RTDs (resistance temperature detectors)
The resistance of platinum rises predictably with temperature. A Pt100 has 100 Ω at 0 °C and, with the standard IEC 60751 curve (α = 0.00385), changes by about 0.385 Ω/°C. Pt1000 has ten times the resistance and sensitivity, which reduces lead-wire error.
R(t) = R₀ · (1 + A·t + B·t²) for 0 … 850 °C
A = 3.9083×10⁻³ B = −5.775×10⁻⁷
Tolerance classes (IEC 60751)
| Class | Tolerance | At 100 °C |
|---|---|---|
| AA | ±(0.10 + 0.0017·|t|) °C | ±0.27 °C |
| A | ±(0.15 + 0.002·|t|) °C | ±0.35 °C |
| B | ±(0.30 + 0.005·|t|) °C | ±0.80 °C |
Convert between resistance and temperature, and see the tolerance at any temperature, with the RTD calculator.
2-wire, 3-wire and 4-wire connections
- 2-wire: the lead resistance adds directly to the sensor reading. With Pt100, 1 Ω of total lead resistance is about +2.6 °C of error. Acceptable only for very short leads or head-mounted transmitters.
- 3-wire: the industry standard. The transmitter compensates lead resistance, assuming all three leads are equal, so use one cable and don't mix conductor sizes.
- 4-wire: the measuring current and the voltage sense use separate pairs, so lead resistance is eliminated. Used for high accuracy and in laboratories.
Thermocouples
Two dissimilar metals joined at the hot end produce a small voltage (the Seebeck effect, µV/°C) that depends on the temperature difference between the hot (measuring) junction and the cold (reference) junction.
| Type | Materials | Typical range | Notes |
|---|---|---|---|
| K | NiCr / NiAl | −200 … 1260 °C | Most common general purpose, ~41 µV/°C |
| J | Fe / CuNi | 0 … 760 °C | Older plants; the iron leg oxidises |
| T | Cu / CuNi | −200 … 370 °C | Good at low temperatures |
| E | NiCr / CuNi | −200 … 870 °C | Highest output (~68 µV/°C) |
| N | NiCrSi / NiSi | −200 … 1260 °C | More stable than K at high temperature |
| R, S | Pt-Rh / Pt | 0 … 1480 °C | Furnaces, high accuracy; expensive |
| B | Pt-30Rh / Pt-6Rh | 600 … 1700 °C | Almost no output below 50 °C |
Cold junction compensation and cables
The transmitter measures the terminal temperature and adds the equivalent mV (cold junction compensation). This only works if the thermocouple metal continues all the way to the transmitter terminals: use the correct extension or compensating cable of the same type. Ordinary copper cable in between creates a new junction and a new error.
RTD vs thermocouple
| RTD (Pt100) | Thermocouple | |
|---|---|---|
| Range | −200 … 600 °C (850 °C max) | −200 … 1700 °C |
| Accuracy & stability | Better | Lower; drifts with age |
| Response | Slower | Faster (small junction) |
| Vibration | More fragile | Rugged |
| Cost | Higher | Lower |
Thermowells
The thermowell protects the sensor and lets it be removed without depressurising the line. Two things matter:
- Mechanical design: fluid flowing past the well sheds vortices. If the shedding frequency approaches the well's natural frequency, it can resonate and break off. Thermowells for gas and high-velocity services must be checked to ASME PTC 19.3 TW-2016 (wake frequency, stress and pressure). Use shorter, thicker or tapered wells, or velocity collars, where needed.
- Response & immersion: immerse the tip at least 10× its diameter, or to the pipe centreline, and use a spring-loaded sensor so the tip touches the bottom of the bore.
Quick troubleshooting
- Reading goes to maximum or minimum: open sensor or broken wire. Check the transmitter's burnout direction (upscale or downscale).
- RTD reads high by a constant offset: lead resistance (2-wire), or unequal leads in a 3-wire connection.
- Thermocouple reads wrong by tens of degrees: wrong type configured, reversed polarity, or copper cable in the circuit.
- Noisy or jumping reading: poor screen earthing, earth loops, or a grounded junction picking up interference. Earth the screen at one end only.