Types of Thermal Fuse: Thermal Path Math and Heatsink Sizing Guide

When a power semiconductor or transformer exceeds its thermal limits, silicon degrades, solder joints creep, and catastrophic failure follows. A thermal fuse is the ultimate fail-safe in your thermal management stack. Unlike standard overcurrent fuses that react to $I^2t$ heating, thermal fuses react directly to temperature. However, selecting the right device requires understanding the thermal path from the silicon junction to the ambient air, and how the fuse itself couples to that heat source.

The three primary types of thermal fuse used in electronics are one-shot thermal cutoffs (TCOs), bimetallic thermostats, and resettable positive temperature coefficient (PTC) thermistors. Each operates on a different physical principle and demands a specific mounting strategy to ensure it trips before the protected component dies.

The Core Types of Thermal Fuse and Selection Data

Choosing a thermal protection device is not just about picking a trip temperature. You must account for holding current, thermal mass, reset behavior, and the physical coupling method. The table below breaks down the most common types of thermal fuse with real-world part numbers and bench-tested parameters.

How the Mechanisms Actually Work

A one-shot TCO contains a eutectic thermal pellet that holds a spring-loaded contact closed. When the ambient temperature around the pellet reaches its precise melting point (e.g., 139°C), the pellet liquefies, the spring releases, and the contact permanently opens. You cannot reset it; it must be replaced. Bimetallic thermostats use a stamped disc of two bonded metals with different expansion coefficients. At the trip temperature, the disc mechanically inverts (snaps), breaking the electrical contacts. It snaps back when it cools. Polymer PTCs rely on a matrix of polymer and conductive carbon black. As the polymer heats and expands, it pulls the carbon particles apart, exponentially increasing the device's resistance from milliohms to kiloohms, effectively choking the current.

Thermal Path Math and Heatsink Sizing for Protected Devices

To select the correct thermal fuse, you must first know how hot the protected part actually gets. This requires calculating the junction-to-ambient thermal resistance ($R_{\theta JA}$). The fundamental thermal path equation is:

$T_J = T_A + (P_D \times R_{\theta JA})$

Where:
- $T_J$ = Junction temperature (°C)
- $T_A$ = Ambient temperature inside the enclosure (°C)
- $P_D$ = Power dissipated by the component (Watts)
- $R_{\theta JA}$ = Total thermal resistance from junction to ambient ($R_{\theta JC} + R_{\theta CS} + R_{\theta SA}$)

Worked Heatsink Selection Example

Let's protect an IRFP460 N-channel MOSFET dissipating 25W in a linear power supply. The datasheet specifies a maximum $T_J$ of 175°C, but for long-term reliability, we derate the target $T_J$ to 150°C. The ambient temperature inside our enclosure ($T_A$) is measured at 40°C.

1. Calculate Max Allowable $R_{\theta JA}$:
   $R_{\theta JA} = (T_J - T_A) / P_D = (150 - 40) / 25 = 4.4 °C/W$
2. Subtract Known Resistances:
   The IRFP460 $R_{\theta JC}$ (junction-to-case) is 0.75 °C/W.
   Using a Bergquist Sil-Pad 2000S thermal interface material, $R_{\theta CS}$ (case-to-sink) is roughly 0.5 °C/W.
3. Find Required Heatsink $R_{\theta SA}$:
   $R_{\theta SA} = 4.4 - 0.75 - 0.5 = 3.15 °C/W$

We need a heatsink with a thermal resistance of 3.15 °C/W or lower. A real-world part that fits this profile is the Aavid Thermalloy 530654B02500G, an extruded aluminum TO-247/TO-3P heatsink rated at roughly 2.5 °C/W in natural convection. At a cost of around $6.50 per unit, it provides a comfortable 0.65 °C/W margin. If we mount our SEFUSE SF240E (139°C TCO) directly to the MOSFET tab under this heatsink, the fuse will trip well before the junction hits the 150°C reliability limit, assuming the tab temperature tracks the junction within the calculated $\Delta T$.

Interpreting Derating Curves

A common mistake is assuming a fuse's rated current holds true at elevated temperatures. If you check the Littelfuse PTC datasheets, you will see a hold current ($I_{hold}$) derating curve. A 1206L050YR PTC rated to hold 0.5A at 20°C will only hold roughly 0.25A at 85°C. If your PCB ambient runs hot, you must select a PTC with a higher baseline $I_{hold}$ to prevent nuisance tripping during normal operation. Always read the derating curve, not just the front-page spec.

Enclosure Airflow, Thermal Limits, and Failure Signatures

How Hot is Too Hot?

Silicon dies can theoretically survive up to 175°C or even 200°C for wide-bandgap semiconductors like SiC. However, the packaging and surrounding passive components fail much earlier. According to Texas Instruments thermal design guidelines, keeping $T_J$ below 125°C is standard practice for commercial-grade ICs to prevent accelerated electromigration. Furthermore, standard SAC305 lead-free solder joints begin to exhibit significant creep and grain coarsening at sustained temperatures above 120°C. Electrolytic capacitors follow the Arrhenius equation: their operational lifespan halves for every 10°C increase above their rated temperature (usually 105°C). Therefore, a thermal fuse set between 110°C and 135°C is the sweet spot for protecting mixed-signal and power PCBs.

What Airflow and Enclosure Changes Buy You

If your thermal calculations show you need a massive, expensive heatsink, altering the enclosure airflow is a cheaper alternative. Thermal resistance in forced convection scales roughly with the inverse square root of air velocity.

• Natural Convection (0 LFM): Baseline $R_{\theta SA}$.
• Light Forced Air (100 LFM): A standard 40mm chassis fan moving air across the fins will drop the heatsink's $R_{\theta SA}$ by approximately 30% to 40%.
• High Velocity (400 LFM): Server-style blowers can drop $R_{\theta SA}$ by 60% or more compared to natural convection.

Adding a $5.00 40x40x10mm Sunon MagLev fan to our Aavid heatsink example would drop its effective $R_{\theta SA}$ from 2.5 °C/W to roughly 1.4 °C/W, allowing the MOSFET to safely dissipate nearly 40W instead of 25W without changing the physical footprint.

Failure Signatures of Thermal Stress

When a thermal fuse is absent or improperly coupled, components don't always fail with a dramatic pop. They exhibit insidious failure signatures that you can diagnose on the bench:

1. Parametric Drift (Thermal Runaway): In MOSFETs, as $T_J$ rises, the threshold voltage ($V_{GS(th)}$) drops, but the on-resistance ($R_{DS(on)}$) increases. If the gate drive is marginal, the increasing $R_{DS(on)}$ causes more $I^2R$ heating, creating a runaway loop. The part will test 'fine' with a multimeter when cold, but fail under load.
2. Solder Joint Cracking: Repeated thermal cycling (expansion and contraction) causes micro-cracks in BGA balls or through-hole leads. The signature is intermittent operation that 'fixes itself' when the board cools down, or fails only after the device has been running for 20 minutes.
3. Die Delamination: Extreme thermal shock can cause the silicon die to separate from the copper leadframe inside a TO-220 or D2PAK package. This causes a sudden, massive spike in $R_{\theta JC}$. The case feels cool to the touch, but the junction is silently burning up.

By correctly matching the types of thermal fuse to your specific thermal path, calculating the exact $R_{\theta}$ requirements, and sizing your heatsinks based on real wattage rather than guesswork, you ensure your power electronics survive the worst-case fault conditions without collateral damage to the PCB.