Every watt the electronics dissipate has to leave the enclosure, or the air inside keeps heating until something fails or throttles. In a sealed box the heat has three routes: it conducts into the walls and out to the room, it moves by air circulating inside, and it radiates from the outside surface. For a low-power controller the shell alone often carries it. For a power supply, drive, or charging module it usually does not — and the moment you seal the box against dust and water, you lose the easy fix of just adding vents. So thermal design is a decision made together with the enclosure, the IP rating, and the way the parts are assembled, not a check done at the end.
This guide covers a quick way to estimate whether the shell is enough, the cooling options in order of complexity, the tension between airflow and IP rating, what the assembly process contributes, and what to put on your drawing.
Start with the number that matters: power dissipated inside
Thermal work begins with how many watts turn into heat inside the box, not with how big the enclosure is. Add up the dissipation of each heat source at its worst-case load — for a supply, that is the input power minus the output power delivered (so the loss, not the rating); for a processor or driver, the figure from its datasheet at your operating point. Then decide the maximum ambient temperature the product must survive and the maximum internal air temperature the parts allow. The gap between the two is the temperature rise you have to live within.
Two limits set that ceiling. Each component has a rated operating temperature on its datasheet, and the hottest part usually governs. Separately, a widely quoted rule of thumb is that every 10 °C rise roughly halves the life of an aluminium electrolytic capacitor; the exact figure is on the capacitor datasheet, but the direction is why running a box 15 °C cooler is not a cosmetic gain.
A quick estimate: can the shell carry it?
For a closed enclosure in still air, a first-pass estimate of the heat the walls can shed is:
P ≈ k × A × (T_inside − T_ambient)
where A is the exposed outer surface area in m², k is an effective heat-transfer coefficient in W/m²·K that lumps natural convection and radiation together, and the temperature difference is in kelvin. Enclosure suppliers publish sizing values for k by wall material — figures in the region of about 5–6 W/m²·K for painted sheet steel, higher for bare aluminium, and lower for plastic are commonly used. Treat those as planning numbers and confirm the value with your enclosure supplier’s data.
A worked illustration, using the formula and those planning values — this is a calculation, not a measured result:
| Item | Value |
|---|---|
| Enclosure 400 × 300 × 200 mm, all six faces exposed | A = 2 × (0.4·0.3 + 0.4·0.2 + 0.3·0.2) = 0.52 m² |
| Painted sheet steel, planning value | k ≈ 5.5 W/m²·K |
| Ambient / maximum internal air | 35 °C / 55 °C, so ΔT = 20 K |
| Heat the shell can shed | 5.5 × 0.52 × 20 ≈ 57 W |
If the electronics dissipate 30 W, the shell has margin on paper. If they dissipate 120 W, no amount of tidy layout closes the gap: the box needs a way to move heat out faster. Two cautions on the estimate. It assumes free air on every face, so a box mounted against a wall, in a rack, or inside another cabinet exposes less area and sheds less. And it treats the inside as one uniform temperature, while a hot spot on a power stage can be far above the average air temperature — which is why the component-level check and a physical test still matter.
Cooling options, from simplest to most involved
Work down this list only as far as the heat load forces you.
- Let the enclosure be the heat sink. Thermally bond the hot part to the wall so heat conducts out through the shell — a thermal pad, gap filler or grease between the component (or its heat spreader) and the enclosure wall, with a metal wall and, if needed, external fins. This is the standard route for sealed boxes because it keeps the IP rating intact. Aluminium extrusions and die-cast housings are chosen for exactly this reason; the material trade-off is covered in sheet metal vs injection molding for enclosures.
- Add an internal heat sink and stir the air. A finned sink on the part plus a small internal circulation fan spreads heat through the box air and onto the walls, avoiding hot spots. It adds a fan to the reliability list but does not open the box.
- Forced ventilation with filtered fans. Fan in, exhaust out, through filters. It moves far more heat than the shell can, but the box is now open to the air, so it generally cannot hold a high IP rating, and the filters need cleaning on a schedule.
