Underfloor Heating and the Floor Build-Up

Pipes or cables cast into a screed occupy a defined position in its depth. That alone removes some of the freedom a screed usually has: the layer is no longer simply a thickness to be adjusted, because something inside it has to remain where it was placed and covered as the design assumed.

The floating screed entry sets the relationship out directly — embedded heating locks several decisions together, fixing the screed's depth and its drying behaviour and turning the joint layout into a heating question as well as a structural one.

Where the build-up carries heating, the floor's response is governed by what sits under the pipes as much as by the pipes themselves. Whether heat goes upward or into the structure is decided by the layer beneath, and the entries describe that as a design matter rather than a product one.

This is the same decision as the insulation position in a ground floor, arriving from the other side. A slab inside the insulated envelope stores heat; one outside it responds differently to heat input. Adding embedded heating to a build-up designed without it changes that arrangement rather than adding to it.

A screed contains water when it is laid and gives it up slowly, and a floating build-up can generally only dry upward. The entries are consistent that when a covering may be laid is judged by testing rather than by programme, and that the answer depends on the layers below as much as on the screed.

Embedded heating adds a further stage that belongs to the heating designer and the screed manufacturer together, and it is one of the places a programme most often assumed a shorter period than the build-up allows.

A floating layer will move, and the planned lines dividing it — at doorways, at changes of shape, and where heated areas are zoned — exist because of that. With heating embedded, the screed expands and contracts in service rather than only while curing, so the joint layout is decided with the heating zones and the covering rather than after the covering is chosen.

A movement joint only works if it passes through everything that has to move, including whatever is applied over it. A rigid finish laid across one has eliminated it, and the crack then appears in the finish a short distance from where the joint was.

The covering has to tolerate the movement the build-up is deliberately allowing and the joints running through it, so covering and screed are chosen against one another rather than one after the other. Once heating is embedded, the covering choice is bound to the thermal layer position beneath it as well.

The entries put the determination where it belongs: whether a covering tolerates the movement and the drying regime of a particular build-up is a matter for its manufacturer's documentation, not a general rule.

A doorway is where the isolation is most often bridged and where the covering usually changes. The joint in the screed, the threshold detail and the edge strip all meet in a narrow line that several different trades pass through, and a single hard contact there reconnects the floating layer to the structure.

The perimeter strip has to run the full depth of the screed and stay unbroken, including at doorways and behind fitted units, which makes the heated floor's edges a sequencing problem as much as a detailing one.

Anything embedded in a screed is effectively permanent. A leak in an embedded pipe is found by tracing rather than by inspection, so what goes in and how it is recorded matters more here than in a floor where services stay reachable.

The entries turn that into a design question rather than a handover formality: is anything embedded in the screed, and is its position being recorded for whoever comes next. A floor that cannot be drilled and has no record of why is a constraint the next owner inherits blind.