Thermocouple Types Explained: J, K, T, E, N, R, S and B

Thermocouple Types Explained: J, K, T, E, N, R, S and B is a quick reference for picking the right sensor for a given temperature range, atmosphere and accuracy target. Thermocouples are the most common contact temperature sensor in industrial plant: cheap, rugged, self-powered, spanning cryogenic lines to furnace interiors above 1700°C. Choosing the wrong type, or ignoring cold-junction compensation, is a routine source of bad readings blamed on "sensor drift" when the real fault is specification.

The Seebeck Effect: How a Thermocouple Works

A thermocouple is two dissimilar metal wires joined at the "hot" junction. A temperature difference between that junction and the open reference ends generates a small voltage, typically tens of microvolts per degree Celsius. This is the Seebeck effect, discovered in 1821: the measured EMF is the difference between the two metals' Seebeck coefficients. A thermocouple does not measure temperature directly; it measures a voltage that is a tabulated function of junction temperature difference, shaped by alloy pairing and sheath material.

The Letter Types: Alloys, Ranges and Where Each Belongs

Letter designations (J, K, T, E, N, R, S, B) are standardised alloy pairs with published voltage tables, so a Type K thermocouple from any manufacturer behaves the same. Base-metal types (J, K, T, E, N) dominate industrial use; noble-metal types (R, S, B) suit high-temperature, high-accuracy work.

• Type K (Chromel-Alumel): industrial default, wide range, good oxidising-atmosphere resistance. Avoid sulphurous atmospheres, which cause "green rot" corrosion.
• Type J (Iron-Constantan): higher EMF than K at moderate temperatures, but the iron leg rusts in humid conditions.
• Type T (Copper-Constantan): excellent stability at low and sub-zero temperatures; standard for refrigeration and cryogenics.
• Type E (Chromel-Constantan): highest EMF output per degree of any standard type; usable down to cryogenic ranges.
• Type N (Nicrosil-Nisil): resists the drift affecting Type K in the 300 to 500°C ageing range; a good upgrade where K is unstable.
• Type R and S (Platinum-Rhodium/Platinum): stable to high temperatures, used in glass and ceramics; fragile and sensitive to contamination.
• Type B (Platinum-Rhodium/Platinum-Rhodium): highest standard temperature range, but reads unreliably below roughly 50°C output, unsuitable near ambient.

Cold-Junction Compensation

Thermocouple voltage depends on the difference between the hot and reference junction temperatures, not the hot junction's absolute value. Historically this was solved with an ice bath at 0°C. Modern transmitters instead measure terminal block temperature, usually with a thermistor or RTD, and add a correction voltage electronically, a technique called cold-junction compensation (CJC). A terminal block near a hot cabinet or in direct sun drifts that reference, and every downstream reading is wrong by the same amount.

IEC 60584 Tolerance Classes

IEC 60584-1 defines the tolerance classes referenced in the table above. Class 1 is the tightest standard tolerance, generally achievable only across a limited span; Class 2 is wider and covers the full rated range for most types; Class 3 exists for a smaller set of types at low temperatures. Specifying a class is not optional: it determines whether an installation can resolve the process variation it is meant to control. A furnace trip set at ±5°C is meaningless if the sensor's tolerance band is ±2.5°C or more.

Thermocouple vs RTD: When to Choose Which

Thermocouples win on range (cryogenic through 1700°C plus), ruggedness and fast response. They lose on absolute accuracy and stability compared with a well-specified RTD Pt100, and the millivolt signal is more noise-prone over long runs, so installations often pair it with a transmitter converting to a 4 to 20 mA current loop. Choose an RTD below roughly 500°C where accuracy matters most; choose a thermocouple where range exceeds RTD limits or response time is critical. On rotating machinery, bearing and winding inputs often feed into API 670 machinery protection logic, where sensor type and redundancy are part of the protection philosophy.

Installation and Data Practices That Prevent Bad Readings

Three practices account for most field errors: matched extension-grade wire back to the terminal block, keeping thermocouple cable separated from VFD or high-current power cable, and verifying insertion depth so the tip sits in the process stream, not a boundary layer near the wall. Logging sensor type, tolerance class and recalibration history against the asset in a CMMS such as Fabrico turns sensors into a traceable asset: the team can see whether replacement is due and what tolerance was assumed at the alarm setpoint. Teams evaluating their records can book a Fabrico demo to see how calibration history integrates with work order planning.

Frequently Asked Questions

Which thermocouple type is most accurate?

Type T offers the best accuracy at low to moderate temperatures. For high-temperature accuracy, Type S has historically served as a laboratory reference standard, with Type R comparable in industrial service.

Can I extend a thermocouple with ordinary copper wire?

No. Copper wire introduces additional junctions with uncompensated Seebeck voltages. Extension or compensating cable matched to the thermocouple type must run from the sensor to the cold-junction reference point.

Why does my Type K reading drift after months in service?

Type K is prone to short-range ordering drift in the 300 to 500°C band and to green rot in reducing or sulphurous atmospheres. Type N resists both and is a reasonable upgrade if drift recurs.

Do thermocouples need periodic calibration?

Yes. Unlike RTDs, thermocouples can drift from alloy contamination and grain changes at temperature, not just electrical faults, so periodic verification against a reference standard is good practice for critical loops.