Image in modal.

The wide temperature capabilities, ruggedness and relative low cost of thermocouples make them the most widely used temperature sensors. They serve in myriad applications: industrial processes, electric power generation, furnace monitoring and control, food-and-beverage processing, automotive instrumentation, aircraft engines, home appliances, and rockets, satellites and spacecraft. When it comes to applications that require the ability to withstand high temperatures, small size, fast response, and high vibration or shock resistance, most times, you will find thermocouples providing the temperature measurements.

To better understand the reasons why thermocouples are so widely used, it is important to understand how a thermocouple works. A thermocouple is a sensor that measures temperature. It consists of two different types of metals joined together at one end. When the junction of the two metals is heated or cooled, a voltage is created that can be correlated back to the temperature — making the thermoelectric circuit a simple, robust and cost-effective temperature sensor.

Thermocouples are manufactured in styles such as probes, probes with connectors, transition-joint probes, infrared sensors and bare-wire thermocouples as well as thermocouple wire. Due to the number of models and technical specifications, it is important to understand the basic structure, functionality and ranges to better determine the right thermocouple type and material for an application.

Also, thermocouples are available in different combinations of metals, or calibrations. The most common are the base-metal thermocouples, which are more commonly known as Types J, K, T, E and N. In addition, there are high temperature calibrations known as noble metal thermocouples. These include Types R, S, C and GB. Each calibration has a different temperature range and environment; the maximum temperature varies with the diameter of the wire used in the thermocouple (table 1, navigate using Page in the upper left and scroll bar at the bottom to see all thermocouple types and data).

thermocouples chart

*Click the image for greater detail

Thermocouple accuracy and range depend on:

  • The thermocouple alloys used.
  • The temperature being measured.
  • The sensor construction.
  • The sheath material.
  • The media being measured.
  • The state (liquid, solid or gas) of the media.
  • The diameter of either the thermocouple wire (if it is exposed) or the sheath diameter (if the thermocouple wire is sheathed).
thermocouple sensor

A thermocouple sensor consists of two metals joined at one end. When the junction is heated or cooled, a voltage is created that can be correlated back to the temperature.


Each calibration has a different temperature range. The maximum temperature varies with the diameter of the wire used in the thermocouple.

The only temperature a temperature sensor measures is its own temperature. As a result, the selection of a probe-style sensor vs. a wire-style sensor is a matter of how best to get the thermocouple junction to the process temperature you are trying to measure.

Using a wire-style sensor may be fine if the fluid does not attack the insulation or conductor materials. A wire-style sensor also can be used if the fluid is at rest (or nearly so) and the temperature is within the capability of the materials. But, if the fluid is corrosive, high temperature, under high pressure or flowing through a pipe, then a probe-style sensor — maybe even with a thermowell — will be a better selection.

It all comes down to how to best get the thermocouple junction to the same temperature as the process or material you are trying to measure in order for the measurement to be accurate and, therefore, worthwhile. So, equally important is the type of measuring junction: grounded, ungrounded or exposed (see sidebar).

Keep in mind that the thermocouple measures the temperature but is also part of a control loop. As a result, any heating or cooling can be applied to the process if the measurement rises or falls against the process setpoints that are used. While there are many types of control, the primary methods that will be described in this article are on-off control and PID control.


Due to the number of models and technical specifications, it is important to understand the basic structure, functionality and ranges to better determine the right thermocouple type and material for an application.

On-Off Control: The Simplest Method

Also called hysteresis control, on-off control is the simplest type of control. As expected, on-off controllers switch abruptly between two states with no middle state. They are for use with equipment that accepts binary input. An example would be a furnace that is either completely on or completely off.

On-off controllers only switch output when the setpoint has been crossed. In the case of heating control, the controller switches on when below the setpoint and off when above the setpoint. To prevent rapid cycling of the system, which can cause damage, hysteresis (also called on-off differential) is added to the controller operations. The differential prevents cycling by exceeding the setpoint by a small amount before the controller switches on or off.

On-off controllers often are used in applications that do not require precise control. They also are used in systems that cannot handle having the energy turned on and off frequently, or those where the mass of the system is so great that temperatures change extremely slowly. On-off controllers can be used for temperature alarms.


Thermocouples are used in many applications, including petroleum, steel, automobile, aerospace, telecom, chemicals, electronics, food processing, consumer goods and lumber manufacturing.

PID: Responsive Control

PID control uses three different control terms — proportional (P), integral (I) and derivative (D) — to help the controller’s algorithms provide a more accurate response to deviations from the setpoint.

When a controller receives input that a process variable has varied from the setpoint, instructions are sent to the final control element for correction. For example, a controller receives a signal from a thermocouple that a process temperature is too low. This prompts the controller to turn on a heater to bring it back up to temperature.

Simple on-off control often leads the final control element to overshoot the setpoint —especially when the original deviation was small. Repeatedly overshooting the setpoint causes the output to oscillate around the setpoint in either a constant, growing or decaying sinusoid. The system is unstable if the amplitude of the oscillations continuously increases with time.

PID controllers use the algorithm derived from their three control terms to maintain system stability by limiting overshoot and the resulting oscillation. The proportional variable controls the rate of correction so that it is proportional to the error. The integral and derivative variables are time based and help the controller automatically compensate for changes in the system. The derivative variable considers the rate at which the error is increasing or decreasing while the integral variable uses knowledge of accumulated errors to the length of time the process is not at the setpoint. This information is used to correct the proportional value.

PID controllers generally are considered the most efficient type of controller. They are widely used in industrial settings. Though each of the variables must be tuned to a particular system, PID controllers provide accurate and stable control.

In order to make PID controllers even more responsive to real-world situations, many manufacturers have incorporated fuzzy logic (or fuzzy control) into the instruments. Fuzzy logic is a mathematical system that attempts to emulate human reasoning. Rather than the binary logic of standard controllers, fuzzy logic introduces continuous variables that provide an effective means of capturing the approximate, inexact nature of the real world.

This ability enables controllers with fuzzy logic to make quick, subtle changes that significantly improve response to fast-changing variables independent of the programming done by the operator. For example, as heaters, valves and other final control elements age, they show signs of wear and no longer respond in the same way as they did when new. Fuzzy logic recognizes this and automatically compensates.

Now that you have a basic understanding of how thermocouples work and how they are used within a process loop, it is easy to see why they are used in so many applications.