When designing power electronics, the term resistance temperature formula actually points to two coupled physical realities. The first is the electrical resistance temperature coefficient, which dictates how much a conductor's resistance increases as it heats up. The second is the thermal resistance formula ($R_{\theta}$), which models how effectively heat flows from a silicon junction to the ambient air. Ignoring the interplay between these two formulas is the primary cause of thermal runaway in DIY power supplies and motor drivers.
In this guide, we will bridge the gap between trace-level electrical heating and system-level thermal management. You will learn how to calculate exact junction temperatures, interpret datasheet derating curves, and select physical heatsinks based on hard wattage data rather than guesswork.
The Core Math: Thermal Resistance and the Temperature Coefficient
To manage heat, you must first quantify the thermal path. In electronics, thermal resistance ($R_{\theta}$) is measured in degrees Celsius per Watt (°C/W). It functions exactly like electrical resistance in Ohm's Law, but instead of voltage driving current, a temperature difference ($\Delta T$) drives heat flow (Power, $P$):
$R_{\theta} = \Delta T / P$ or $T_{Junction} = T_{Ambient} + (P \times R_{\theta JA})$
The total junction-to-ambient thermal resistance ($R_{\theta JA}$) is the sum of the series resistances in the heat path: the junction-to-case ($R_{\theta JC}$), the case-to-sink interface ($R_{\theta CS}$), and the sink-to-ambient ($R_{\theta SA}$).
The Electrical Resistance Temperature Feedback Loop
While $R_{\theta}$ manages the heat exiting the part, the electrical resistance temperature formula governs the heat being generated inside your PCB traces and bond wires. For copper, the formula is:
$R(T) = R_0 [1 + \alpha (T - T_0)]$
Where $\alpha$ (the temperature coefficient of resistance for copper) is approximately 0.00393 /°C. If you design a power trace with 10 m$\Omega$ of resistance at 20°C, and poor thermal management allows that trace to reach 100°C, the resistance becomes:
$R(100) = 0.010 \times [1 + 0.00393 \times (80)] = 13.14$ m$\Omega$.
That is a 31.4% increase in resistance. Because resistive heating is calculated as $I^2R$, your trace is now generating 31.4% more heat at the exact same current. This positive feedback loop is the definition of thermal runaway.
Reference Table: Typical Thermal Resistance Values
Before you can calculate a heatsink requirement, you need the baseline $R_{\theta}$ values for your component package and interface materials. The table below provides real-world data based on JEDEC and ASTM testing standards.
| Component / Interface | Typical $R_{\theta}$ (°C/W) | Standard | Engineering Notes |
|---|---|---|---|
| TO-220 (Junction-to-Case, $R_{\theta JC}$) | 1.5 - 5.0 | JEDEC JESD51-14 | Highly dependent on silicon die size and leadframe copper mass. |
| TO-247 (Junction-to-Case, $R_{\theta JC}$) | 0.4 - 1.2 | JEDEC JESD51-14 | Standard for high-power MOSFETs and IGBTs; massive copper tab. |
| D2PAK (Junction-to-Board, $R_{\theta JB}$) | 15.0 - 25.0 | JEDEC JESD51-8 | Relies entirely on PCB copper pour area; no traditional heatsink tab. |
| Arctic Silver 5 (Thermal Paste) | ~0.13 (at 0.001" BLT) | ASTM D5470 | BLT = Bond Line Thickness. Requires high mounting pressure for best results. |
| Bergquist Sil-Pad 900 VHM | ~1.8 (at 50 psi) | ASTM D5470 | Electrically isolating. Thicker BLT than paste, but prevents shorting to the sink. |
Heatsink Selection and Derating Curves in Practice
Let's apply the thermal path math to a concrete scenario. Suppose you are building a linear power supply using an LM317T in a TO-220 package. The circuit drops 12V to 5V at 1A. The power dissipated by the regulator is $P = (12V - 5V) \times 1A = 7W$.
The LM317 datasheet specifies a maximum junction temperature ($T_{J(max)}$) of 125°C. If your workshop ambient temperature ($T_A$) is 35°C, your maximum allowable temperature rise ($\Delta T$) is $125°C - 35°C = 90°C$.
Calculating the Required $R_{\theta SA}$
Using the thermal resistance formula, the maximum total thermal resistance allowed is:
$R_{\theta JA(max)} = 90°C / 7W = 12.8°C/W$.
Now, we subtract the known internal resistances. From the table above, the LM317 TO-220 $R_{\theta JC}$ is typically 4.0°C/W. If we use a Bergquist Sil-Pad for electrical isolation, $R_{\theta CS}$ is roughly 1.8°C/W. The required sink-to-ambient resistance is:
$R_{\theta SA} = R_{\theta JA(max)} - R_{\theta JC} - R_{\theta CS}$
$R_{\theta SA} = 12.8 - 4.0 - 1.8 = 7.0°C/W}$.
