Why still air is an excellent insulator

The reason an unventilated enclosure behaves like a flask.

Still air is an excellent insulator because air itself conducts heat poorly, and as long as it is not allowed to move in a loop it cannot make up for that poor conduction the way it can when it is free to circulate, which is exactly the condition inside a sealed, unventilated enclosure.

Moving air carries heat, still air resists it

Air is a poor conductor of heat on its own, worse than almost any solid an engineer would choose on purpose, but a moving body of air rarely relies on conduction alone. Warm air near a hot surface expands slightly, becomes less dense than the air around it, and rises, while cooler, denser air sinks in to take its place, and that continuous loop carries most of the heat away from a hot object sitting in open air.

The loop needs room to exist, a path upward for the warm air to leave along and a separate path for cooler air to arrive along. Inside a sealed enclosure with no vents, no fan and only a small enclosed volume, there is nowhere for such a loop to form. The air right next to a hot component warms up, and with nowhere to go and nothing arriving to displace it, it sits there, blocking further heat from leaving the surface behind it nearly as effectively as a deliberate layer of insulating foam.

Forty times worse than glass

Still air conducts heat about forty times less readily than window glass and nearly eight thousand times less readily than aluminium. That is why a layer of trapped air is so often chosen deliberately as an insulator, in a flask, a double-glazed window, a wall cavity or a winter coat. Most insulating materials, foam and fibreglass and down among them, are mainly a way of holding air still, with the solid part there only to stop the air from circulating.

The moment the same air is allowed to circulate freely, carried along by a fan or given an open path in and out of an enclosure, it becomes the efficient heat carrier that moving air is, shifting far more heat than the still version of the same gas could manage through conduction alone.

Bubble wrap and a sealed case

A sheet of bubble wrap insulates a fragile parcel remarkably well for something that is mostly air, and the reason lies in each little bubble being sealed off from its neighbours. Air trapped inside one bubble cannot rise out of it and be replaced by air pulled in from somewhere else, because the bubble has no opening for either to happen through. Heat crossing the sheet has to crawl across each sealed pocket by conduction alone, through air denied any chance to organise itself into a circulating current, and that is slow enough that the whole sheet behaves as an effective insulating layer despite being built almost entirely from air.

A sealed electronics enclosure works on precisely the same principle at a larger scale, one large sealed pocket in place of hundreds of tiny ones, with the same denial of any loop for the air to circulate through. The bubble wrap is keeping the cold out of the parcel, and the enclosure is keeping the heat in with its own components, but the physics doing both jobs is identical.

One gap is sometimes worse than none

An enclosure built to protect a circuit board from dust, moisture or knocks can end up holding warmth in the way a flask holds warmth in a drink, unless a deliberate path is designed for heat to leave by some route other than the still air inside the case. That route might be a pair of vents, one low and one high, that let a convective loop form, a fan that forces one, or a metal panel that conducts heat from the components straight to the outside of the case without asking the air inside to carry it there.

A design that half-solves the problem can make things worse. A single small gap cut into an otherwise sealed case lets warm air spill out without giving cooler air a separate path in, breaking the sealed enclosure's insulating advantage for dust and moisture without establishing a real circulating loop in its place. Making an enclosure bigger does little on its own either, since a bigger sealed volume still has no path for its air to loop and simply holds a larger, equally stagnant body of warm air around whatever is generating the heat.

Light loads and heavy loads

Still air is perfectly adequate when the job is resisting a small, steady trickle of heat. A sealed enclosure holding a component that produces very little heat can rely on still air and a metal case to shed that small amount by conduction and radiation alone, no vents or fans required, in the same way a flask copes with a small amount of heat loss over many hours.

Push the heat load up far enough, though, and the temperature difference needed to force that heat out through still air and a metal wall becomes larger than the components inside can safely tolerate. At that point no amount of clever enclosure shaping avoids the need to move air, or some other fluid, through the space. The insulating trick that keeps a lightly loaded enclosure comfortably cool is the same trick that will cook a heavily loaded one if nothing changes to let the extra heat out, and the designer's job is to know which of the two situations a given box is in before the lid is screwed down.