Thermal Bridging or Thermal Bypass: Telling Two Problems Apart

Bridging is conduction. Where an insulating layer is interrupted, by a structural member passing through it, by a fixing, by a frame set out of line with it, by a slab running to the outer face, or simply by the geometry of a corner, the interruption carries a share of the flow out of all proportion to the area it occupies. The record calls that disproportion the whole point.

A bypass is movement. Insulation slows heat largely by holding air still, and where air can move against, behind or through the layer, that stillness is lost locally and heat travels with the air rather than through the material. The record's key sentence is that the layer can be complete, undamaged and exactly what was specified while behaving as though a part of it were absent.

The bridging record puts it as a test anyone can apply: if closing gaps and stopping air moving around the insulation resolves the problem, it was a bypass, and bridging is conduction through solid material which remains after every gap has been closed. The bypass record puts the mirror image: the practical test is whether adding an enclosing lining or a perimeter closure changes anything at all.

That is a diagnostic order rather than a repair instruction. It says which question to ask first, and it says that the answer to the cheaper question determines whether the expensive one needs asking. It does not say what to do about either, because what either means for a specific building is a professional assessment.

The record describes what has to be true for a bypass to exist, and it is useful because it tells you where to look. The path is usually a gap behind the insulation connected to another gap elsewhere, at a wall head, a floor perimeter, a party junction, an eaves line, a duct route or a partition head, so that a loop is formed. The driver is a temperature difference across that loop.

The consequence is that neither gap is remarkable alone, and the loop is what matters. Because the gaps belong to different elements and different packages, no single trade sees the loop, and it appears only when someone traces the voids through a whole section. That is why the record says loops form between elements rather than within them.

This is the point that most often surprises people. The record states that a loop can circulate entirely within the construction without ever exchanging air with outside, so it can exist in a building that performs well under pressurisation testing. It names as a misunderstanding the assumption that a building performing well under such a test is free of bypass.

The air barrier record adds the other half from its own side: a test reads the envelope as a whole and cannot say which junction is leaking or whether a bond will last. Neither record is dismissing testing; both are saying it answers a different question from the one a cold surface is asking.

The bridging record is equally direct about the opposite error. Insulation added to the field of a wall acts on the area where heat was already moving slowly, and does nothing to the path running around it, so an assembly can be improved substantially in the field and left almost unchanged in behaviour because the junctions were never part of the exercise.

It goes further. Because the element concentrates flow precisely because the insulation around it is resisting, increasing the resistance of the field alone makes the concentration at the interruption more pronounced rather than less. Improving the field and improving the junction are separate exercises that have to be done together.

The record includes a third case that neither name obviously covers. At an external corner, at a parapet, at a projecting slab or at a recessed opening, the surface losing heat is larger than the surface receiving it from the room, so the interior surface runs colder even where the insulation line is unbroken.

It names the consequence for how you look for it: bridging cannot be read off a specification and has to be read off a section. It also names the misunderstanding that a bridge must pass right through the wall, when a step, a thinning or a corner produces the same concentration.

There is one position where the two problems meet and the resolution is a named component. Where ventilation is deliberately required next to insulation, as at an eaves or a ventilated void, the bypass record says a baffle holds the ventilation path open while keeping moving air off the insulation face, and describes it as not an accessory but the component that lets both requirements hold.

It also states the two ways of getting that wrong. Push insulation into the ventilation path and the ventilation stops. Leave the face open to the moving air and the insulation is being asked to work in a draught it was never enclosing. Both look like insulation correctly installed.

Both records send the consequence to the same place: the internal surface. The bridging record notes that the effect usually appears as moisture rather than as energy, because the interior surface at a bridge is colder than the surfaces around it, which is where surface condensation and mould appear first and where an occupant notices something long before any measurement is taken.

It then makes the qualification that matters for any proposed remedy. A colder interior surface only produces condensation if the air meeting it carries enough moisture, so bridging, internal humidity and ventilation act together, and treating any one of them alone can move the visible problem elsewhere rather than remove it. That is why the assessment belongs with a professional considering them together.