Active Tests of Thermocouple

Manufacturers routinely apply quality control to their thermocouple products, yet users are responsible locally to assure that no occasional defect, error, or change has occurred in manufacture or installation. Several simple and routine test procedures are available to the user.

Symptom and Occurrence of Inhomogeneity

Thermoelectric inhomogeneity, though now widely misunderstood, has long been recognized as a very real and commonplace problem of thermocouples applied in damaging circumstances.

Modern production of thermoelements now results in only slight, but measurable, irregular longitudinal variation of Seebeck coefficient, ∆(T, x), distributed along the full length of a thermoelement. Maximum deviations of ∆(T) typically are a very small fraction of the standard tolerance. The risk of significant inhomogeneity in as-received new wire material is slight.

Inhomogeneity that is significant in thermocouple thermometry occurs as a broad localized distinctive deviation from the initial uniform Seebeck coefficient. Recognizable drift during measurement of a temperature expected to be constant is always a symptom of progressing inhomogeneity. Inhomogeneity rarely is authentically tested in commercial practice or by calibration laboratories.

Profound inhomogeneity very often results from prolonged use at high temperature, in a harsh environment, or during fabrication. Most often, significant inhomogeneity is caused or aggravated by excessive temperature so it usually occurs broadly distributed adjacent to the measuring junction precisely where it has the greatest adverse effect in thermometry when junction temperature varies.

Less often, it results from defective insulation, metallurgical interaction between thermoelements or with their insulant, evaporation of alloy constituents, or from exposure to higher temperature regions or chemical attack remote from the measuring junction.

Conventional NIST-traceable precise and costly calibrations might sincerely but incorrectly be certified even if a degraded thermocouple can accurately be measured to be locally severely inhomogeneous and of indefinite uncertainty as actually applied in thermometry. Most troublesome inhomogeneity is unrecognized even by experienced thermocouple users.

Therefore, inhomogeneity remains a commonplace, very real, but usually discounted problem. Display of the distribution of such inhomogeneity requires a special test.

Authentic Inhomogeneity Test Methods

A consequentially inhomogeneous segment has, for a test temperature step, ∆T at T_a, a deviant Seebeck emf, ∆E(x), which is significant relative to initial tolerance. The objective of inhomogeneity testing is to identify locations, longitudinal distributions, and the relative amount of deviations from the initial Seebeck coefficient along thermoelements.

The inhomogeneity test is used only to disqualify for calibration or thermometry. It is not a recalibration. Inhomogeneity identified by scanning a 100°C step above ambient temperature is adequate to disqualify a thermocouple for use or recalibration.

While the relative magnitude of excessive inhomogeneity at moderate temperature might be greater or less at higher temperature, inhomogeneity significant at a temperature of use is also apparent at moderate temperature. It is not necessary to test inhomogeneity at the temperature of intended use.

Several authentic thermoelectric inhomogeneity tests, simple to sophisticated, have been developed and extensively demonstrated. Effective primitive methods were practiced by pioneers.

Valid testsrange from methods of many-cm spatial resolution to advanced methods that have accurately resolved inhomogeneity with spatial resolution of a few millimeters. Their practical need has been very clearly proved, yet few authentic inhomogeneity tests are now performed.

Consequential inhomogeneity is spatially distributed; therefore, all authentic inhomogeneity tests of thermocouples require moving a very narrow temperature step between two temperature plateaus scanned progressively along the paired thermoelements while spatially monitoring the voltage.

This appropriate step temperature pattern is best applied by progressive immersion of a thermocouple from ambient temperature into an isothermal-heated deep liquid bath of high thermal conductivity while recording the voltage as a function of position. This step compares the Seebeck emfs of widely separated segments of nearly equal length.

The Seebeck coefficients of the two segments are likely most dissimilar if the segments are well separated. The finest resolution is by stepped-progressive immersion of the thermocouple into a deep isothermal liquid bath of high thermal conductivity, such as of a eutectic alloy of gallium, indium, and tin. That alloy is liquid at room temperature. Heated to near 100°C, the non damaging bath imposes a narrow step of temperature above ambient temperature at the liquid interface.

