PT100 RTD vs Type K Thermocouple vs Thermistor: Choosing the Right Temperature Sensor for Singapore Applications

Sensor comparison at a glance

Before selecting any temperature sensor for a Singapore process, pharmaceutical, or laboratory application, the operating range, accuracy class, linearity, and calibration requirements must all align with the application's demands. This table summarises the key parameters for side-by-side evaluation.

Why sensor selection matters before calibration

The most accurate calibration certificate in the world cannot compensate for choosing the wrong sensor type. A Type K thermocouple placed in a pharmaceutical stability chamber (25°C, ±0.5°C tolerance per ICH Q1A) will drift faster and read less accurately than a PT100 RTD in the same application. Not because the thermocouple is broken, but because it is a fundamentally less appropriate technology for that operating range and accuracy requirement.

This matters for calibration scheduling as well. A PT100 in a temperature-controlled room with stable conditions may remain within tolerance for 12–18 months between calibrations. A Type K thermocouple in the same room, with its higher drift rate and inherent cold-junction dependence, may need calibration every 6 months to maintain the same level of measurement confidence, at significantly higher cost over time.

The correct sequence is: select the sensor for the application requirements first, then calibrate appropriately. Retrofitting a higher-accuracy calibration programme onto the wrong sensor type is inefficient and ultimately cannot solve the problem that better sensor selection would have prevented.

PT100 RTD. How it works and why it is Singapore's process standard

The PT100 is a platinum resistance thermometer: a precision platinum element whose electrical resistance changes predictably with temperature. At 0°C, the resistance is exactly 100Ω (hence "PT100"). As temperature rises, resistance increases at approximately 0.385Ω per °C. A relationship defined by the IEC 60751 standard using the Callendar–Van Dusen equations.

This standardisation means that all PT100 sensors from any manufacturer following IEC 60751 are interchangeable within their tolerance class. Class A accuracy is ±0.15°C at 0°C, ±0.35°C at 100°C, and ±0.80°C at 300°C. Class B is approximately twice those tolerances. A PT100 Class A in a pharmaceutical stability chamber is a known, auditable, specification-backed choice. Unlike a non-standardised sensor whose performance depends entirely on the manufacturer's individual calibration.

For Singapore applications, the PT100 is the dominant choice across pharmaceutical GMP, food cold-chain, HVAC chilled water, and process manufacturing. The reasons are practical: near-linear response simplifies instrument scaling and fault-finding; stability over time reduces calibration frequency and cost; absence of a cold junction reference eliminates a common source of systematic error; and the wide ecosystem of transmitters, data loggers, and indicators that accept IEC 60751 inputs means the sensor integrates cleanly with existing instrumentation infrastructure.

PT100 connection methods , 2-wire, 3-wire, 4-wire

The connection method between a PT100 sensor and its measuring instrument has a direct and significant effect on measurement accuracy. Particularly for long cable runs common in industrial and building services installations.

2-wire connection

The two connection cables carry both the excitation current and the voltage measurement signal. The resistance of the cables (typically 0.1Ω per metre for copper wire), adds directly to the measured resistance and appears as a systematic positive temperature error. For a 10-metre cable, this error is approximately 2.6°C at 0°C. 2-wire connection is acceptable only for very short cable runs (under 2 metres) or applications where measurement uncertainty greater than 1°C is tolerable. It should not be used for GMP, HACCP, or laboratory applications.

3-wire connection

A third cable allows the instrument to measure the resistance of one connection leg and subtract it from the total measurement, on the assumption that both legs have equal resistance. In practice, if both cables are the same gauge and length (a reasonable assumption for most installations), 3-wire connection reduces lead resistance error to a small residual. This is adequate for most industrial process applications and is the most common PT100 connection method in Singapore building services and manufacturing.

4-wire (Kelvin) connection

Separate pairs of terminals handle current injection and voltage measurement independently. Because no current flows through the voltage measurement leads, their resistance contributes zero error. 4-wire connection completely eliminates lead resistance error regardless of cable length, temperature coefficient of the cable, or mismatch between the two legs. This is required for laboratory calibrations, pharmaceutical GMP critical measurements, and any application where ±0.1°C accuracy is needed with cable runs over 5 metres.

Type K thermocouple. How it works and Singapore applications

A thermocouple generates a small voltage (the Seebeck voltage), at the junction of two dissimilar metal wires. The voltage is proportional to the temperature difference between the measurement junction (the tip of the probe, inserted into the process) and the reference junction (typically inside the measurement instrument, at a known or measured ambient temperature). The instrument converts this voltage to temperature using the NIST thermocouple reference tables for the specific thermocouple type.

