A sealed enclosure sheds heat only through its walls, so the internal temperature settles at ambient plus a rise set by the watt loss inside and the surface area outside. Estimate the loss honestly, work out the rise, and you will know whether the box survives, whether the electronics need derating, or whether the box needs to be bigger.
What a sealed box can and cannot do with heat
An enclosure with no vents and no active cooling has exactly one heat path: conduction through the walls, then convection and radiation from the outer surface. Nothing else moves the heat out. The air inside just circulates the energy around until the walls warm up enough to carry it away.
That sets a hard ceiling. A small sealed box might get rid of 10-20 W before the interior climbs 20 degrees C above ambient. The bigger the surface area for a given heat load, the lower the rise, which is why a shallow wide box cools better than a compact cube holding the same board.
Sealing also means no air exchange, so the box cannot dump humidity or vapours either. That trade-off is deliberate on washdown and submersible installs, and it is why the cooling decision inside is different from a vented enclosure: any vent path trades sealing for airflow, so decide it deliberately rather than by accident.
Where the watts actually come from
Start with the power budget, not the board size. Every watt that goes in and does not come out as useful work turns into heat. A linear regulator is the usual offender.
Take a real example: a board fed 12 V at 5 A, delivering 3.3 V at 2 A. Input power is 60 W. Useful output is 6.6 W. The difference, roughly 53 W, becomes heat inside the box. Nothing about the enclosure changes that number; only the circuit design does.
The same load on a switching converter runs far cooler. A switcher at 85-90% efficiency losing 15% of a 6.6 W output dissipates around 1 W. That single part choice moves the box from thermally impossible to trivial.
Count everything that dissipates: regulators, drivers, relays, the transformer of a doorbell-style supply, even long runs of wiring. Then add a margin, because nameplate ratings and reality rarely match.
Estimating the temperature rise
For a sealed box in still air, a workable estimate uses the exposed surface area and the allowable rise above ambient. As a rule of thumb, a metal enclosure in free air dissipates roughly 1-2 W per square foot of surface per degree C of rise, with the lower end typical for painted or brushed finishes indoors.
Worked through: a stainless wall-mount box around 16 x 12 x 8 in has a surface area of roughly 8 square feet. At 1.5 W per square foot per degree C, it carries about 12 W per degree C of rise. A 53 W loss inside that box pushes the interior roughly 4-5 degrees C above ambient, which most electronics shrug off. The same 53 W in a compact 8 x 6 x 4 in box, with maybe 2.5 square feet of surface, produces a rise well over 15 degrees C on top of the hottest day the site sees.
Two corrections matter. First, temperature-sensitive parts sit at the hot spots, next to the regulator, not at the box average. Second, in direct sun the box ambient is not the air temperature; solar gain can add tens of degrees, and a sun shield reduces that before it reaches the walls. The outdoor electrical enclosure guide covers siting and shielding in more detail.
What the heat does to the electronics
Semiconductors tolerate heat up to a point and then fail faster as it climbs. Silicon junctions are usually rated to 125 degrees C or 150 degrees C, but the part's life at the limit is short. Every sustained 10 degrees C reduction roughly doubles the life of many electronic components, an Arrhenius rule that holds well enough for planning.
The parts that complain first are electrolytic capacitors, whose lifetime approximately halves for each 10 degrees C above their rating, and any regulator with a thermal shutdown. A switching regulator that hits its shutdown limit does not fail dramatically; it cycles, the output dips, and the downstream logic browns out in ways that look like software bugs.
Precision analog circuits drift with temperature too. Reference voltages and op-amp offsets move by microvolts or millivolts per degree, so a box that cycles 30 degrees C between night and day produces readings that wander, even though nothing has failed.
How to size the box for the heat
Treat the heat budget as an input to sizing alongside the space budget. The guide to sizing an electrical enclosure walks through the full method; the thermal steps are these.
- Add up the heat loss in watts at worst case, including regulator losses, relay coils and anything else that warms up.
- Find the highest ambient the enclosure will see, and add solar gain if it sits in the sun.
- Pick a maximum internal temperature the most sensitive part can live with, minus a margin of 10-15 degrees C.
- Divide the heat load by the allowable rise and by a transfer figure of 1-2 W per square foot per degree C to get the surface area the box needs.
- Choose the smallest standard size that meets that area with room to spare, or specify a custom one.
If the required area is unrealistic, the fix is inside the box, not around it: replace the linear regulator with a switcher, move a hot supply outside the sealed volume, or keep hot parts away from electrolytics and thermal interfaces on the board.
If nothing brings the load low enough, active cooling that keeps the seal intact, such as a heat pipe or air conditioner, is the next step. For many low-power boards the honest answer is simpler: a bigger box costs little and solves the problem outright.
What to check before you commit
Before ordering, confirm the arithmetic against reality. Power the assembly in free air at the hottest expected ambient and measure the board temperature near the regulator. Then measure again inside the chosen box; the difference tells you the real wall penalty, not the estimate.
Check the ratings match the job. Our stainless wall-mount enclosures are rated NEMA 4X / IP66 by the manufacturer; not UL listed, and for prolonged submersion the spec calls for a Type 6P design rather than a standard sealed box.
The enclosure selector filters by environment, size and inside layout, and carries the requirements to a custom enclosure quote when no standard size fits. A custom box lets you set the depth and proportions to maximize surface area for the same volume, and it is delivered in 10 business days from drawing approval. For the spec side, the NEMA 4X vs IP66 guide explains what those ratings do and do not promise, and the IP rating chart shows how the codes compare.
If the real question is whether to vent instead of seal, work that out before you order, because the rating and the cooling strategy have to be decided together. Once the heat load, the ambient and the required rating are on paper, the box specification usually writes itself.
Questions
How much heat can a sealed enclosure get rid of?
Only what conduction through the walls and convection and radiation from the outer surface can carry. A rough rule is a few watts per square foot of exposed surface per 10 degrees C of rise above ambient, depending on material and finish. A small sealed IP68 box might manage 10-20 W for a 20 degrees C rise; a large stainless box can dissipate considerably more because its surface area grows faster than its volume.
Does a regulator on a PCB count as heat even when the load is small?
Yes. A linear regulator dissipates the difference between input and output power. Dropping 12 V to 3.3 V at 2 A burns off about 17 W regardless of how small the board is. A switching converter loses far less, typically 10-20% of the output power, which is why it is the usual choice inside a sealed box.
Can I just drill a vent hole to cool a sealed box?
A vent breaks the sealing you chose the box for. Once there is a hole, water, dust and insects have a path in, and the box no longer meets its rating. If the electronics genuinely need airflow, use a breather vent and accept the rating change deliberately, or move to a larger box or cooler components instead.
Do electrolytic capacitors really fail faster when the box runs hot?
Yes. A common rule of thumb is that electrolytic capacitor life roughly halves for every 10 degrees C rise above the capacitor's rated temperature. Electrolytics are usually the shortest-lived parts on a board already, so sustained high internal temperature often shows up first as dried-out caps and drifting regulator output.