The bath test is well adapted to MIMS thermocouples. It is applied to other thermocouple forms by enclosing them temporarily in a thin close-fitting insulating plastic sheath. The spatial resolution depends on immersion speed, step duration, and thermal properties of the liquid. Even water-based liquids below their boiling point can be used but with lower spatial resolution. Such advanced tests well demonstrated by several laboratories have used specialized equipment so they have not been accepted by the thermometry community.

However, the Australian National Measurement Laboratory has routinely preceded calibration by one authentic inhomogeneity test. The test uses progressive entry of the thermocouple into a lengthy isothermal oven while monitoring the emf. That produces an authentic inhomogeneity test with spatial resolution of several cm. The vertical immersion tests restrict the length of the test subject.

A similar authentic test is performed on individual thermoelements of any length. A thermoelement is run through a heated liquid bath or furnace with widely separated entry and exit points. This imposes two spaced temperature steps of equal width and temperature different from ambient temperature. The effect of such progressive immersions is well visualized by the T/X sketch of the Functional Model.

False Inhomogeneity Tests 

Comparison Test: Some procedures are misunderstood to be inhomogeneity tests. For example, two or more samples taken from well-separated locations along a batch of thermocouple wire are calibrated. Calibrations are of indefinite local segments of thermoelements.

Agreement between two calibrations only suggests that homogeneity is within thermocouple tolerance, but it effectively misunderstands troublesome inhomogeneity to be global rather than localized. It is not an authentic inhomogeneity test. 

Spike Test: Another commonplace quality procedure also is misunderstood to be an inhomogeneity test. This “spike test” is the very opposite of a true inhomogeneity test. The thermocouple pair, of any length, must be electrically joined at the end opposite the thermocouple monitor.

That junction is held at any fixed temperature. While monitoring the temperature indication, a symmetric very narrow swept spike of temperature is applied by a narrow-tipped soldering iron, flame, or a transverse heated wire that is slowly scanned along the thermocouple.

This commonplace practice inappropriately compares the difference between Seebeck coefficients of two very closely spaced segments that are most likely to be similar even in severely inhomogeneous thermoelements.

The narrower the temperature spike, the worse the test as an indicator of consequential inhomogeneity. Its widespread misunderstanding as an inhomogeneity test has misled many to discount inhomogeneity as a practical significant problem in accurate thermoelectric thermometry.

Hidden Junction Test

The temperature spike test does serve a very different practical purpose. As the very narrow temperature source scans past a location, a localized spike of temperature indication reveals and locates a real junction.

This quick and simple procedure actually is an important test for hidden real junctions. Such defects do infrequently occur where different thermoelements, spliced or connected hidden within a cable assembly, are different though of the same nominal type or even of a different type.

Otherwise, concealed, they might be connected in crossed polarity. Such consequential defects, unless discovered by this important test, are very difficult to recognize during setup and thermometry.

Thermocouple Instability Test

The most obvious occurrence of thermocouple inhomogeneity is that which, unrecognized, is produced in the traditional thermocouple test of long-term drift of sensitivity. The measuring junction, and adjacent lengths of thermoelements, is fixed deeply immersed (more than 20 probe diameters), within an essentially isothermal zone of a furnace. Oven temperature and environment remain constant.

Thermoelectric instability is measured as a sustained progressive change of thermocouple voltage that, with exposure time, drifts toward the tolerance limit and beyond. Costly long-duration thermocouple instability tests, conducted by suppliers rather than by users, are usually performed over durations from weeks to months.

The traditional instability test result is misperceived as if a test of overall rather than localized thermocouple sensitivity. Peculiarly, the emf drift observed is not from those segments that are uniformly degrading within the isothermal furnace.

Actually, as understood from the Functional Model and T/X sketch, the material degrading isothermally within the furnace contributes little of the observed voltage. Rather, the observed drift of emf is of thermoelement segments at the oven entry that span the temperature difference between furnace and ambient temperature remote from the measuring junction.

Therefore, the uniform change of Seebeck emf of the segments within the furnace actually is much greater than is reported. Likewise, tolerance is exceeded earlier than observed in this test.

Inhomogeneity Test Incidental to Instability Test

A simple in-place inhomogeneity test, complementing the instability test, can reveal the actual inhomogeneity pattern and improves the significance of the drift test at little additional cost.

Beginning the instability test, voltages E₁(x) are recorded as the measuring junction, and thermoelements are progressively advanced from ambient temperature to dwell at the drift test position. That data document, for new material, the initial temperature profile of the entry temperature step and of the oven.

