Distinctive Thermocouple Noise Problems

Thermocouple emfs are small. Fine temperature resolution requires substantial amplification. For quasi-static thermometry, modern noise suppression techniques available to designers of data loggers overcome ordinary coupled electrical noise to routinely achieve as small as 0.1 °C temperature resolution and long-term stability (much smaller than required by thermocouple tolerance).

Thermocouple measurement of transient temperatures requires more careful attention to classical noise reduction techniques because thermoelectric circuits present distinctive problems.

Distinctive Thermocouple Noise Problems

Electromagnetic Noise

In thermocouple probes of MIMS construction, ceramic-bead-insulated thermocouples, and some paired insulated thermocouple wires, the paired thermoelements are not twisted and are spaced uniformly apart in a plane.

Thermocouples are unbalanced lines because the paired legs have very different resistance. Lengthy and larger diameter thermocouple probes present a significant circuit loop area to couple magnetic noise fields. Even though MIMS probes are entirely metal-enclosed, the sheaths are very thin and of low magnetic permeability. They are scarcely effective for EM shielding.

These features prevent the use of some classic techniques for the rejection of EM noise. To reduce troublesome EM noise, the measuring thermocouple probe should be of the smallest feasible diameter and of minimum length. In instances where the EM source is localized and identified, the orientation of the probe relative to the source might be arranged to minimize EM coupling.

For rejection of EM noise, extension leads should always be of twisted-pair construction, with a pitch small enough to reject high-frequency noise. Some special thermocouple connectors incorporate a ferrite core intended to suppress transient noise.

Electrostatic Noise

ES noise is more easily reduced than is EM noise. Shields of low resistance, though thin, can be effective if properly connected. Shields of plated copper braid are commonly used and are effective for noise of moderate frequency.

Continuous shields, such as the MIMS sheaths and aluminized Mylar® film, are effective for low-frequency ES noise and are more effective than braid shields for very-high-frequency noise. ES shields must be continuous (without gaps and holes as in braid) for maximum effectiveness. Shield continuity must be maintained across connectors.

Optimum benefits from shielding require use with thermocouple monitors that provide three-wire input and multiplexers that, individually for each thermocouple, switch both the signal pair and their separate shield lead.

Grounding and Shielding

Dynamic thermocouple thermometry can require monitoring with rapid frequency response in electrically noisy environments. The technically well-founded principles of noise control for electric circuits apply equally to thermoelectric circuits. Well-designed thermocouple instruments carefully isolate the circuitry from other components and control ground reference points.

A secondary overall shield, isolated from inner shields and separately grounded, can improve the rejection of EM noise. The circuit grounding of thermocouple is complicated by intimate thermal contact of the measuring junction with an earth-grounded conducting test subject, as for an intrinsic thermocouple, is needed for rapid transient thermometry.

Ordinary coaxial instrumentation cable must not be used as passive-type extension leads because the outer braid and center conductors—as thermoelements—have very different Seebeck coefficients. If not at uniform temperature, such cable can be a significant source of Seebeck dc noise emf.

Figure 1 illustrates appropriate grounding for several of the many possible situations. Because the measuring junction is not a localized site of source emf, the point of thermocouple grounding must be carefully considered.

Circuit grounding should be at one point only and never at any point of the thermocouple circuit other than at measuring or reference junctions. Electric contact, and particularly shunting or shorting, to a thermoelement at any points at different temperatures between measuring and reference junctions can introduce spurious Seebeck emf from incidental unrecognized thermoelements or ground potential differences.

Feasibility of a grounding method is usually dictated by the internal design of monitors and data loggers. Instrument designers address noise suppression in different ways. Simple and inexpensive line powered thermocouple thermometers often have only single-ended inputs with the negative lead internally connected to the power ground.

This references the sensor grounding also to power ground. In conflict, a “grounded-junction”-style thermocouple might be required for fast transient thermometry. In use, initially isolated thermocouples might short to the protective sheath or test subject. Although this does change the grounding configuration, resulting errors due to thermoelectric effects then usually are predominant.

Filtering

Most common thermometry is of static or low-frequency temperature changes. This allows designers to complement the user-provided passive rejection of EM and ES noise by proper grounding and shielding. Internal filtering provided by manufacturers of data acquisition systems can effectively reject high-frequency noise.

Some systems provide very effective averaging or “double-slope integration” techniques for low-frequency thermometry. Routine resolution and stability to 0.1 μV variation over a period of days are achieved by many modern thermocouple data systems.

However, these noise rejection methods restrict sampling to intervals no shorter than 10–20 ms that are unsuited to very fast transient thermometry.

• 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