Automation Glossary • Thermocouple

What Is a Thermocouple?

Merobix Engineering • • 7 min read

The thermocouple is the temperature sensor for the hot, rugged end of the process - flares, furnaces, fired heaters, and exhaust streams where an RTD would not survive. It is also one of the oldest and most elegant sensing principles in instrumentation. This guide explains what a thermocouple is, how the Seebeck effect makes it work, the common types, and why cold-junction compensation matters.

Back to Blog

Thermocouple in one line: A thermocouple is a temperature sensor made of two different metal wires joined at one end. The junction generates a small voltage that varies with temperature (the Seebeck effect), and reading that voltage - after cold-junction compensation - gives the temperature.

How a Thermocouple Works

When two dissimilar metals are joined and that junction is at a different temperature from the point where the wires are read, a small voltage appears - typically a few millivolts. This is the Seebeck effect. The voltage depends on the temperature difference between the sensing (hot) junction out in the process and the reference (cold) junction back at the instrument. Measuring that voltage and knowing the metals used lets the electronics calculate the process temperature.

Because the output depends on a temperature difference, the instrument must know the temperature of its own reference junction to compute the absolute process temperature - this correction is called cold-junction compensation, and every thermocouple transmitter or input card performs it automatically.

Thermocouple Types

Different metal pairs are standardized into lettered types, each with a temperature range and characteristic. Type K (nickel-chromium / nickel-alumel) is the most common general-purpose type, covering a very wide range and used heavily on heaters and flares. Type J (iron / constantan) suits moderate ranges. Type T (copper / constantan) is accurate at low temperatures. Types E, N, and the noble-metal types R, S, and B extend to specialized or very high-temperature service.

Thermocouples connect using matching extension wire - cabling made of the same or compensating alloys - so the reference junction is correctly located at the instrument. Using ordinary copper wire instead would introduce an uncompensated junction and a measurement error.

Thermocouple vs RTD in Oil and Gas

Thermocouples handle much higher temperatures than RTDs, respond faster, tolerate vibration and shock, and cost less - which is why they own the flare, furnace, and combustion measurements in a facility. RTDs are more accurate and stable at moderate temperatures and dominate process and custody points. In the SCADA chain, a thermocouple wires into a temperature transmitter or a controller's thermocouple input, which converts the millivolt signal into a tag that SCADA reads and trends.

Specifying a Thermocouple: More Than Picking a Letter

The type letter fixes the wire alloys, but the specification you hand a supplier carries several more decisions. Junction construction comes first: a grounded junction is welded to the sheath tip for faster response but couples the measurement to plant ground, an ungrounded junction is isolated and immune to ground loops at some cost in speed, and an exposed junction is fastest and most fragile. The tradeoffs are laid out under thermocouple junction types. Sheath material and diameter follow from the process atmosphere and the response you need, and most process installations put the assembly in a thermowell so the element can be pulled without breaking containment.

Accuracy is specified by tolerance class rather than by numbers pulled from the air: the IEC 60584 tolerance classes define the permitted deviation for each type, and the class you order should follow from what the measurement is for - control, protection, or plain monitoring. For anything feeding a trip, the redundancy and voting arrangement usually matters more than buying a tighter class.

Installation Details That Decide the Measurement

Immersion depth is the quiet error source. Too shallow, and heat conducted along the sheath pulls the junction toward ambient, so a hot process reads low. Immerse deep enough that stem conduction is negligible for the geometry - follow the assembly manufacturer's recommendation rather than a rule of thumb - and spring-load the element against the bottom of the thermowell so the junction stays coupled to the process, not to the pipe wall.

Wiring errors are the other half. Extension wire must match the thermocouple type, polarity must be preserved at every terminal, and runs should be shielded and routed away from power conductors - a millivolt-level signal shares no cable tray comfortably with a drive output. Every added terminal is a potential unplanned junction, so keep splices few and use the correct compensating connectors. A quick pass through verifying polarity and extension wire catches most of these faults before anyone starts doubting the sensor itself.

Failure Modes and What They Look Like on the Trend

Thermocouples fail open more often than shorted: thermal cycling and vibration eventually crack the wire, and the reading is then driven by the input card's burnout detection - upscale or downscale per configuration - so the failure is obvious provided burnout action was configured deliberately rather than left at whatever default shipped. Subtler failures matter more. Moisture inside the sheath shunts the millivolt signal and drags the reading toward the temperature at the wet spot, so the trend drifts credibly instead of failing loudly.

Drift also comes from the wire itself. Prolonged exposure at temperature changes the alloys, and because the Seebeck voltage is generated along the temperature gradient rather than at the tip alone, an inhomogeneous section of wire produces an error that appears and disappears as the gradient moves along the sheath. That is why a thermocouple can pass a bench check and still misread installed. When a reading is doubted, compare against a reference measurement at the same point and check the cold-junction reading before condemning the element.

A Five-Minute Field Sanity Check

  1. Confirm the transmitter or input card is configured for the correct thermocouple type.
  2. Compare the cold-junction temperature the instrument reports against the actual terminal temperature.
  3. Verify polarity and extension wire type at every accessible terminal.
  4. Disconnect and measure the loop's millivolt output; compare it against the reference table for the expected process temperature.
  5. Inspect the sheath and thermowell for damage, moisture, or shallow immersion.

Five checks, no special tools beyond a multimeter and a reference table, and they resolve the large majority of suspect-thermocouple callouts without pulling the element. Anything still unexplained after this list is a genuine candidate for replacement or for a closer look at the input electronics.

Frequently Asked Questions

How does a thermocouple work?

Two dissimilar metals joined at a junction produce a small voltage that changes with the temperature difference between that junction and the reference point (the Seebeck effect). Reading that voltage, with cold-junction compensation applied, yields the temperature.

What is the most common thermocouple type?

Type K (nickel-chromium / nickel-alumel) is the most widely used. It covers a broad temperature range, is inexpensive and rugged, and suits general industrial service including heaters and flares in oil and gas.

Thermocouple or RTD - which should I use?

Use a thermocouple for very high temperatures, fast response, ruggedness, and low cost - typical of flares and furnaces. Use an RTD when accuracy and long-term stability matter at moderate temperatures, such as process and custody measurement.

Why does my thermocouple read in the wrong direction?

Reversed polarity somewhere in the run: the process heats up and the indicated value falls. A full reversal at the instrument is obvious, but a reversal at an intermediate terminal block partially cancels and produces a subtler, temperature-dependent error, which is why polarity is verified end to end rather than only at the transmitter.

Should a suspect thermocouple be recalibrated or replaced?

Field practice usually replaces the element. The wire itself drifts and cannot be adjusted, and elements are inexpensive compared with the labor of characterizing one. Calibration effort goes into the transmitter and input card, and into verifying the installed measurement against a reference. Anything feeding custody or safety functions follows the site's QA program.

More in Instrumentation & Measurement
Calibrate a Thermocouple  •  Commission a Thermocouple Loop  •  Verify Thermocouple Polarity and Extension Wire  •  IEC 60584 thermocouple tolerance classes  •  Thermocouple vs RTD Selection  •  All Instrumentation & Measurement →
Free SCADA operator training
Merobix University - 70 video lessons & 261 quiz questions, from first login to compliance reporting. No demo call required.
Start free →