Eventually, at the end of the drift test, the exit profile, E₂(x), is recorded as the measuring junction is first advanced a distance greater than the width of the entry step into the furnace. That documents the final profile of the entry step with undamaged material.

Then, the thermocouple is slowly withdrawn until the measuring junction dwells at ambient temperature. The difference between the two profiles, E₁(x) and E₂(x), from each test subject documents the actual test induced pattern of inhomogeneity and the relative magnitude of the actual change of the degraded material.

Unlike the formal inhomogeneity test conducted near room temperature, this complementary test is performed at the elevated test temperature that produced the inhomogeneity that appeared as drift.

A special automated calibration furnace, Model FUR-1200, offered by the Electronic Development Labs, Inc. is well adapted to such complementary programmed in-place inhomogeneity test as it allows programmed computer-controlled immersion and withdrawal of the test subject.

It also is used to document the effect on stability of cyclic rather than fixed temperature. MIMS assemblies are degraded differently by varying rather than constant temperatures.

Thermoelement Identification

Color codes apply to extension-grade wires, to insulated thermocouple-grade wire, and in principle to bare thermocouple wire. Bare wires are not indelibly color-coded. An installer very skilled in electronics might not be familiar with the distinctive thermocouple codes or their consequence.

Thermocouples of the same letter-designated type but of different international standards have different international color codes for both type and polarity. The prudent user, before installation or application, will confirm material-type identification of each thermoelement independent of the color code.

Definite type identification must be by a combination of methods. A single identification method may be indefinite. No method fits all materials or circumstances.

Visual Identification: Type TP thermoelements are of copper. Bare, they are visually distinguished by their distinctive reddish color. JP thermoelements are of iron and have a distinctive matte gray cast. Other base-metal alloys and platinum and its alloys, if bare, are not easily distinguished by appearance.

Usually they all have a similar bright silvery appearance unless, if heat treatment or if bared from compressed mineral insulation, fabrication has produced a roughened gray matte surface appearance.

Magnetic Identification: A type JP (iron) leg is strongly magnetic. Type KN is only weakly magnetic. All other standard thermoelement materials are nonmagnetic. JP and KN thermoelements can be distinguished from the others by testing the relative attraction to them of a small magnet.

Resistive Identification: Thermocouple assemblies or cables usually have paired thermoelements that are of the same length and wire size. If the temperatures of the ends are about the same, the resistance of each thermoelement and the ratio between leg resistances are diagnostic for identification of the materials. Room temperature resistivities of thermoelements are given in the Table. To view that table,Please Follow the Link.

If the identity of one thermocouple or a thermoelement of a batch of like material, wire size, and length is determined, the others can be confirmed by comparison of loop resistances.

Resistance measurement of isolated individual thermoelements requires that both ends be accessible at the same location and temperature. For assembled thermocouples with inaccessible measuring junctions, only the isothermal loop resistance can be measured and compared with calculated loop resistances.

This quantitative loop resistance measurement is complicated by Seebeck emf that can affect apparent resistance if the circuit ends are at a substantially different temperature.

Thermal Identification: Less conveniently, a pair can be identified by the terminal voltage or indicated temperature for a known temperature of measuring junction and the reference junction. Definite identification does not require the accuracy of formal thermometry.

Reference junction temperature may be applied by simple ice bath or compensation by the monitor. An independently known temperature, at least 200 °C, applied to the measuring junction for identification is required.

Otherwise, uncertainty of the thermocouple calibration and of the imposed temperature would make the emf distinction of thermocouple pairs unreliable, such as E from J and K from N or T by using a 100 °C boiling water bath.

Very similar types R and S (that have identical color codes) can be reliably distinguished only at much higher temperatures or by formal calibration.

• Thermocouple Working Principle
• Absolute Seebeck Effect
• Basic Thermocouple Circuits
• Extensions of Thermocouple
• Functional Model of Thermoelectric Circuits
• T/X Sketch of Dual-Reference Junction Circuit
• Applications of Functional Model
• Characteristics of Thermocouples
• Thermocouple Hardware
• Thermocouple Junction Styles
• Active Tests of Thermocouple
• Calibration of Thermocouples
• Thermocouple Thermometry Practice
• Distinctive Thermocouple Noise Problems