You must select a heatsink rated for 7.0°C/W or lower in natural convection. A suitable off-the-shelf part is the Wakefield-Vette 637K (rated ~6.5°C/W) or the Aavid Thermalloy 577102B04000G (rated ~8.5°C/W, which would require slight forced air). Picking a random 'TO-220 heatsink' from a bin without checking the °C/W rating at your specific wattage is how you end up with a melted solder joint.
Interpreting Datasheet Derating Curves
Datasheets rarely give you a single wattage number; they provide a derating curve. This curve shows maximum allowable power dissipation on the Y-axis and case/ambient temperature on the X-axis.
- The Flat Region: Up to 25°C, the part can handle its absolute maximum rated power (e.g., 50W for a TO-247 MOSFET).
- The Slope: Above 25°C, the line slopes downward to zero watts at $T_{J(max)}$ (e.g., 150°C). The slope represents the derating factor. For a 50W part with a 125°C span (150 - 25), the derating factor is $50 / 125 = 0.4 W/°C$.
- The Calculation: If your heatsink maintains the case at 80°C, you must derate the part: $50W - (0.4 W/°C \times (80°C - 25°C)) = 28W$ maximum allowable dissipation.
How Hot is Too Hot? The Reliability Limit
Silicon survives up to 150°C or 175°C, but operating at the absolute maximum $T_{J}$ is a guarantee of premature death. According to the TI Application Report SNVA419C on thermal management, component reliability follows the Arrhenius equation. As a practical rule of thumb derived from MIL-HDBK-217F, the operational lifespan of a semiconductor halves for every 10°C increase in junction temperature above 85°C.
If you want a 10-year lifespan from your power electronics, design your thermal path to keep $T_J$ below 100°C under worst-case ambient and maximum load conditions. 125°C is the 'cliff edge' where the silicon shuts down via internal thermal protection, but 100°C is the engineering target for reliability.
Enclosure Airflow, Failure Signatures, and System Changes
Once you have sized the heatsink for open-air bench testing, you must account for the real-world environment. Enclosures and airflow drastically alter the $R_{\theta SA}$ value you calculated.
What Airflow and Enclosure Changes Buy You
Natural convection relies on the chimney effect. When you place a calculated heatsink inside a sealed NEMA enclosure, the ambient air inside the box heats up, effectively raising your $T_A$ and destroying your $\Delta T$ headroom. A sealed plastic enclosure can easily add 20°C to 30°C to the internal ambient temperature.
Introducing forced convection (a fan) changes the thermal dynamics entirely. Here is how airflow impacts a standard extruded aluminum heatsink:
| Cooling Method | Airflow (LFM) | Effective $R_{\theta SA}$ | Use Case |
|---|---|---|---|
| Natural Convection | 0 LFM | 10.0 °C/W | Passive audio amps, low-power LED drivers. |
| Low Forced Air (40mm fan) | 100 LFM | 6.5 °C/W | Desktop PC VRMs, enclosed 3D printer boards. |
| High Forced Air (80mm fan) | 400 LFM | 3.2 °C/W | Server power supplies, high-current motor controllers. |
Note: LFM = Linear Feet per Minute. Data based on standard 2-inch extruded aluminum fin profiles.
Adding a modest 40mm fan (like a Noctua NF-A4x20) pushing 100 LFM across your Wakefield-Vette heatsink buys you a 35% reduction in thermal resistance, allowing you to use a smaller, cheaper, and lighter physical heatsink.
Failure Signatures of Thermal Stress
When the resistance temperature formula works against you and thermal management fails, components do not always let out 'magic smoke' immediately. They degrade through specific physical mechanisms:
- Electromigration: At high current densities and temperatures above 100°C, metal atoms in the silicon's internal interconnects literally migrate in the direction of electron flow. This creates voids (leading to open circuits) and hillocks (leading to short circuits). It is the primary cause of death in overclocked CPUs and overstressed linear regulators.
- Thermal Cycling Fatigue (CTE Mismatch): Every material has a Coefficient of Thermal Expansion (CTE). Silicon, copper leadframes, and FR4 fiberglass all expand at different rates. If your thermal design allows the part to swing from 40°C (idle) to 110°C (load) repeatedly, the mechanical shear stress will crack the BGA solder balls or snap the wire bonds inside the IC package.
- Thermal Runaway (The Positive Feedback Loop): As discussed in the electrical resistance formula section, if a PCB trace or a bipolar junction transistor (BJT) heats up, its resistance or current draw increases. If the heatsink cannot dissipate this new, higher wattage, the temperature spikes exponentially until the silicon melts or the bond wires fuse open.
For a deeper dive into how PCB layout affects these thermal pathways, the All About Circuits heatsink selection guide provides excellent visual breakdowns of thermal pad via stitching.
Mastering thermal management requires treating heat flow with the same mathematical rigor as current flow. By calculating your $R_{\theta}$ path, respecting the electrical resistance temperature coefficient, and designing for a 100°C junction limit, you ensure your electronics survive long past their warranty period.
