What are the types of thermocouple wires?

Thermocouple wire is sold by a letter. K, J, T, E, N, R, S and B are the eight letter designations in common industrial use, and each one is a published pairing of two different metals with its own voltage table, its own temperature range and its own permitted error. The letter is the only thing the instrument needs in order to turn the millivolts into a temperature, which is why it is the first line on any specification.

What the Letter Fixes, and What It Leaves Open

The letter answers four questions at once: which two alloys are inside, which temperature range the pairing was characterised for, how much error is allowed when the wire is new, and what colour the insulation should be. Those four are the reason the letters exist, because a buyer in one country and an instrument in another can agree on a reading without exchanging tables.

What the letter does not answer is where that wire may be used. The nominal range on a datasheet is the span over which the pairing behaves, and it is not the limit of the installation. The conductor gauge, the insulation, the sheath and the atmosphere all take temperature off the top, and those belong to the buyer, since the standard has nothing to say about them.

The Eight Letter Types

Four of the eight are base metal pairs, built on nickel, iron and copper, and four are noble metal pairs built on platinum and rhodium. The two groups are priced far apart and behave differently, and the split is the first thing to know.

Three facts about that table are worth isolating, because each one changes a decision. Type E has the highest output of any standard type, so a small temperature change produces a larger signal and the instrument has an easier job; that is why it is found on test benches more than on general process work. Type B has the lowest output of the eight, and its published data notes that it reads the same at 0 °C and 42 °C, which is why it cannot be used at the cool end at all. And Type K covers the widest span of the base metal types, which is most of the reason it is the default.

The tungsten-rhenium pairs are not in this table on purpose. Type C and Type D use W5Re/W26Re and W3Re/W25Re respectively, and they are not letter designations in the same sense: they are covered by ASTM E1751, the guide for non-letter designated combinations, and they sit outside the letter-designated tables. A spec that calls for a tungsten pair is a different conversation from one that calls for a letter.

The Range on the Datasheet Is Not the Range in Service

A nominal range is quoted for the alloy pair. The published upper limits below are for protected thermocouples, in a closed-end sheath, and they differ by conductor gauge. The same type that is rated to 1,250 °C on paper is rated to 760 °C in the finest gauge listed.

A second range matters as well, and it is the one that causes arguments. The tolerance tables do not cover the full nominal span. Under ASTM E230 the published limits apply to Type J from 0 to 760 °C, Types K and N from 0 to 1,260 °C, Type T from 0 to 370 °C, Type E from 0 to 870 °C, Types R and S from 0 to 1,480 °C, and Type B from 870 to 1,700 °C. Outside those spans the wire is not being bought against a published limit, whatever the nominal range says, and the percentage figures apply to readings above zero.

Tolerance: Two Systems, Two Sets of Numbers

Europe and North America describe the same wire in different languages, and the two are not interchangeable on a purchase order. North America uses grades, called standard limits and special limits, and quotes each as the greater of a fixed number of degrees or a percentage of the reading. IEC 60584 sorts wire into classes, numbered 1, 2 and 3, and quotes each against a temperature range.

The phrase the greater of matters more than it looks. A standard-grade Type K reading 500 °C may be out by ±3.75 °C, because 0.75 % of 500 is larger than the fixed 2.2 °C; special-grade wire at the same point may be out by 2.0 °C. Below about 293 °C the fixed figure takes over and governs instead. So the error a buyer is entitled to is not a constant, and a claim about accuracy that does not name the temperature is not a claim at all.

Two further traps sit in the comparison. Type J is covered to 750 °C under IEC and 760 °C under ASTM, so a datasheet written to one standard and read against the other will look wrong by 10 °C. And the IEC classes are ranges, so the same Class 2 wire carries two different numbers depending on where in its span it is used. When wire is quoted, the standard has to be named next to the class or the grade, or the number means nothing.

The Colour Code, and the Mistake It Causes

Insulation colour is meant to identify the type without reading the printing, and the two standards behind it disagree. Under ANSI MC96.1 and ASTM E230 the negative wire is always red and the positive wire carries the type colour. Under IEC 60584 the negative wire is white instead, which changes every pairing.

A yellow-and-red Type K cable in a North American panel and a green-and-white one on imported equipment are the same thermocouple in two dialects. Two mistakes follow from that, and both are made at the terminals, not at the drawing board. The first is polarity: reversing the legs of a pair inverts the reading, and the result looks like a calibration problem. The second is extension cable. Every junction between unlike metals adds its own voltage, so ordinary copper cable run between the sensor and the instrument introduces junctions that were never in the calculation. Extension cable has to match the type letter the whole way, and the standard has to be named when the cable is ordered.

What Each Type Fails At

No type fails at random. Each has a known weakness, and most of them show up as a change in the reading, with nothing visibly broken.

Choosing a Type, in Five Checks

The order of the checks matters, because the first two rule types out and the rest only choose between what is left.

