Thermal derating is the mathematical reduction of a power component's maximum current capacity as its operating temperature rises above the standard 25°C baseline. When evaluating electrical engineering products like power MOSFETs, solid-state relays, or DC-DC buck converters, this parameter dictates what the component can actually handle inside your specific enclosure versus what the marketing headline claims. It changes the safe continuous current limit of a power semiconductor in a real-world environment, forcing you to calculate heat dissipation rather than just reading a single number off a box. The most common mistake makers and junior engineers make is confusing the 'Absolute Maximum Drain Current' (tested at a 25°C case temperature with an infinite, perfect heatsink) with the realistic continuous current limit at 50°C or 80°C ambient air.
The Datasheet Illusion in Electrical Engineering Products
If you buy a classic IRFZ44N MOSFET, the first page of the datasheet proudly advertises a 49A continuous drain current. If you push 49A through that TO-220 package without a massive active cooling setup, the silicon will violently fail. This is because silicon has a Positive Temperature Coefficient (PTC). As the junction temperature rises, the on-resistance (Rds(on)) increases, which generates more heat, which increases the resistance further—a feedback loop that ends in thermal runaway.
To properly evaluate these components, you must look at the Normalized Rds(on) vs. Temperature graph in the datasheet. For most standard silicon MOSFETs, at a junction temperature of 100°C, the Rds(on) is roughly 2.0 to 2.5 times higher than it is at 25°C. This means your power dissipation (calculated as I² × R) more than doubles as the part gets hot.
Worked Numeric Example: Sizing a Motor Controller FET
Let's run the math on a real-world design. You are building a 24V DC winch controller that pulls a continuous 20A. You select the IRFZ44N because it is cheap and readily available.
- Baseline Resistance: The datasheet lists Rds(on) at Vgs = 10V as 17.5 mΩ (0.0175 Ω) at 25°C.
- Baseline Power Dissipation: P = I² × R. Therefore, 20² × 0.0175 = 7 Watts.
- Thermal Resistance (Free Air): The Junction-to-Ambient thermal resistance (RθJA) for a TO-220 package without a heatsink is typically 62 °C/W.
- Temperature Rise: 7W × 62 °C/W = 434 °C rise above ambient.
Assuming a 25°C room, the junction temperature would theoretically hit 459°C. Since the maximum rated junction temperature (Tj) is 175°C, the part will destroy itself in seconds. You must add a heatsink. Let's calculate the maximum allowable thermal resistance for the heatsink to keep the junction under 150°C (leaving a 25°C safety margin) in a 40°C ambient environment (typical inside an enclosure).
- Max Allowed Delta T: 150°C (Target Tj) - 40°C (Ambient) = 110°C.
- Hot Rds(on) Adjustment: At 150°C, Rds(on) is roughly 2.2x higher. 0.0175 Ω × 2.2 = 0.0385 Ω.
- Hot Power Dissipation: 20² × 0.0385 = 15.4 Watts.
- Total Max RθJA: 110°C / 15.4W = 7.14 °C/W.
Subtracting the Junction-to-Case (1.5 °C/W) and Case-to-Sink (0.5 °C/W with thermal paste) resistances, your heatsink must have a thermal resistance of 5.14 °C/W or lower. This requires a substantial extruded aluminum heatsink, completely changing the physical footprint of your design.
Where You Meet This In Practice
You will run into thermal derating constraints whenever you switch high-current DC loads in confined spaces. Common scenarios include:
| Application | Typical Load | Thermal Trap | Better Component Choice |
|---|---|---|---|
| 3D Printer Heated Beds | 12V / 10A - 15A | Standard RAMPS boards use surface-mount FETs with poor PCB thermal vias, leading to melted connectors. | External solid-state relay (SSR) or a dedicated MOSFET module with an integrated heatsink. |
| E-Bike Motor Controllers | 48V / 30A+ | Enclosed in aluminum extrusions that trap heat; ambient inside the case easily reaches 70°C on hills. | Parallel arrays of low-Qg (gate charge) FETs like the IRFB3077 to spread the thermal load. |
| Automotive Auxiliary Lighting | 14.4V / 20A | Under-hood ambient temperatures routinely exceed 85°C in summer traffic, severely derating standard FETs. | Automotive-grade high-temp FETs (AEC-Q101 qualified) with high Tj max ratings (175°C+). |
For a deeper look into how gate drive voltage affects these thermal calculations, the All About Circuits guide on MOSFET selection provides excellent baseline theory on matching Vgs to your microcontroller's logic levels.
Real-World Scenario Walkthrough: The Melted Terminal Block
The Setup: A hobbyist is designing a 12V PWM dimmer for a 15A off-road LED light bar. They choose the IRLB8721, a popular logic-level MOSFET with an incredibly low Rds(on) of 8.7 mΩ at 4.5V Vgs. They solder it to a small perfboard, connect it to an Arduino Nano for PWM control, and seal the entire assembly inside a small, waterproof plastic project box mounted under the vehicle's dashboard.
The Numbers: At 15A and 25°C, the power dissipation is 15² × 0.0087 = 1.95W. Looking at a TO-220 free-air RθJA of 62 °C/W, the expected temperature rise is roughly 120°C. The hobbyist assumes 1.95W is 'low power' and skips the heatsink.
The Outcome: After 20 minutes of driving with the lights on, the MOSFET desoldered itself from the perfboard, failed short-circuit, and blew the vehicle's main 20A fuse. The plastic project box warped from the heat.
What Went Wrong: The hobbyist ignored the environment. The ambient temperature inside a sealed plastic box under a dashboard in the summer quickly rose to 60°C. Because the box was sealed, there was no convective airflow, effectively raising the thermal resistance far beyond the 62 °C/W free-air rating. As the junction temperature climbed past 100°C, the PTC effect kicked in, raising the Rds(on) and generating even more heat. The junction exceeded 175°C, the silicon avalanche broke down, and the part shorted drain-to-source. For more on how packaging affects heat transfer, Sparkfun's transistor application tutorial covers the physical limitations of through-hole versus surface-mount packages.
FAQ: Selecting Power Semiconductors
Q: Can I just parallel two identical MOSFETs to double my current capacity and halve the heat?
A: Yes, but with caveats. Because of the positive temperature coefficient, MOSFETs naturally current-share relatively well (if one gets hotter, its resistance rises, pushing current to the cooler one). However, you must ensure they share the same gate drive signal with identical trace lengths to prevent one from turning on slightly faster and taking the initial inrush current spike. Always use a separate gate resistor (e.g., 10Ω) for each paralleled FET to prevent high-frequency gate oscillation.
Q: Why do logic-level MOSFETs usually have higher Rds(on) than standard 10V-drive FETs?
A: Logic-level FETs (designed to fully turn on at 3.3V or 5V from an ESP32 or Arduino) require a thinner gate oxide layer and a denser, shallower cell structure. This physical geometry inherently increases the on-resistance compared to a standard FET optimized for a 10V-12V gate drive. If you have a 12V gate driver available, always choose a standard-threshold FET for lower conduction losses.
Q: How do I accurately measure junction temperature on the bench without an infrared camera?
A: You cannot easily measure the internal silicon junction directly. Instead, measure the case temperature using a thermocouple epoxied to the top of the TO-220 tab. Then, use the formula: Tj = Tc + (P × RθJC). If you measure a case temp of 90°C, and you know you are dissipating 10W, and the Junction-to-Case thermal resistance is 1.5 °C/W, your internal junction is at 90 + (10 × 1.5) = 105°C. This is the method professional engineers use to validate thermal models.
