Heat Sink Calculator helps estimate junction temperature, power at a target T_J, and required heat sink resistance from ambient conditions and thermal path values for design checks
Junction Temp (T_J)
110.00 °C
The maximum predicted temperature at the silicon junction under steady-state load.
Thermal Resistance Breakdown
1.70 °C/W (Total)
- Internal Path (Rcase)
- 29.41 %
- External Path (R1 + R2)
- 70.59 %
Total system resistance mapped to internal component vs. external cooling solution pathways.
Thermal Gradients
85.00 °C (Total ΔT)
- Rise Across Case
- 25.00 °C
- Rise Across External
- 60.00 °C
The exact temperature rise expected across each major section of the thermal series circuit.
Critical Surfaces
85.00 °C (Case Tc)
- Heat Sink Base (Ts)
- 75.00 °C
- Sink to Ambient Diff.
- 50.00 °C
Predicted temperatures directly on the outer casing of the component and the mating sink interface.
Limits & Derating
40.00 °C (Max Ambient)
- Max Power @ T_J,max
- 58.82 W
- Thermal Margin
- 15.00 °C
Thermal limit checks use the editable device junction limit entered above.
The Thermal Resistance Chain: From Silicon to Air
Heat moves through a component the same way current moves through a series circuit. Power takes the place of current, temperature difference takes the place of voltage, and each material the heat crosses — die, case, interface, heat sink, air — acts like a resistor. Stack those resistors end to end and the total temperature rise from ambient air to the silicon junction is just power times total resistance.
The calculator above is built on one equation, rearranged three different ways depending on what you already know:
$$T_J = P \times (R_{JC} + R_1 + R_2) + T_A$$
Here $P$ is the power the device dissipates in watts, $T_A$ is ambient air temperature, and $R_{JC}$, $R_1$, and $R_2$ are the three resistances in the chain: junction-to-case (fixed by the component’s die and package), case-to-sink (the mounting interface — thermal pad, grease, or a dry joint), and sink-to-ambient (the heat sink itself, plus whatever airflow is moving across it). $T_J$ is the resulting junction temperature.
Solve that same equation for $P$ instead of $T_J$ and you get the wattage a device can dissipate before it hits a target junction temperature:
$$P = \frac{T_J – T_A}{R_{JC} + R_1 + R_2}$$
And solve it for the sink-to-ambient term and you get the maximum heat sink resistance a design can tolerate for a given power and junction limit:
$$\theta_{SA,req} = \frac{T_{J,max} – T_A}{P} – R_{JC} – R_1$$
Three unknowns, one relationship. Which version runs depends only on which two of $T_J$, $P$, and $\theta_{SA}$ you’re trying to find.
Sizing a Heat Sink for a 50 W TO-220 Regulator
Take a device dissipating 50 W, sitting in 25°C ambient air, rated for a 125°C maximum junction temperature. Its datasheet lists a junction-to-case resistance of 0.5°C/W. It’s mounted with a thermal pad rated at 0.2°C/W, bolted to a heat sink rated at 1.0°C/W in the airflow it will actually see.
Add the three resistances: 0.5 + 0.2 + 1.0 = 1.7°C/W total. Multiply by power and add ambient: 50 \times 1.7 + 25 = 110°C. That’s the predicted junction temperature — 15°C below the 125°C limit, which is the thermal margin.
Walking the same numbers down the chain shows where each degree gets spent:
| Point in the chain | Temperature | Resistance to next point | Drop across it |
|---|---|---|---|
| Junction (T_J) | 110.00°C | R_JC = 0.50°C/W | 25.00°C |
| Case (T_C) | 85.00°C | R1 = 0.20°C/W | 10.00°C |
| Sink (T_S) | 75.00°C | R2 = 1.00°C/W | 50.00°C |
| Ambient (T_A) | 25.00°C | — | — |
Notice that the heat sink itself accounts for 50 of the 85 total degrees. The internal path — junction to case — only accounts for 25. That split, roughly 29% internal to 71% external in this example, is typical of a well-cooled design: the component’s own construction is fixed, so almost all of the engineering happens in the mounting and the sink.
Why 125°C on the Datasheet Isn’t a Suggestion
The moment predicted $T_J$ climbs past the maximum junction temperature you entered, the calculator switches to a Thermal Limit Exceeded warning. There’s no cushion built into that check and no rounding. If the math says 126.4°C against a 125°C limit, that’s a failed design, not a close call.
That number isn’t arbitrary and it isn’t universal either — it’s whatever your specific component’s datasheet lists as its rated maximum, which is why the field is editable rather than fixed. Silicon devices commonly land somewhere between 125°C and 175°C depending on the die and package, so plug in the actual figure from your part’s datasheet, not the calculator’s 125°C default.
