PCB Thermal Management: Copper, Thermal Vias, and Heatsinks

Why Heat Is the Enemy of Electronics

PCB thermal management diagram showing a QFN chip over a thermal via array carrying heat into a copper pour and ground plane, with a finned heatsink and TIM layer above the part

Imagine you build a board that runs perfectly on the bench, then a week later inside a sealed enclosure it starts failing at random. The usual culprit is neglected thermal management. Every component that consumes power turns some of it into heat, and if that heat finds no path out it drives up the junction temperature (Tj) inside the silicon itself — the hottest and most critical point in any semiconductor.

Heat shortens life dramatically. The approximate Arrhenius rule states that roughly every +10°C rise in temperature can halve a component's expected life. An electrolytic capacitor rated 10,000 hours at 45°C may survive only a fraction of that at 85°C.

Every part has a maximum Tj (often 125°C or 150°C). The closer you get to it:

The smaller your safety margin and the more intermittent failures appear.
You must derate the current — run the part below its nominal rating.
Failure mechanisms accelerate: diffusion, solder-joint cracking, capacitor dry-out.

Golden rule: design so the hottest point on the board stays at least 20°C below its rated maximum. Thermal margin is an investment in reliability.

Heat-Transfer Paths

Heat leaves a board through three physical paths. Knowing them tells you where to spend your design effort.

Path	Mechanism	Importance on a PCB
Conduction	Transfer through solids (copper, vias, solder)	Dominant path inside the board
Convection	Moving air carries heat away	Key for heatsinks and ventilated enclosures
Radiation	Thermal emission from hot surfaces	Usually weak at low temperatures

On a bare PCB, conduction is the real hero: heat moves from the chip body into its pads, into the connected copper, then through vias to inner and bottom layers. Copper conducts heat hundreds of times better than the insulating FR-4 resin.

Convection only works once heat reaches a large surface (a copper area or heatsink) in contact with air. That is why conducting heat from the hot spot to a wide area is always the first step.

Copper as a Heatsink

The simplest heatsink available to you is the copper already on your board. A copper pour connected to the hot component's pad spreads heat over a wide patch, increasing the contact area with the air.

Core techniques:

Large copper area: connect the thermal pad to the biggest copper region you can on that layer, then carry it to other layers through vias.
Thicker copper: standard copper is 1oz (about 35µm). Using 2oz (about 70µm) roughly doubles the cross-section, improving both heat spreading and current capacity.
Symmetric distribution: avoid choking heat into a narrow corner; let copper extend in all four directions from the part.

Do not isolate the thermal pad with narrow thermal-relief spokes the way you would a normal solder pad. A power pad needs a solid copper connection to move heat freely.

A copper area of a few square centimetres on both layers can lower θJA noticeably compared with a small, isolated pad.

Thermal Vias

When the chip is a QFN or DPAK with a bottom thermal pad, the question becomes: how do we move heat from under the body to the other layers? The answer is thermal vias — an array of plated holes piercing the thermal pad down to a ground plane or copper layer beneath.

Practical guidance:

Regular array: arrange vias in an even grid under the pad, with typical spacing of 0.5mm to 1.2mm.
Size: a common small hole is 0.3mm drill with a 0.6mm pad; many small vias beat a few large ones.
Count: more vias means lower thermal resistance — a 3×3 or 4×4 array is common under power chips.

Type	Description	Effect
Filled	Plugged with conductive resin or copper and capped	Best heat transfer, blocks solder wicking
Tented	Closed with solder mask, not filled	Cheaper, but solder can escape through it

Open vias under a QFN pad can wick away solder during reflow and create a poor joint. Specify filled-and-capped via-in-pad (filled & capped) for critical pads.

Thermal Placement and Airflow

Thermal design starts with placement, before you route a single trace. Spread the heat instead of stacking it.

Space out hot parts: do not cram the voltage regulator, the power stage, and the processor into one corner; spread them so the board can breathe.
Protect the sensitive from the hot: keep precision sensors, voltage references, and electrolytic capacitors away from heat sources — heat dries a capacitor out and kills it early.
Airflow direction: if a fan exists, arrange hot parts along the air path so one hot part does not preheat the one downstream of it.
Sealed enclosure: inside an unventilated case the ambient temperature (Tambient) climbs significantly. Calculate using the in-box temperature, not room temperature.

Place hot parts near a board edge or near ventilation openings, and keep an open copper path that carries their heat outward.

Heatsinks, TIM, and MCPCB

When dissipation exceeds what copper alone can handle, you turn to dedicated solutions.

Heatsink: a finned aluminium block that hugely increases the surface area exposed to air. It mounts on top of the hot part.
Thermal interface material (TIM): paste or pad placed between the component and the heatsink to expel insulating air from the tiny gaps. Without it, a heatsink loses most of its benefit.
Metal-core PCB (MCPCB): a board with an aluminium base instead of FR-4, conducting heat from the part straight into the metal base. Ideal for high-power LEDs and power stages.

Solution	When to use it
Copper + vias	Low-to-medium power (up to a few watts)
Heatsink + TIM	High-power parts on a normal board
Aluminium MCPCB	High-power LED lighting, continuous power currents

Never forget the TIM. A heatsink clamped onto dry metal (with an air gap) can perform worse than no heatsink at all.

Estimating Temperatures

To estimate Tj you need two numbers from the datasheet: the thermal resistances.

θJA (junction-to-ambient): resistance from the junction to surrounding air, in °C/W. Depends on the board and ventilation.
θJC (junction-to-case): resistance from the junction to the component case — used when fitting a heatsink.

The basic equation:

Tj = Tambient + P × θ

Worked example: a regulator dissipating P = 1.5W, with θJA = 50°C/W, in an enclosure whose internal temperature is Tambient = 40°C:

Tj = 40 + 1.5 × 50 = 115°C

That is dangerously close to the 125°C limit — a thin margin. Fixes: add copper area and vias (lowering θJA), add a heatsink (moving to the θJC path), or reduce the dissipated power.

When heat sources multiply or the enclosure is complex, the simplified equation cannot give an accurate picture. Then reach for finite-element thermal simulation, or measure real temperatures with a thermal camera on a prototype.

Summary

Thermal management is not a luxury but a condition for board reliability. Remember the practical rules:

Heat shortens life (+10°C ≈ half the life) and pushes Tj toward its limit — leave margin.
Conduction through copper and vias is the dominant path inside the board.
Use large copper pours and 2oz copper as a free heatsink, with a solid (not relieved) connection.
Place a filled thermal-via array under QFN/DPAK thermal pads.
Space hot parts, protect sensitive ones, and respect airflow and enclosure temperature.
For high power: heatsink + TIM, or an aluminium MCPCB.
Estimate Tj = Tambient + P×θ from the datasheet, and simulate thermally when in doubt.