thermowells serve several important functions:

1. Protect the sensor

The thermowell protects the delicate thermocouple or RTD sheath from the process.

2. Allow maintenance without shutdown

The sensor insert can be removed while the thermowell remains in place.

This means the plant does not necessarily need to be shut down just to replace or inspect the sensor.

3. Prevent contamination

During maintenance, the thermowell prevents air, dirt, or other contaminants from entering the process.

4. Provide thermal contact

The thermowell transfers heat from the process medium to the sensor.

Thermowell installation

Thermowells may be:

• Screw-fitted
• Flanged

The dimensions often follow standards. The chapter mentions that the internal diameter is normally around 7 mm, which fits sensors with sheaths up to about 6 mm diameter.

The length depends on the application.

A key practical point:

The thermowell must be long enough to reach the part of the process whose temperature you actually want to measure.

If it does not reach the liquid or fluid properly, even a perfect sensor will give a bad reading.

Thermal response time: why probes can be slow 🐢

Thermocouples and RTDs themselves can respond quickly.

But in real installations, the sensor is inside:

1. A sheath
2. Mineral packing
3. A thermowell
4. Possibly an air gap

All these layers add thermal resistance and thermal mass.

The result is a slower response.

The chapter says the probe assembly can behave like a first-order system with a typical time constant of about:

• 0.5 to 1.0 min
• In extreme cases, up to 5 min

What does time constant mean?

A time constant describes how quickly a sensor responds to a temperature change.

After one time constant, the sensor has moved about 63% of the way toward the new temperature.

So if a thermowell assembly has a 1-minute time constant, it does not instantly show the real process temperature after a sudden change.

This matters in control systems. A slow temperature measurement can make control sluggish or unstable.

Improving thermal contact

The chapter describes two ways to improve response:

1. Spring loading

The sheath is spring-loaded so the sensor tip touches the bottom of the thermowell.

This improves heat transfer.

2. Heat transfer oil

A small amount of heat transfer oil can be added in the gap between the sheath and thermowell.

This reduces the insulating effect of air and improves heat transfer.

16.8 Final practical comments 🛠️

The final section emphasizes a very important engineering lesson:

Sensor accuracy is useless if the sensor is installed in the wrong place.

For example, if the probe does not reach into the liquid whose temperature is being measured, the system may measure air, wall temperature, or an unrepresentative region instead.

This can lead to wrong process decisions.

The chapter stresses that it is cheaper and less embarrassing to solve these issues during design and specification than after the plant is already running.

Head-mounted transmitters 📡

The chapter also mentions that modern systems often use head-mounted transmitters.

These are small electronic circuits mounted in the sensor head cap.

They can convert the raw sensor signal into standard industrial signals such as:

• 1–5 V
• 4–20 mA

For thermocouples

The head-mounted transmitter can provide:

• Amplification
• Filtering
• Linearisation
• Scaling

But the thermocouple measurement is still relative to local ambient/reference temperature.

For RTDs

The transmitter provides the bridge-type measurement needed to convert resistance into temperature.

The benefit is that the signal sent back to the control system is stronger and more standardized.

Instead of sending a tiny millivolt thermocouple signal or delicate resistance measurement over long distances, the system sends a robust industrial signal.

Thermocouple vs RTD: the key comparison ⚖️

A simple way to remember:

Thermocouple = rugged and cheap voltage sensor.RTD = accurate and stable resistance sensor.

Most important takeaways 🧠

1. Temperature measurement is essential in process industries.
2. Thermocouples work using the Seebeck effect. Two different metals produce a small voltage when their junctions are at different temperatures.
3. Thermocouples measure temperature difference, not absolute temperature directly. The reference junction temperature must be known or compensated.
4. Thermocouple type matters. Types E, J, K, R, and T differ in materials, temperature range, accuracy, and chemical resistance.
5. Correct thermocouple cable is important. Compensating cable prevents measurement errors when extending wires to a cabinet.
6. RTDs measure temperature through resistance. Platinum RTDs, especially Pt100, are the standard industrial choice.
7. Pt100 means 100 Ω at 0°C. At 100°C, it is about 138.5 Ω.
8. RTDs are more accurate than thermocouples but more expensive.
9. RTD wiring matters. Two-wire is simple but affected by lead resistance. Three-wire is common. Four-wire is best for high accuracy.
10. Self-heating can distort RTD measurements. Too much measurement current can warm the sensor itself.
11. Thermowells protect the sensor and allow maintenance without process shutdown.
12. Thermowells slow down the response. The sensor may respond quickly by itself, but the full assembly can be much slower.
13. Probe placement is critical. A highly accurate sensor in the wrong place gives useless data.
14. Head-mounted transmitters convert raw signals into standard industrial outputs. This makes signal transmission easier and more reliable.

Simple mental model 🌡️

Think of temperature measurement like asking: “How hot is the process really?”

A thermocouple answers by saying: “Based on the tiny voltage created between two metal junctions, here is the temperature difference.”

An RTD answers by saying: “Based on how much the platinum resistance changed, here is the temperature.”

A thermowell says: “I will protect the sensor, but I may slow down its response.”

And the engineer must ask: “Is the sensor actually touching or sensing the part of the process I care about?”

That last question is often the difference between a useful measurement and a misleading one.