Coolant Temperature Sensors: NTC vs PT1000 and Why the Resistance Table Matters
How a coolant temperature sensor is specified, why the resistance curve must match the ECU calibration, and the two checks that find most faults.
The resistance curve is the calibration
Almost every coolant sensor on the road is an NTC thermistor: resistance falls as temperature rises. A typical curve reads around 2.5 kΩ at 20 °C and about 180 Ω at 90 °C. But the exact values are not standard — they are whatever the engine calibration expects, and two sensors that look identical and fit the same bore can be 30 % apart at operating temperature.
An ECU that receives a resistance 30 % below expectation believes the engine is hotter than it is. It enriches the mixture, retards ignition, and may run the electric fan early. None of that produces a fault code, because the value is inside the plausible range. The car simply runs worse and nobody can say why.
PT1000 elements are a different animal
Platinum resistance elements — PT100, PT1000 — are used where the measurement has to stay accurate across a very wide range or survive sustained high temperature: exhaust gas, battery packs, industrial coolant loops. They are more linear and far more stable over life than an NTC, and correspondingly more expensive.
They are not interchangeable with an NTC sensor in the same application. The signal conditioning in the ECU assumes one curve or the other. A PT1000 substituted into an NTC circuit will read plausibly and be wrong by a large margin — the same failure as a mis-specified NTC, just in the other direction.
The two checks that find most faults
First, measure resistance cold and compare it to the published value at the ambient temperature you measured. A sensor that is within 5 % cold is usually within spec everywhere, because the curve is a physical property of the element rather than something that drifts independently.
Second, watch the reported temperature on a scan tool while the engine warms from cold. It should rise smoothly and level off near the thermostat rating. A reading that jumps, plateaus early, or tracks 15 °C above a second sensor reading the same coolant is a wiring or ground problem far more often than a failed element.
Common questions
Are all coolant temperature sensors the same thread?
No. Bore thread, seat angle and the depth of the sensing tip all vary by engine family. A sensor that threads in but seats 3 mm proud can read the wrong layer of the coolant flow.
Why does my fan run with a cold engine?
Most often a sensor reading high, but check the wiring and the fan relay before replacing the sensor. A corroded ground on the sensor circuit reads as high resistance, which an NTC converts into a low-temperature reading — and some strategies run the fan as a fail-safe on implausible input.
Does coolant type affect the sensor?
Not the reading, but silicate-heavy coolants coat the sensing tip over years and slow its response. The static resistance stays correct; the dynamic lag is what degrades.
Need the part number matched?
Send the OE number and the target vehicle. We return the cross-reference, the datasheet and a price — usually inside one working day.