If you are asking which temperature device uses resistance to register temperature change, the direct answer is twofold: Thermistors (specifically NTC and PTC types) and Resistance Temperature Detectors (RTDs). Both rely on the predictable shift in electrical resistance as a function of temperature, but they do so using entirely different materials and physical mechanisms.
As a builder or electrical designer, you don't just select these sensors to read a display. You use them to validate your thermal management math, ensure your power semiconductors aren't cooking themselves inside an enclosure, and trigger shutdown routines before silicon melts. Below, we break down the sensors, the thermal path math they help you verify, and how to spec a heatsink based on real-world wattage.
Which Temperature Device Uses Resistance to Register Temperature Change?
The choice between a thermistor and an RTD comes down to your required temperature range, linearity, and budget. Thermistors are semiconductor-based and highly sensitive but non-linear. RTDs are metal-based (usually platinum), highly linear, and stable over massive temperature swings.
| Feature | NTC Thermistor (e.g., Vishay NTCLE100E3103) | PTC Thermistor (e.g., Bourns MF-MSMF) | RTD PT1000 (e.g., TE Connectivity) |
|---|---|---|---|
| Resistance Behavior | Resistance drops as temp rises (Negative Coefficient) | Resistance spikes sharply at a threshold (Positive Coefficient) | Resistance rises linearly as temp rises (Positive Coefficient) |
| Typical Base Value | 10kΩ or 100kΩ at 25°C | Variable (often used as resettable fuses) | 1000Ω at 0°C |
| Linearity | Highly non-linear (requires Steinhart-Hart equation) | Non-linear (switch-like) | Highly linear (approx 3.85 Ω/°C for PT1000) |
| Operating Range | -55°C to +125°C (up to 300°C for specialized glass) | -40°C to +85°C | -200°C to +850°C |
| Cost (2026 Est.) | $0.10 - $0.25 per unit | $0.30 - $0.60 per unit | $8.00 - $15.00 per probe |
| Best Use Case | 3D printer hotends, battery pack monitoring, ESP32 ambient sensing | Overcurrent/overtemp protection, motor winding protection | Industrial process control, precision lab equipment, HVAC |
For most hobbyist and prosumer electronics workbench projects, the 10kΩ NTC thermistor (with a B-value of 3950K) is the default choice. It interfaces easily with a microcontroller's ADC via a simple voltage divider. For precision mains-voltage or industrial applications, the PT1000 RTD is the standard, usually read via a dedicated SPI ADC like the MAX31865.
Thermal Path Math: Validating Sensor Readings with Rθ
Once you have your resistance-based sensor strapped to a component, what do the numbers actually mean? You are measuring the result of the thermal resistance network, denoted as Rθ (theta), measured in °C/W. This is the thermal equivalent of Ohm's Law: Temperature Rise = Power Dissipation × Thermal Resistance.
The critical metric for any power semiconductor (like a MOSFET or voltage regulator) is the junction temperature (TJ). The formula is:
TJ = TA + (PD × RθJA)
Where:
- TA = Ambient temperature inside the enclosure (°C)
- PD = Power dissipated by the part (Watts)
- RθJA = Total thermal resistance from Junction to Ambient (°C/W)
RθJA is a sum of three distinct physical barriers:
- RθJC (Junction-to-Case): Internal to the silicon package. For a standard TO-220, this is typically ~1.5 °C/W. You cannot change this.
- RθCS (Case-to-Sink): The interface material. Bare metal-to-metal is terrible (~2.0 °C/W). A thin layer of Arctic Silver thermal paste drops this to ~0.3 °C/W. A silicone thermal pad might be ~1.0 °C/W.
- RθSA (Sink-to-Ambient): The heatsink itself. This is the only variable you control via part selection and airflow.
Heatsink Selection and Derating: A Real-World MOSFET Example
Let's apply this math to a real scenario. You are switching a 15A load using an Infineon IRFP460N N-channel MOSFET. The datasheet states the absolute maximum TJ is 175°C. However, as outlined in Texas Instruments' thermal metrics application notes, running silicon at its absolute limit guarantees premature death. We want to keep TJ under 110°C for long-term reliability.
The Wattage Basis: The MOSFET has an RDS(on) of 0.026Ω. At 15A continuous, power dissipation (PD = I²R) is 5.85W. Let's round to 6W for safety margin. Assume worst-case enclosure ambient (TA) is 45°C.
Calculating Required RθSA:
- Max allowed RθJA = (110°C - 45°C) / 6W = 10.8 °C/W
- We know RθJC is 1.0 °C/W and RθCS (with thermal paste) is 0.3 °C/W.
- Therefore, our heatsink (RθSA) must be: 10.8 - 1.0 - 0.3 = 9.5 °C/W or lower.
