PT100 Sensors — Tolerance Classes, 2/3/4-Wire Connection and PT100 vs PT1000
Standards explainer compiled by cBallast. 3-page PDF; the full content is on this page.
Compiled by
cBallast
Subject
Temperature sensors
Document type
Standards explainer
Reference
2026 edition
Issued
2026-10-18
Pages
3
Format
Web page and PDF
What a PT100 actually is (100 Ω of platinum at 0 °C per IEC 60751), how much accuracy class AA, A and B really buy, why lead resistance dictates the 2-, 3- or 4-wire decision, when PT1000 makes the problem vanish, and the checks that separate a drifting sensor from a bad connection.
The platinum contract
A PT100 is a platinum resistor laser-trimmed to 100.00 Ω at 0 °C with a standardised slope (≈0.385 Ω/°C), defined by IEC 60751 — which is why any instrument reading "PT100" understands any maker's sensor. Accuracy is bought by class:
2-wire: lead resistance adds straight onto the reading — 0.4 Ω of cable reads as +1 °C. Acceptable only for short runs and rough duty.
3-wire: the instrument measures one lead and subtracts it twice — cancels leads if they match. The industrial standard, good to a corroded terminal's first surprise.
4-wire: separate current and sense pairs, lead resistance vanishes entirely — the class-A-and-better companion.
PT1000 sidesteps it: ten times the base resistance makes lead ohms ten times less significant — why building automation and compact transmitters went PT1000. Instrument input must match; the slope scales identically.
Field diagnostics in three measurements
≈110 Ω at room temperature (PT100) between element wires = sensor plausible; open or megohms = broken element; a few ohms to the sheath = moisture ingress — the usual death of trace-heated and washdown installations.
Reading jumps when the cable moves = connection, not sensor: terminal corrosion defeats 3-wire compensation asymmetrically.
Slow response: a thermowell adds seconds to minutes — spring-loaded inserts, contact paste and correct immersion depth (10× diameter as the rule of thumb) recover most of it.
Quoted within 24 hours against the nameplate code or the duty: insertion length, process connection, class, wiring and head/transmitter choice — with 4–20 mA head transmitters (see the MBS pressure-side companion) where the cable run argues against raw RTD wiring.
Frequently asked questions
How accurate is a standard PT100?
Class B — ±0.3 °C at 0 °C, widening to ±0.8 °C at 100 °C per IEC 60751. Class A roughly halves it, AA again — but wiring and installation errors routinely exceed class differences, so buy class only after fixing both.
When do I need a 4-wire PT100?
When the accuracy target is tighter than the lead-resistance uncertainty — class A and better duty, long or uncertain cable runs, laboratory work. 3-wire covers ordinary plant control; 2-wire only short, rough-duty runs.
Can I connect a PT1000 to a PT100 input?
No — the instrument would read wildly high. The input must be configured (or built) for the element type; many modern transmitters accept both via configuration, which is the clean migration path.
Why does my PT100 read a few degrees high, consistently?
In 2-wire connection, that is the cable: every 0.39 Ω of loop resistance adds one apparent degree. Move to 3- or 4-wire, or offset-calibrate the loop if rewiring is impossible.
How deep must a sensor sit in the pipe?
Rule of thumb: immersion of at least ten times the probe diameter, tip in the flowing core — short immersion reads the pipe wall, not the fluid. Thermowell response adds on top; spring-loaded tips and paste close the gap.