- Heat exchangers and enclosure air conditioners. An air-to-air heat exchanger keeps the internal air separate from the external air and exchanges heat across a barrier, so the box stays sealed. An enclosure air conditioner goes further and can hold the inside below ambient. Both add size, cost, power draw and a maintenance item.
Passive first is the sound default: it has no moving parts to fail. Each step down the list buys capacity at the cost of complexity, so the useful question is the smallest option that covers the worst case with margin.
Where thermal design and IP rating collide
The two requirements pull in opposite directions. An IP65-or-better enclosure cannot have open vents, which rules out option 3 and pushes the design toward conduction to the wall or a sealed heat exchanger. A breathable vent membrane fitted to a sealed box equalises pressure so seals are not stressed as the air heats and cools, but it passes almost no heat, so it is a pressure fix rather than a cooling fix.
Vents are also shielding leak paths, so a box with an EMC requirement needs honeycomb or small-hole vent panels; see EMI/EMC shielding in box build enclosures.
This is why the IP rating and the heat load need to be settled together. If your product must survive washdown or outdoor rain, plan for conduction or a sealed exchanger from the start; do not design for a vented fan box and then ask for IP65. The rating itself, and why the whole assembled box has to hold it, is explained in the guide to IP ratings and ingress protection for enclosures.
Two more environmental loads are easy to miss. Outdoor enclosures gain heat from sunlight, so the ambient the box sees is higher than the air temperature — shading, a sunshield or a lighter finish is part of the thermal design. And altitude thins the air, which reduces convective cooling and matters if the product ships to high sites.
What the box-build process contributes
A thermal design on paper depends on things that happen on the assembly line. The parts of the build that most often decide whether the heat actually reaches the wall:
- Thermal interface materials. A pad that is the wrong thickness, or a grease layer applied unevenly, leaves an air gap that acts as insulation. The pad thickness, compression and placement have to be defined and repeatable.
- Fastener torque and sequence. The contact pressure between a part and its heat spreader or the wall depends on how it is fastened. A defined torque and tightening order keep that contact consistent from unit to unit.
- Airflow paths. Cable bundles routed across a fan inlet, or a bracket placed in front of an outlet, can undo a layout that simulated well. Cable routing is a thermal issue as well as a mechanical one, which is one reason the cable harness integration is planned with the layout rather than after it.
- Service access. Filters, fans and exchangers are wear items. If they cannot be reached without stripping the unit, they will not be maintained.
Reviewing these interfaces at the drawing stage is part of design for manufacturing on a box build, and the same review happens during enclosure integration, where the thermal path, the seals and the fastening are checked together.
Verifying it: measure, do not assume
A calculation and a simulation are estimates; the product still has to be tested. A practical thermal verification runs the finished, closed unit at the maximum specified ambient and worst-case load, with temperature sensors on the hottest components and on the inside air, and waits until the readings stop climbing before judging the result. Check the margin against each component’s datasheet limit rather than against the average. For products expected to see repeated hot and cold cycles in service, environmental stress screening exercises the assembly across temperature to expose weak interfaces, and the functional test plan for a box build covers how the unit is exercised while it is hot.
What to put on your drawing and RFQ
- Power dissipated inside the box, in watts, at worst-case load — per major heat source if you know it.
- Maximum ambient temperature and whether the unit sits in sun, in a rack, or against a wall.
- Maximum allowed internal air temperature and the component limit that sets it.
- The IP rating, because it decides whether vented cooling is even an option.
- Orientation and mounting, since natural convection depends on it.
- The cooling approach you intend — conduction to wall, heat sink, fan, exchanger — and the thermal-interface material and fastening torque if they are specified.
- Service and maintenance expectations, including whether filters or fans must be replaceable in the field.
Heat is a system property: it comes from the electronics, leaves through the enclosure, and depends on how the two are put together. State the watts, the ambient and the IP rating up front, choose the simplest cooling that covers the worst case, and verify it on the finished box. If you are working through a sealed power or control unit, our power and control systems work regularly hits this trade-off, and the same reasoning applies to EV charging station enclosures and energy storage enclosures, where the heat load is large and the enclosure is sealed.
For how heat, DIN rail layout and the 24 V DC supply interact in an industrial control enclosure, see box build for industrial automation.