The Heatsink Pick: A Aavid Thermalloy 530624B02500G (approx. $6.50 on Mouser) has a natural convection RθSA of about 8.5 °C/W. This fits our budget with 1 °C/W to spare.
Interpreting the Derating Curve
If you look at the IRFP460N datasheet, the Power Derating Curve shows a linear drop-off. At a case temperature (TC) of 25°C, it can handle 230W. By the time TC hits 175°C, the allowed power is 0W. The slope of this line is exactly 1 / RθJC. If your NTC thermistor strapped to the metal tab reads 80°C, you can cross-reference this curve to know exactly how much headroom you have left before the silicon throttles or fails.
What Airflow and Enclosure Changes Buy You
If your thermistor alarms show the tab is still running too hot, you have two mechanical levers:
- Forced Air: Adding a 40mm Sunon MagLev fan pushing 10 CFM across the Aavid heatsink will drop its RθSA from 8.5 °C/W down to roughly 3.5 °C/W. This drastically flattens the thermal gradient.
- Enclosure Venting: A sealed IP65 enclosure will cause TA to rise as the internal air absorbs heat. Adding filtered intake/exhaust louvers with a passive chimney effect prevents TA from creeping above room temperature, preserving your ΔT (Delta T) headroom.
Failure Signatures: How Hot is Too Hot and What Breaks First
How hot is too hot? While silicon survives up to 175°C, the Arrhenius equation dictates that the failure rate of electronic components roughly doubles for every 10°C increase above a baseline (usually 85°C or 105°C). Running a TO-220 at 140°C isn't instantly fatal, but it will age the part exponentially faster.
Never parallel MOSFETs in linear (analog) mode without individual source resistors. Unlike switching mode, a MOSFET's threshold voltage (VGS(th)) has a negative temperature coefficient. If one FET gets slightly hotter, it hogs more current, gets hotter still, and enters thermal runaway until it violently shorts. Always use an NTC thermistor on the heatsink to monitor the array and trip a hardware comparator if temps exceed 90°C.
When thermal limits are repeatedly violated, you won't always see a melted package. Look for these subtle failure signatures of thermal stress:
- Solder Joint Creep: Repeated thermal cycling (expansion and contraction) causes micro-cracks in the solder connecting the component leads to the PCB, leading to intermittent high-resistance connections.
- Wire Bond Lift-Off: Inside the IC package, the microscopic aluminum wires bonding the silicon die to the leadframe can detach due to CTE (Coefficient of Thermal Expansion) mismatch, resulting in a sudden, permanent open circuit.
- Delamination: In TO-247 or multi-chip modules, the epoxy/plastic body can separate from the copper leadframe, destroying the thermal path and causing instantaneous over-temp failures.
Frequently Asked Questions
Which temperature device uses resistance to register temperature change in a 3D printer hotend?
Most modern 3D printers (like those running Klipper or Marlin) use either a 100kΩ NTC glass bead thermistor (like the Semitec 104NT-4-R025H42G) or a PT1000 RTD. The NTC is cheap and perfectly adequate for the 0-280°C range of standard hotends. The PT1000 is preferred for high-temperature engineering materials (like PEEK or Ultem) because it doesn't suffer from the severe non-linearity and ADC resolution loss that NTCs experience at the top end of their range.
How does a resistance-based temperature sensor differ from a thermocouple?
A resistance device (thermistor/RTD) requires an external excitation current to measure the voltage drop across the changing ohmic resistance. A thermocouple (Type K, J, etc.) generates its own tiny millivolt signal via the Seebeck effect—the physical phenomenon where two dissimilar metals joined at a junction produce a voltage proportional to the temperature gradient. Thermocouples are better for extreme heat (up to 1200°C), while resistance devices are vastly superior for precision in the -50°C to +300°C range.
Can I use a standard multimeter to test which temperature device uses resistance?
Yes. Set your multimeter to the Ohms (Ω) range and connect the probes to the sensor leads. For an NTC thermistor, pinch the bead or apply mild heat from a heat gun; the resistance should drop smoothly (e.g., from 10kΩ down to 2kΩ). If the resistance climbs steadily as you apply heat, you are holding a PTC thermistor or an RTD. If the reading stays completely flat regardless of heat, the sensor is dead or it's a solid-state IC sensor (like a DS18B20) which outputs digital data, not raw resistance.
What causes a resistance temperature sensor to fail or drift out of calibration?
The most common cause of drift in NTC thermistors is moisture ingress through microscopic cracks in the epoxy coating, which alters the semiconductor's baseline resistance. For RTDs, physical shock or severe vibration can strain the fine platinum wire wound inside the ceramic core, causing a permanent shift in the 0°C baseline. Always use potted or stainless-steel sheathed probes in humid or high-vibration environments like motor housings or outdoor solar charge controllers.