What makes it a hard limit rather than a soft one is what happens on the other side of it. Push a junction past its rated maximum and you’re not looking at reduced efficiency the way you might with a long, thin wire.
You’re looking at accelerated wear-out, unpredictable thermal shutdown behavior, or outright failure, depending on the part. The margin the calculator reports — the gap between predicted $T_J$ and your entered limit — is the number worth watching, not just the pass or fail flag.
What Actually Moves Junction Temperature
Power dissipation and total resistance scale the result linearly, so doubling either one doubles the temperature rise above ambient. But the three resistance terms are rarely equal contributors, and knowing which one to attack matters more than the raw math.
$R_{JC}$ is baked into the component. You can’t change it without swapping parts. $R_1$, the case-to-sink interface, is often the cheapest lever — a better thermal pad or properly applied grease can shave tenths of a degree per watt off that term for a few dollars.
$R_2$, the sink-to-ambient path, is usually the biggest number in the stack and the one most sensitive to airflow: a heat sink rated at 1.0°C/W in still air might drop to a fraction of that under forced convection, which is why fan placement changes a thermal budget more than almost any other single decision.
Ambient temperature shifts the whole curve up or down without changing any resistance at all — a device that runs fine at a 25°C bench ambient can run hot inside a sealed enclosure sitting at 45°C, even with identical hardware.
A few edge cases are worth knowing about. Resistance values are allowed to be zero — a mounting with no interface material at all, for instance — but power has to be strictly positive; zero or negative wattage isn’t a valid thermal load.
Asking the calculator to solve for power at a target junction temperature that’s below ambient returns a physics error, since heat won’t flow from a colder junction into a warmer surrounding without violating the second law.
And solving for the required heat sink resistance when the internal path alone already consumes the entire junction-to-ambient budget returns a not-feasible result — no heat sink, however good, fixes a design where the component and its mounting are the problem.
Common Questions About Heat Sink Thermal Resistance
What is thermal resistance in a heat sink calculation?
It’s a measure, in °C per watt, of how much a material or interface resists the flow of heat. A lower number means a given wattage produces a smaller temperature rise across that material, which is why heat sink datasheets advertise their resistance value as the headline spec.
How do I calculate the heat sink size I need for a given wattage?
Start from your component’s junction-to-case resistance and your mounting interface resistance, subtract both from the total budget your power and temperature limit allow, and whatever’s left is the maximum sink-to-ambient resistance you can afford. Any heat sink rated at or below that number, in the airflow you’ll actually have, will keep you under your junction limit.
What’s the difference between junction-to-case, case-to-sink, and sink-to-ambient resistance?
Junction-to-case is fixed by the component’s internal construction and comes straight off its datasheet. Case-to-sink depends on the mounting interface — pad, grease, or bare metal contact. Sink-to-ambient depends on the heat sink itself and the air moving across it, and it’s usually the term with the most room to improve.
Why does junction temperature rise when ambient temperature rises?
Junction temperature is ambient temperature plus a fixed rise caused by power flowing through resistance. Since the rise itself doesn’t change with ambient, any increase in ambient air temperature adds directly, degree for degree, onto the junction temperature.
What happens if a component runs above its maximum junction temperature?
Depending on the part, you’ll see accelerated long-term wear, erratic thermal shutdown, or immediate damage. It isn’t a graceful degradation the way a slightly undersized wire is — the rated maximum exists because the manufacturer tested reliability up to that point and not beyond it.
Does thermal paste or a thermal pad actually change the result?
Yes — that interface is exactly the case-to-sink resistance term in the equation. A poorly applied or dried-out interface can add several tenths of a degree per watt, which on a high-power device translates directly into extra junction temperature.
How much does airflow reduce the resistance I need from a heat sink?
Substantially. Moving from natural convection to even modest forced airflow commonly cuts a heat sink’s effective resistance by half or more, since forced air continuously replaces the warm boundary layer clinging to the fins. That’s why two heat sinks with identical published resistance ratings can behave very differently once one of them sits in front of a fan.
Can this kind of calculation be used for pulsed or intermittent loads?
Not directly. A steady-state resistance model assumes constant power dissipated over a long enough time that temperatures stop changing, which understates peak junction temperature for short, high-power pulses. Pulsed loads need a transient thermal impedance curve instead of a single resistance figure.
Steady-State, Not Transient
This calculation assumes constant power and enough time for temperatures to settle. It doesn’t account for thermal mass, duty-cycled or pulsed loads, multiple heat sources sharing one sink, or non-uniform heat spreading across the sink’s surface — all of which can push real peak junction temperature above what a steady-state number predicts.