RTD and PT100 temperature measurement is a method of inferring temperature from the predictable change in electrical resistance of a pure metal, most commonly platinum. Resistance temperature detectors (RTDs) are the workhorse sensor for process temperature from about -200°C to 850°C, the range where accuracy and long-term stability matter more than response speed. Understanding the resistance-temperature relationship, wiring configuration, and tolerance class is essential for specifying a sensor that actually delivers the accuracy a process needs.
Platinum, like most pure metals, has electrical resistance that increases almost linearly with temperature. A precision resistance element is wound or deposited in a known pattern so its resistance at any temperature can be predicted from a standard curve. Passing a small, stable excitation current through the element and measuring the voltage drop gives a resistance value that a transmitter converts to a temperature reading using the Callendar-Van Dusen equation. Because the relationship is a physical property of the metal rather than a fragile calibration curve, RTDs hold their accuracy for years if the element is not mechanically or chemically damaged.
The "PT" designates platinum; the number is the nominal resistance in ohms at 0°C. A PT100 reads 100.00 ohms at 0°C; a PT1000 reads 1000.00 ohms at 0°C. Both follow the same temperature coefficient, so a PT1000 simply produces ten times the signal change per degree.
The temperature coefficient of resistance, alpha, defines the average fractional resistance change per degree Celsius between 0°C and 100°C. Two curves dominate industrial use:
Mixing curves in a transmitter configuration is a common root cause of a temperature reading that is systematically off by several degrees despite a sensor that measures correctly at the calibration point. Always confirm the alpha value matches between sensor and transmitter, not just the PT100/PT1000 base value.
Because an RTD measures resistance, any resistance in the connecting wires adds directly to the measured value unless the circuit compensates for it. Copper lead wire has resistance around 0.08 ohms per metre for a single 0.5 mm² conductor, and that resistance changes with ambient temperature along the cable run, adding further drift.
| Configuration | Lead error compensation | Typical accuracy impact | Common use |
|---|---|---|---|
| 2-wire | None; lead resistance adds directly to reading | On a PT100 (alpha 0.00385), each ohm of total lead resistance shifts the reading by roughly 2.6°C; a 1.2 ohm run (about 50 m of light-gauge copper both ways) can add 3°C of error | Short runs, low-cost HVAC, PT1000 sensors where the error fraction is small |
| 3-wire | Bridge-style cancellation assumes equal resistance in two leads | Residual error only from lead-to-lead mismatch; typically small when both leads are matched in length and gauge | Standard for industrial process transmitters |
| 4-wire | Full Kelvin sensing removes lead resistance from the measurement entirely | Near-zero lead error, limited only by transmitter accuracy | Calibration labs, custody transfer, critical process loops |
For any run beyond a few metres on a PT100, 3-wire is the practical minimum. Four-wire is reserved for reference-grade measurement because the extra conductor rarely justifies its cost on routine process points.
IEC 60751 defines tolerance classes that state the maximum permissible deviation from the standard curve at a given temperature. The tolerance is not a flat number; it widens with temperature.
| Class | Tolerance formula (°C) | Tolerance at 0°C | Tolerance at 400°C |
|---|---|---|---|
| Class AA (1/3 DIN) | ±(0.10 + 0.0017 × |t|) | ±0.10°C | ±0.78°C |
| Class A | ±(0.15 + 0.002 × |t|) | ±0.15°C | ±0.95°C |
| Class B | ±(0.30 + 0.005 × |t|) | ±0.30°C | ±2.30°C |
Class B is the general industrial default and is adequate for most control loops. Class A or AA is specified where tight control bands or custody-transfer accuracy are required, for example on a fired-heater outlet or a fiscal metering skid. Class AA elements typically cost more and are more sensitive to installation stress, so overspecifying tolerance on a non-critical loop adds cost without benefit.
The excitation current used to measure resistance also dissipates power in the element (I²R), raising its temperature slightly above the process it is measuring. Self-heating is characterized by a dissipation constant in mW/°C, which varies widely with element construction, sheath design, and whether the sensor sits in a fast-moving fluid or still air; well-mounted industrial sensors in flowing liquid typically dissipate more power per degree of self-heating (a higher constant) than a bare element in still air. In liquids with good heat transfer, self-heating error is usually under 0.05°C; in still air or a poorly conducting sheath, it can exceed 1°C. Keeping excitation current low (typically under 1 mA) and ensuring good thermal contact between the sensor tip and the process, including correct immersion depth and a well-fitted thermowell, keeps this error negligible. Poor thermowell fit is also one of the most common causes of slow thermal response and reading lag flagged during a plant's condition monitoring review.
Feeding an RTD signal into a CMMS-linked monitoring system lets a maintenance team trend actual sensor drift over time rather than reacting only when a reading looks obviously wrong. Where sustained drift or an out-of-tolerance reading is detected, an automatically generated work order in a platform like Book a Fabrico demo routes calibration or replacement tasks before the deviation affects product quality or triggers a trip.
RTDs and thermocouples solve the same problem with different tradeoffs, and the choice usually comes down to range, accuracy, and response time.
For general process control between minus 200°C and 500°C where accuracy and long-term stability drive value, an RTD is almost always the better default choice.
Both use the same platinum resistance-temperature curve and coefficient options, but a PT100 reads 100 ohms at 0°C while a PT1000 reads 1000 ohms at 0°C. PT1000 produces a larger resistance change per degree, which makes 2-wire lead-resistance error much less significant over long cable runs.
The most common cause is an alpha coefficient mismatch, configuring a transmitter for 0.00392 when the installed sensor uses the IEC 60751 standard of 0.00385, or vice versa. This produces a temperature-dependent error that grows away from 0°C even though the sensor itself is healthy.
Yes, for the large majority of industrial loops. Three-wire wiring cancels lead resistance as long as both lead conductors are matched in length and gauge, leaving only a small residual error. Four-wire is reserved for calibration-grade or custody-transfer applications.
Class B (±0.30°C at 0°C, widening with temperature) suits general process monitoring and most control loops. Class A or AA should be specified only where the control band is tight or the measurement feeds a fiscal or safety-critical calculation, since tighter classes cost more and are more sensitive to mechanical stress during installation.
Connexion
Blog
