Why Concrete Floor Slabs Crack and Move

Concrete placed today is a different element next month. It shrinks as it cures, giving up the water it was mixed with, and afterwards it goes on responding to temperature. Those are not defects; they are what the material does, and a floor is designed around them rather than in spite of them.

The question a design answers is therefore not whether the slab will move but where the consequence of that movement will appear.

A slab free to shorten shortens. A slab held still while it tries to shorten builds up force against whatever is holding it, and the material resolves that force by cracking somewhere it chose for itself. That is why the perimeter condition matters as much as the joint layout: the compressible strip at the edge exists so the slab can move without bearing hard against the surrounding structure.

The post-tensioned slab entry states the same principle in its most extreme form, where the freedom to shorten is the mechanism rather than a tolerance, and anything that holds the plate still takes part of the force that was intended for the concrete.

A layer laid over a deck either acts with it or independently of it, and that single decision follows from what the layer is for. The library states the consequences in both directions: a bonded topping cracks if the deck moves, and an unbonded one curls as it shrinks. Neither is a defect in itself — each is what that arrangement does — and the failure is in choosing the wrong one for the purpose.

This is why a crack in a floor is rarely diagnosable from the crack alone. What the layer was arranged to do has to be established first.

Two lines in a concrete floor can look identical and mean opposite things. A construction joint is the plane where one pour was stopped and the next cast against it, and the finished element is meant to behave as though the joint were not there. A movement joint is a deliberate discontinuity provided so parts of a construction can move relative to each other.

The library gives a physical tell rather than an appearance test. A construction joint is normally crossed by continuous reinforcement, because it is meant to act as one member; a movement joint is deliberately free of anything restraining the movement it exists to permit. It also notes that the sawn or formed contraction joint in a slab is a movement joint under a third name, and not a pour break.

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 a movement joint has eliminated it. The movement still occurs, and the crack appears in the finish, usually a short distance from where the joint was, and usually in a position nobody would have chosen.

The same applies upward through a build-up. Where a joint in the structure has to be carried up through an isolated layer above, and where that layer may run over, is decided from the structural layout rather than from the screed layout.

Joint positions are determined by where movement concentrates, which follows from the materials and the geometry rather than from convenience. That is why the design question in the entries pairs joints with the room layout: not because the layout decides the joints, but because the two have to be reconciled before a covering is chosen.

Joint positions are also where the programme and the design negotiate. A pour break that suits the plant and the working day may not be a break the designer has allowed, and the resolution belongs to the designer.

A joint layout is a design output for one floor, calculated against its materials, geometry and exposure. No entry describing a construction in general can supply it, which is why four of them raise it as a question to take to a professional and stop there. The useful move for a reader is to ask the question early enough that the answer can still shape the finish.