Type K uses chromel (Ni-Cr alloy, positive leg) and alumel (Ni-Al alloy, negative leg). It is the most widely used thermocouple type worldwide. Valued for its wide operating range (-200°C to +1260°C), robust construction, low cost, and excellent availability of calibrated extension wire, connectors, and instrumentation.

In Singapore industrial settings, Type K thermocouples are the default choice for heat treatment furnaces (700–1050°C), engine exhaust gas temperature monitoring, autoclave internal probes during biological indicator challenge testing, soldering and reflow oven temperature profiling, and kiln monitoring in ceramics and glass production. Their fast response (millisecond response for bare-wire junctions), ability to be fabricated into fine-gauge probes that fit tight spaces, and tolerance of mechanical shock make them the practical choice where PT100s would be too fragile or too slow.

Type K drift above 500°C. The major limitation

The most important characteristic of Type K thermocouples that affects calibration planning is their tendency to drift when operated continuously above approximately 500°C. This is not a calibration failure. It is a physical change in the sensor material that alters its thermoelectric output.

Above 500°C, the chromel leg undergoes preferential oxidation at the grain boundaries near the wire surface. Chromium, which gives chromel its high Seebeck coefficient, migrates to the surface and oxidises. Leaving a depleted zone in the wire that generates a different voltage for the same temperature. A Type K probe that reads correctly at 1000°C in January may read 5–15°C high by June under continuous operation, even with correct cold-junction compensation and accurate instrumentation.

The practical consequence for Singapore heat treatment and furnace operators is straightforward: Type K thermocouples above 600°C require both a defined calibration interval (typically 6 months) and a planned physical replacement cycle (typically annual). Calibration can detect and document the drift; only replacement eliminates it. For applications requiring better stability above 700°C, Type N thermocouples (Nicrosil-Nisil) are significantly more resistant to this mechanism and are increasingly specified in aerospace heat treatment (AMS2750) and materials testing.

NTC thermistors. Pharmaceutical and laboratory use

NTC (negative temperature coefficient) thermistors are semiconductor devices whose resistance decreases exponentially with increasing temperature. Within their optimal range of 0–100°C, a high-quality NTC thermistor can achieve accuracy of ±0.01°C (roughly 10 to 50 times better than a PT100 in the same range), because the large resistance change per degree allows very high-resolution measurement with simple circuitry.

This precision comes with important constraints. The response is highly non-linear: the resistance of a 10kΩ thermistor at 0°C might be 32kΩ, falling to 10kΩ at 25°C and 4kΩ at 50°C. Each thermistor type has its own characteristic curve (the Steinhart–Hart equation), and even sensors nominally identical may have individual variations requiring individual calibration. Thermistors are also fragile, sensitive to moisture ingress, and unsuitable for industrial environments involving vibration, chemical exposure, or temperatures above 150°C.

In Singapore, thermistors are the appropriate technology for: precision laboratory incubators (37°C, ±0.1°C), pharmaceutical stability chambers calibrated to ICH Q1A (25°C or 40°C, ±0.5°C), pharmaceutical cold-chain data loggers (2–8°C), NIST-traceable precision thermometers used as working reference standards, and medical equipment including patient temperature monitors and dialysis machine temperature controls.

The calibration approach for thermistors differs from PT100s and thermocouples. Because the individual thermistor cannot easily be removed and calibrated separately, calibration is best performed on the complete measurement system (the thermistor in its housing, connected to its signal conditioning and readout instrument), as a single unit. Unitest calibrates precision thermometers, pharmaceutical data loggers, and temperature measurement systems containing thermistors as complete instruments against ITS-90 traceable references.

Self-heating error. Affects both PT100s and thermistors

Both PT100 RTDs and thermistors are resistive sensors. Passing electrical current through them causes heating via the I²R effect. The sensor element dissipates power as heat, raising its temperature slightly above the actual temperature of the medium being measured. This systematic positive error is called self-heating error.

For PT100 sensors, the standard excitation current is 1mA. At this current level, a well-immersed industrial PT100 element dissipates approximately 0.1µW of power. Producing a self-heating effect of 0.02–0.05°C in typical process conditions. In poorly stirred or stagnant media (dead gas pockets, low-velocity ducts), self-heating can be higher and must be characterised during sensor selection. Reducing excitation current to 0.3mA or 0.1mA reduces self-heating but also reduces signal level and increases susceptibility to noise. A trade-off that precision laboratory transmitters manage automatically.

For NTC thermistors, the problem is more significant. A 10kΩ thermistor at 25°C carrying 1mA dissipates 10mW (far more than a PT100), which would cause severe self-heating and measurement error. Thermistor-based instruments use very low excitation currents, typically 10–100µA, and may use intermittent excitation (pulsed measurement) to minimise the total energy deposited in the sensor element. Well-designed pharmaceutical data loggers and precision thermometers specify their self-heating error as part of the instrument specification.

