Calibrating a PT100 sensor means comparing its readings against a known temperature reference and correcting any deviation. The process can be as simple as checking the sensor in an ice bath at 0 °C or as thorough as running multiple test points in a dry-block calibrator with a certified reference probe. The method you choose depends on the accuracy you need and the equipment you have available.
How PT100 Tolerance Classes Work
Before you calibrate, it helps to know what “accurate enough” looks like. The international standard IEC 60751 defines three tolerance classes for platinum resistance sensors, each allowing a different margin of error:
- Class AA: ±(0.1 + 0.0017 × |t|) °C
- Class A: ±(0.15 + 0.002 × |t|) °C
- Class B: ±(0.3 + 0.005 × |t|) °C
In those formulas, |t| is the temperature in degrees Celsius (ignoring the sign). So a Class A sensor at 100 °C is allowed to drift up to ±0.35 °C before it’s considered out of tolerance. A Class B sensor at the same temperature gets ±0.8 °C of leeway. Knowing your sensor’s tolerance class tells you how tight your calibration needs to be and whether a simple one-point check will do or you need a multi-point procedure with a high-accuracy reference.
Three Calibration Methods Compared
There are three common approaches, each trading off speed, cost, and accuracy.
Electronic Simulation Only
A temperature simulator feeds the exact resistance value your readout device expects at a given temperature. You’re testing whether the electronics (transmitter, PLC input, display) interpret that resistance correctly. This is the fastest method because electrical signals require no stabilization time. The drawback: your actual PT100 probe stays untested. If the sensor element has drifted, you won’t catch it. This method works as a quick electronics check but does not constitute a full system calibration.
Dry-Block Calibrator Without a Reference Probe
A dry-block (or dry-well) calibrator heats or cools a metal insert to a stable, known temperature. You insert your PT100 into the block and compare its reading to the temperature shown on the calibrator’s built-in display. This tests the sensor and electronics together as a complete system, which is a significant improvement over simulation alone. The limitation is that your accuracy can only be as good as the calibrated accuracy of the dry-block unit itself.
Dry-Block Calibrator With a Reference Probe
This is the most accurate approach. You place both a certified reference probe and your PT100 into the same dry-block (or a liquid calibration bath) and compare readings directly. Because the reference probe has its own traceable calibration certificate, you’re no longer limited by the dry-block’s display accuracy. This method costs more and takes longer to set up, but it produces results you can trace back to national measurement standards.
Equipment You’ll Need
The specific tools depend on which method you’re using, but a full calibration setup typically includes:
- Temperature source: A dry-block calibrator, a liquid calibration bath, or (for a single-point check) an ice bath.
- Reference probe: A calibrated platinum resistance thermometer with a current, traceable calibration certificate. This is optional if you’re relying on the dry-block’s internal sensor, but strongly recommended for tighter accuracy.
- Readout instrument: A precision thermometer readout or a high-resolution digital multimeter capable of measuring resistance to at least 0.01 ohm. Some calibration setups use a dedicated handheld temperature calibrator that both sources and measures temperature signals.
- Appropriate insert: If using a dry-block, you need a metal insert with bore holes sized to fit your PT100 probe snugly. Air gaps between the probe and the insert wall reduce heat transfer and introduce error.
Step-by-Step Dry-Block Calibration
This procedure covers the most common in-house approach: calibrating with a dry-block and a reference probe.
Start by inserting both the reference probe and your PT100 into the dry-block. Keep them close together, ideally arranged in a circle with the reference probe in the center. This minimizes any temperature gradient between the two sensors. Connect both probes to their respective readout instruments.
Set the dry-block to your first test temperature and wait for the reading to stabilize fully. Rushing this step is the most common source of error. Once stable, record the reference probe’s temperature, your PT100’s temperature reading, and the equipment settings. If you’re doing a higher-level calibration, also record the raw resistance value from the PT100, since you may need it later to calculate new correction coefficients.
Repeat this at each test point. For industrial PT100 sensors, three to five points spread across the sensor’s working range is typical. At each point, compare the PT100 reading to the reference and calculate the error. If the error falls within the tolerance for your sensor’s class, the sensor passes. If it doesn’t, you can either adjust the connected instrument’s settings to compensate or replace the sensor.
For higher-precision work, the recorded resistance values can be used to recalculate the sensor’s characteristic curve. Platinum sensors follow a relationship described by the Callendar-Van Dusen equation, which uses three coefficients (commonly labeled A, B, and C) to convert resistance into temperature. The C coefficient only applies below 0 °C, so for sensors used exclusively above freezing, the math simplifies to just two coefficients. Your calibration software or readout device may let you enter updated coefficients directly, which effectively fine-tunes the sensor’s accuracy across its entire range.
Ice Point Check as a Quick Verification
If you need a fast, low-cost sanity check rather than a full multi-point calibration, the ice point method is hard to beat. Water and ice in equilibrium produce a naturally stable 0.00 °C reference point, no specialized equipment required.
Fill an insulated container (a vacuum flask works well) with finely crushed ice made from distilled or deionized water. Add just enough distilled water to fill the gaps between ice pieces without submerging the ice completely. The ratio matters: too much water and the bath temperature will rise above zero; too much ice with air pockets and you get uneven temperatures. You want a dense, slushy mixture.
Submerge the PT100’s sensing tip into the center of the ice bath. Make sure it doesn’t touch the container walls or bottom, since these can be slightly warmer than the ice-water mix. Wait for the reading to stabilize, then note how far it deviates from 0 °C. Even trace impurities in the water will lower the freezing point slightly, so using high-purity water is important if you want results better than ±0.1 °C.
An ice point check gives you one data point. It confirms whether the sensor reads correctly at 0 °C but tells you nothing about accuracy at 200 °C or 400 °C. For sensors used across a wide range, combine this with at least one or two higher-temperature test points.
How Often to Recalibrate
There’s no single universal interval. Calibration frequency depends on how the sensor is used, how harsh the environment is, and how much accuracy your process demands.
Sensors in high-temperature furnaces, chemical reactors, or other extreme conditions experience more physical stress and are more prone to drift. They generally need calibration more often than sensors sitting in a climate-controlled room. Continuous-use sensors also drift faster than ones used intermittently. In pharmaceutical or medical applications, where measurement precision is tightly regulated, more frequent calibration is expected.
A practical starting point is to follow the manufacturer’s recommended interval, then adjust based on your own calibration history. If a sensor consistently passes with minimal drift over several cycles, you can reasonably extend the interval. If it’s borderline or trending toward its tolerance limit, shorten it. Tracking the drift at each calibration gives you the data to make that call confidently rather than guessing.