The Temperature Reading That Drifted (Thermocouple)
We measured a temperature (a heater, 100°C). We used a thermocouple (type K). On the floor, the reading drifted (it read 95°C, not 100°C). The problem: the thermocouple’s cold junction (the connection to the panel) wasn’t compensated. The thermocouple measures the difference (hot vs cold junction). The cold junction (at the panel, which warms up) needs compensation. We either used a PT100 (no cold junction, simpler) or added the compensation (in the transmitter). We switched to a PT100 (a RTD, no cold junction). The reading was stable. The mistake was using a thermocouple (without cold junction compensation) for a low temperature (100°C, where a PT100 is better).
Temperature sensor: PT100 vs thermocouple — pick the right one. This article covers the comparison.
The Two Sensors
PT100 (RTD)
A PT100 (a resistance temperature detector) is a platinum resistor (100 Ω at 0°C). The resistance changes with temperature (predictable). It’s accurate (±0.1°C). It’s stable (no drift). It’s for a low temperature (-50 to 200°C). It needs a transmitter (to convert the resistance to a signal). No cold junction (simple).
Thermocouple (Type K)
A thermocouple (two metals, joined) generates a voltage (proportional to the temperature difference). It’s for a high temperature (up to 1200°C). It’s less accurate (±1°C). It needs cold junction compensation (the cold junction, at the panel, must be measured/compensated). For a high temperature (a furnace), it’s the choice.
| Sensor | Range | Accuracy |
|---|---|---|
| PT100 (RTD) | -50 to 200°C | ±0.1°C (stable) |
| Thermocouple (K) | 0 to 1200°C | ±1°C (needs compensation) |
Step 1: The Temperature (Range)
Is the temperature low (under 200°C)? Use a PT100 (accurate, stable). Is it high (over 200°C, a furnace)? Use a thermocouple. The reading that drifted was a thermocouple on a 100°C (low) application. Use a PT100 (better for low).
The temperature rule: Low (under 200°C) → PT100 (accurate, stable). High (over 200°C) → thermocouple. The reading that drifted was a thermocouple on 100°C. Use a PT100. Don’t use a thermocouple for a low temperature (it needs compensation, and it’s less accurate).
Step 2: The Accuracy (PT100)
A PT100 is accurate (±0.1°C). For a process that needs precision (a temperature control, ±1°C), use a PT100. A thermocouple (±1°C) is less accurate (for a high temp, where precision is less critical).
Step 3: The Cold Junction (Thermocouple)
A thermocouple needs cold junction compensation (the cold junction, at the panel, is measured). The transmitter (or the PLC) adds the compensation. Without it, the reading drifts (the panel warms). For a thermocouple, use a transmitter (with built-in compensation).
Step 4: The Cable (Extension)
A thermocouple needs special extension wire (the same metals, type K). A PT100 needs a 3-wire (or 4-wire) cable (to compensate the lead resistance). For a long run, the PT100’s lead resistance matters (use 3-wire). The thermocouple’s extension wire is special (and expensive).
A Temperature Checklist
- What is the temperature? (°C?)
- Is it low (under 200)? (PT100?)
- Is it high (over 200)? (Thermocouple?)
- Is the accuracy needed? (PT100?)
- Is the cold junction compensated? (Thermocouple?)
- Is the cable right? (3-wire? Extension?)
- Does the reading drift? (Test?)
- Is the transmitter right? (Signal?)
- Is the probe installed? (Immersion?)
- Is the response time OK? (Fast?)
The Bottom Line
Temperature sensor: PT100 vs thermocouple matches the sensor to the range. The reading that drifted was a thermocouple on 100°C. Use a PT100 (low, accurate). For high (over 200), thermocouple (with compensation). The reading that was stable wasn’t the hottest sensor — it was a PT100.