Type K thermocouples are voltage-generating devices. No excitation current flows through the measurement circuit, and self-heating is therefore not a concern in thermocouples. This is one advantage of thermocouple technology in high-resistance or thermally sensitive environments.

Calibration approaches for each sensor type

The calibration method, interval, and traceability requirements differ across the three sensor types, and choosing the wrong calibration approach for a given sensor can produce a certificate that looks valid but does not reflect the sensor's actual in-service performance.

PT100 RTD calibration

PT100 sensors are calibrated by comparison in a controlled temperature environment. Typically a liquid bath (oil or water) for best uniformity, or a dry block calibrator for field-portable work. The PT100 under calibration is immersed alongside a reference thermometer (a calibrated SPRT or a secondary reference PT100 traceable to the NMC) and readings are compared at multiple temperature setpoints across the operating range. Alternatively, the PT100 element can be measured for resistance at controlled temperatures and the resistance converted to temperature using IEC 60751 tables, but this requires precise resistance measurement equipment and careful temperature control.

For pharmaceutical GMP applications, calibration is performed at a minimum of three points across the operating range, including the critical control points (e.g. 2°C and 8°C for a cold room, 25°C and 40°C for a stability chamber). The measurement uncertainty of the calibration must be stated on the certificate, and the expanded uncertainty must be significantly smaller than the application's allowable tolerance.

Type K thermocouple calibration

Thermocouple calibration is performed by comparison in a controlled environment appropriate to the temperature range: a liquid bath or dry block for measurements below 300°C; a tube furnace with temperature-controlled insert for measurements above 600°C. The thermocouple under calibration is placed alongside a reference SPRT (or reference thermocouple of higher accuracy) and readings are compared at the setpoints of interest. Cold-junction compensation must be correct during calibration. Errors in cold-junction measurement appear directly as systematic errors in the thermocouple reading.

For high-temperature furnace thermocouples, multi-point calibration across the full operating range is important. The thermocouple sensitivity (µV/°C) varies by up to 20% across the Type K range, and a single-point calibration at one temperature cannot correct for non-linearity at other temperatures. Where thermocouple drift is suspected (continuous high-temperature service), calibration should be performed on the actual probe that has been in service. Not a new probe of the same type.

NTC thermistor calibration

As noted above, thermistors are best calibrated as part of their complete instrument. The data logger, precision thermometer, or monitoring system as a whole. The calibration consists of immersing the probe in a liquid bath alongside a reference thermometer, recording both at each setpoint, and determining the instrument's error and uncertainty at each point. The certificate documents the instrument's performance, not the individual thermistor element's resistance curve. For pharmaceutical data loggers used in stability chambers or cold rooms, calibration at the actual operating temperature (25°C, 40°C, or 5°C) produces the most relevant uncertainty estimate for compliance purposes.

Singapore-specific sensor selection guide

The following recommendations reflect the sensor selection decisions most frequently made correctly (and incorrectly), in Singapore industrial, pharmaceutical, food, and laboratory applications.

Pharmaceutical cold room (2–8°C): PT100 Class A, 4-wire or 3-wire depending on cable run, connected to a pharmaceutical-grade transmitter or data logger. Annual calibration at a minimum of three temperature points. SAC-SINGLAS accredited certificate required for HSA GMP audit compliance.

Food blast freezer (-35°C): PT100 or Type T thermocouple (Type T has superior low-temperature accuracy compared to Type K below -40°C and is NIST-recommended for cryogenic applications). Stainless steel sheathed probe with food-grade fittings. 12-month calibration interval for HACCP critical control point instruments.

HVAC chilled water system (5–20°C): PT100 3-wire immersion sensor with transmitter (4–20mA output to BAS). Cable runs in building services typically exceed 20 metres, making 3-wire connection the minimum. 12-month calibration interval for commissioning and maintenance documentation under BCA Green Mark or ISO 50001.

Heat treatment furnace (800–1050°C): Type K thermocouple (or Type N for better stability). 6-month calibration interval; annual probe replacement under a planned maintenance schedule. For aerospace or automotive heat treatment, refer to AMS2750 (aerospace) or CQI-9 (automotive) for mandatory calibration intervals and system accuracy tests.

Precision laboratory incubator or stability chamber (0–100°C): NTC thermistor in calibrated precision thermometer or data logger. Annual calibration at operating temperature against ITS-90 traceable reference. Expanded uncertainty of calibration should be less than one-fifth of the application tolerance to provide adequate measurement confidence margin.

Autoclave sterilisation monitoring (121°C, 134°C): PT100 for the continuous process temperature display; Type K thermocouple probes (or thermocouple data loggers) for biological indicator challenge load distribution studies. The PT100 process sensor is calibrated annually; challenge test probes are calibrated before each qualification run or per the site's SOP.