1. Temperature range. Pick a type whose published range leaves headroom above the peak, then reduce it for the gauge, the insulation and the sheath. The nominal figure is the starting point, not the answer.
2. Atmosphere. Oxidising conditions suit K, N, R and S. Vacuum, inert and reducing conditions suit J. Moisture favours T and rules out J. Hydrogen and reducing gases rule out the platinum types.
3. Accuracy. Decide whether standard limits are enough before pricing the special grade, and remember that the result is only as good as the reference junction handling and the instrument behind it.
4. Instrument and cable. The controller, transmitter or logger has to be set to the exact letter, and the extension cable has to match it to the terminals.
5. Colour standard. Confirm whether the site reads ANSI or IEC colours before the wire is cut.

Those five checks produce a short specification, and the specification is what the order should carry.

Questions Buyers Ask

Which thermocouple type is the most common?

Type K, by a wide margin. It covers the widest span of the base metal types, it is the cheapest of the four general-purpose pairs, and almost every controller and transmitter supports it out of the box. What rules it out is the atmosphere: it is the wrong choice in vacuum, or in a furnace that alternates between oxidising and reducing conditions.

Which type is the most accurate?

Types R and S, on paper. Their special limits are ±0.6 °C or 0.1 % of the reading, the tightest of the eight, and Type S is often used as a reference. Among the base metal types Type T is the tightest at ±0.5 °C or 0.4 %, which is part of why it is used for cold-chain and laboratory work. Paper accuracy is not field accuracy: the reference junction, the extension cable and the instrument all add error.

What does the colour of thermocouple wire tell me?

If you know which standard it was coloured to. Under ANSI the negative leg is red and the positive leg takes the type colour, so a Type K pair is yellow and red. Under IEC the negative leg is white, and the same Type K pair is green and white. A pair that looks unfamiliar is more likely to be the other standard than the wrong sensor.

What is the difference between standard limits and special limits?

The permitted error, and about half of it. For Type K the standard limit is ±2.2 °C or 0.75 % and the special limit is ±1.1 °C or 0.4 %, whichever of the two figures is greater at the temperature in question. IEC describes the same wire with class numbers instead of grades, and the numbers differ slightly, so the standard has to be named on the order.

Can I substitute Type N for Type K?

Not without changing the rest of the loop. Type N was developed as a more stable alternative to K under sustained heat, and it carries the same published limits, but it is a different alloy pair with a different voltage table. The transmitter, controller or logger has to be set to N, and the extension cable has to be Type N as well, or the reading is taken off the wrong table.

Why does the same wire have two different temperature limits?

Because the nominal range describes the alloy pair and the service limit describes the installation. A Type K pair is characterised from −200 to 1,250 °C, but the published upper limit for a protected thermocouple falls from 980 °C in 20 AWG to 760 °C in 30 AWG, and the ASTM tolerance tables cover a narrower span than the nominal range in every case. The limit that applies is the lowest of the three.

Summary

The eight letter types are a shorthand for four things at once: the alloy pair, the range the pairing was characterised for, the permitted error, and the insulation colour. Type K is the general-purpose default, T and E cover the cold end and the high-resolution end, N is the more stable base metal type above 1,000 °C, and R, S and B take the extremes at a platinum price. What the letter does not settle is the limit in service, which the gauge, the construction and the atmosphere decide.

DLX Alloy has been making nickel alloys at Changzhou, Jiangsu since 2002, on a 12,000 m² site that holds ISO 9001 certification and is audited by SGS. Thermocouple wire is supplied in the eight letter types above, as bare wire and as mineral insulated cable: send the type, the grade or class, the range and the construction for a confirmed quotation.

References

1. ASTM E230/E230M, temperature-electromotive force tables for standardized thermocouples. Source for the limits of error, for the spans over which those limits apply, and for the colour code in its ANSI form.
2. IEC 60584-1, thermocouples: EMF specifications and tolerances. Source for the class system and for the tolerance table.
3. IEC 60584-3, thermocouples: extension and compensating cables, tolerances and identification system. Source for the IEC colour code and the white negative leg.
4. ANSI MC96.1, temperature measurement thermocouples. Source for the ANSI colour assignments, in which the negative leg is red for every type.
5. ASTM E1751/E1751M, guide for temperature-EMF tables for non-letter designated thermocouple combinations. Source for the position of the tungsten-rhenium pairs, Type C and Type D, outside the letter designations.
6. TE Connectivity comparison of thermocouple types. Source for the trade names, the application ranges, the design specification tables and the recommended upper limits by wire gauge.
7. Published thermocouple type chart and tolerance comparison, Tripoint Technical Services. Source for the limits of error by grade, the worked example of the greater-of rule, the ANSI against IEC colour comparison and the failure modes by type.
8. Product data held by DLX Alloy. Source for the forms supplied.

The figures quoted are published values for new wire and are not acceptance limits. A thermocouple lot is accepted against the ordered type, grade or class, range and construction, and the tolerance figures describe new wire only: they say nothing about drift after the sensor has been in service, which is a calibration question.