Pt100/Pt1000 RTD Resistance to Temperature Calculator

Callendar–Van Dusen method

The positive-temperature branch uses A and B; the sub-zero branch adds C(T−100)T³. The inverse is solved only inside −200°C to 850°C.

Primary reference: NIST/BIPM Guide on Secondary Thermometry for Industrial Platinum Resistance Thermometers.

Pt100 and Pt1000 reference table (IEC 60751, α = 0.00385)

The table below is generated from the same Callendar–Van Dusen coefficients the calculator applies, so you can sanity-check a reading or plug a nominal value into a PLC scaling block without opening the calculator. Values are nominal ideal-sensor resistance; a real element will differ within its tolerance class.

Two properties of the curve matter more than any single row:

• Linearity is an approximation. The average slope from 0 to 100 °C is 0.38506 Ω/K, but the local slope falls from roughly 0.39 Ω/K near 0 °C to roughly 0.29 Ω/K at 600 °C. A fixed 0.385 gain in a PLC block therefore reads low at the top of the range.
• Resistance never reaches zero. Even at a perfect 0 K the model returns a positive value, which is why a reading of 0 Ω or a wildly negative temperature almost always means an open sensor, a shorted lead, or a mis-wired channel rather than a real process condition.

Tolerance classes — why your reading differs from the table

IEC 60751 defines four tolerance classes for industrial platinum elements. The tolerance is a temperature-equivalent band, not a fixed ohms offset, so the allowable resistance error grows as the element warms.

Read the class as a floor on your uncertainty budget, not a ceiling. On a Class A element at 100 °C you already accept ±0.35 °C from the element alone, before any of the following are added:

• Lead and cable resistance. A two-wire connection adds the loop resistance of both copper conductors directly to the measurement. A three-wire connection compensates only when the two current leads are genuinely matched; a four-wire connection removes the effect almost entirely. This is usually the largest single error on a two-wire PT100 run.
• Transmitter or channel error. An analogue channel carries its own gain and offset spec, commonly 0.1 % of span or worse, and it applies on top of the element tolerance.
• Reference-junction and cold-junction handling. RTDs are absolute-resistance devices, so there is no cold junction to compensate — but the excitation current reference and the ADC reference do drift.
• Self-heating. Excitation power (I²R) warms the element above the surrounding medium. In still air the effect is several times larger than in a flowing liquid; check the dissipation constant in the sensor datasheet if you are chasing tenths of a degree.

Common mistakes when converting RTD resistance

• Treating the reciprocal of the linear gain as exact. Dividing (R − 100) by 0.385 is the single most common shortcut, and it is wrong by roughly 2 °C at 100 °C and much more above 400 °C. Use the inverse curve in the calculator instead.
• Substituting Pt1000 without rescaling the channel. A 1000 Ω element needs a different excitation current and a different input range; many channels saturate long before 100 °C and then report a flat, plausible-looking value.
• Ignoring the class in the uncertainty statement. Quoting a converted value to 0.01 °C from a Class B element overstates the result by two orders of magnitude. Report the element class alongside the number.
• Using this curve to correct a real sensor. The A/B/C coefficients describe a nominal industrial platinum element. A calibration certificate with its own coefficients is what you need if you are correcting systematic error.
• Extrapolating past the stated range. The inverse solver is bounded to −200 °C to 850 °C. Pushing a resistance from beyond that range into the curve produces a confident but meaningless temperature.
• Blaming the sensor for a wiring fault. A reading that is pinned at the bottom of the range, or that swings with cable movement, is a lead or termination problem, not a resistance-to-temperature conversion problem.