When Adding Insulation Makes an Assembly Worse

In each of these assemblies, some part of the depth is doing a job that is not thermal. A ventilated void carries moisture away. An eaves inlet feeds that void. A gap above rafter insulation is a drying route. A wall cavity gives water somewhere to run down and out. In each case the insulation being added occupies exactly that part.

The result is not a trade-off between two thermal outcomes but a substitution of one function for another, and the one being removed is usually a moisture function. That is why the symptom tends to appear as dampness rather than as a disappointing heating bill.

In a cold deck flat roof the insulation sits between or below the joists, and a clear air space is left between the top of the insulation and the underside of the deck. The record describes the competition directly: push the insulation up to the deck and the ventilation path disappears while the roof looks fully insulated; hold it down to preserve the path and the thermal layer is the thinner of the two claims on the depth.

What makes this case severe is that the deck above is impermeable, so everything that condenses in the void has to leave the way it came, as vapour in moving air. There is no drying route upwards at all, and the failure appears as moisture on the underside of a deck that nothing can dry, staining a ceiling from within rather than as water coming in from outside.

In a ventilated pitched roof the insulation has to continue right out over the wall head to meet the wall insulation, which is exactly where the ventilation opening also needs the space. The junction record puts the contradiction in one sentence: insulation must be continuous, and an air path must remain open, in the same few hundred millimetres.

Resolving both at once is what an eaves detail is for. The failure mode is ordinary and invisible — loft insulation pushed out to the wall as far as it will go, with the inlet closed behind it — and the roof still has its openings, still has its insulation and no longer has a working void.

The ventilated pitched roof record lists boarding out, recessed lights, an unsealed hatch, new cables through the ceiling line and an extract duct terminating in the space as ordinary domestic acts that each remove part of the arrangement. Boards laid on ceiling joists compress the insulation beneath them and cap the joist bays so air can no longer move freely across the space.

Read individually those look harmless. Read as a system they are the failure, which is why the record treats storage in a roof space as a change to the roof assembly rather than a use of an empty space.

In a rafter-line insulated roof, filling the rafter depth completely removes the vented space above the insulation, and the drying duty transfers to whatever is left: the vapour behaviour of the underlay and any counter-batten cavity above it. The record names the combination that removes every route at once — a closed underlay, no space beneath it and full-depth insulation — and states that these are not independent selections.

The counter-batten record makes the same point from the covering side. Where the underlay is laid over a continuous deck there is no drape, the batten bears on it along its whole length and each crossing becomes a small dam, so the counter-batten is what restores a drainage path and an air space that other parts of the build-up often rely on.

A wall cavity breaks the path water would otherwise follow across the wall and gives moisture reaching the back of the outer leaf somewhere to run down. The cavity wall record is careful about what filling changes: not that the void always loses its ability to drain, but that resistance to water crossing the wall passes from an open space to the fill's own behaviour.

That is why the retrofit record turns the question back onto the wall. A leaf that passes little water can be backed by almost anything; a leaf that wets through under driving rain depends entirely on the fill refusing to pass that water on, and on the fill not settling later to leave a gap down which water can run.

An internal lining does not occupy a void, but it follows the same rule from the other direction: it removes the wall's inward drying route while making the construction behind it colder. Moisture already reaching that construction now has less energy available to drive it back out.

The record adds a second-order effect worth knowing. The more the insulation cools the wall, the more it matters that internal moisture cannot reach it; but the tighter that layer is made, the more completely the wall is asked to dry outward alone. Where the outward route is already weak, the two requirements work directly against each other.

There is a related failure that is not about filling a space but about leaving one open. Insulation slows heat largely by holding air still, so where air can move against, behind or through the layer, heat travels with the air instead. A layer with an open face is not partly effective: along that face air moves freely and carries heat past the material entirely.

A bypass needs a path and a driver: a gap behind insulation connected to another gap elsewhere, so that a loop is formed, and a temperature difference across it. Neither gap is remarkable alone. The record notes that a loop can circulate entirely inside the construction, so a building can perform well under pressurisation testing and still have one.