Nearly every enquiry we receive that turns out to be a mismatch starts here. A drawing says "expansion joint" on a line that is a saw cut in a slab, or a BOQ has one item covering both a 6 mm groove and a 100 mm structural separation, and the rate that comes back is either absurd or meaningless. It is worth ten minutes to separate them properly.
The three, in one table
| Construction joint | Control joint | Expansion / movement joint | |
|---|---|---|---|
| What it is for | Managing where one pour stops and the next begins | Deciding where shrinkage cracking will happen | Letting two parts of the structure move relative to each other |
| Does it allow movement? | No — the opposite. It is designed to transmit force across | Contraction only, and a small amount | Yes — open, close, shear and vertical offset |
| Depth | Full depth of the element | Partial — typically a quarter to a third of the slab depth | Full depth, right through the structure |
| Reinforcement | Continuous through | Usually continuous | Discontinuous — the two sides are separate structures |
| Typical width | None — concrete against concrete | 3–10 mm groove or saw cut | 25–500 mm, from the calculation |
| Who decides it | The contractor, within the engineer's rules | The engineer or the flooring designer | The structural engineer of record |
| What covers it | Nothing — it is invisible when done well | A sealant or a light insert profile | A manufactured joint system |
Construction joint — a pour boundary, not a movement joint
A concrete pour has a practical size limit: how much can be placed, compacted and finished before the earlier concrete sets. A construction joint is simply where one pour stops so the next can start.
The design intent is that it should behave as though it were not there. Reinforcement runs continuously through it, the face of the first pour is prepared — roughened, cleaned, sometimes with a bonding agent or a water bar — and the finished element transmits shear and moment across the line exactly as monolithic concrete would.
Its position is normally the contractor's to choose, but within limits the engineer sets: at points of low shear, away from critical sections, and not where the engineer has said not to. It is a construction sequencing decision that has structural consequences, which is why it appears on method statements rather than on the movement joint schedule.
It is not our scope and it should never appear on a joint supplier's BOQ. If one does, something has been mislabelled upstream.
Control joint — deciding where the crack goes
Concrete shrinks as it dries, and a restrained slab that shrinks will crack. That is not a defect and it is not preventable — the question is only whether the crack appears in a straight planned line or wanders across the floor.
A control joint (also called a contraction joint) is a deliberate line of weakness that answers that question. A groove formed in the wet concrete or, more commonly, a saw cut made early — within hours of the pour, before shrinkage stress has built up — reduces the section along that line so it becomes the weakest place in the slab. The crack forms there, below the cut, out of sight, and stays tight because the aggregate interlocks across it and the reinforcement usually runs through.
Two things follow that are worth knowing:
- The timing is critical and the depth is not arbitrary. Cut too shallow and the slab cracks somewhere else anyway; cut too late and it has already cracked. A quarter to a third of the slab depth is the usual target, and the window for cutting is measured in hours.
- A control joint does not accommodate expansion. It manages contraction. If the slab needs to grow, a control joint offers it nothing, because the two sides are still connected by aggregate interlock and steel.
Control joints are spaced far more closely than expansion joints — a few metres in a ground-bearing slab, against tens of metres for a movement joint. A drawing with lines at 5 m centres labelled "expansion joint" is almost always a control joint layout that has been mislabelled, and pricing it as movement joints produces a number nobody can use.
Expansion joint — a real separation
An expansion joint, or movement joint, is a complete break through the structure. Slab, beam, wall, finish and reinforcement all stop, and the two sides are structurally independent from that line onward. It exists so that thermal movement, shrinkage, settlement and seismic drift have somewhere to go instead of arriving as damage.
Because the gap is real and open, something has to cover it — and that cover has to let the gap keep changing width while carrying traffic, staying watertight, holding the fire compartment line and sitting flush enough to be safe to walk on. That assembly is the joint system, and it is what we manufacture.
Widths run from 25 mm on a modest thermal joint to 500 mm on a seismic separation between two tall blocks. How that number is arrived at is worked through in calculating the gap width, and where the joints go in the first place in how far apart should expansion joints be.
Two more that belong in the same conversation
- Isolation joint. A separation between a slab and something rigid passing through or beside it — a column, a wall base, a machine plinth, a lift pit — so the slab can move without the column restraining it. Narrow, full depth through the slab, and the reason a floor tented near a column is so often a floor with no isolation joint at that column.
- Movement joint in the finish. The joint in the tiling, screed or stone. It must sit over the structural joint, full depth of the finish, and it is usually the tiling contractor's scope rather than the structural one — which is exactly why it gets left out. See why large tiled and stone floors tent.
Why the labels matter commercially
Three concrete consequences, all of which we see:
- Unusable tender pricing. One BOQ line covering "expansion joints" across a project that contains both saw cuts and structural separations cannot be priced sensibly. Bidders either price the cheap interpretation and claim variations later, or price the expensive one and lose. Neither serves the client.
- The wrong product installed. A line labelled expansion joint but detailed as a control joint gets a sealant. When the structure moves, the sealant fails, and the investigation finds a specification that never asked for a movement joint in the first place.
- Scope falling between trades. The structural joint is on the structural drawings; the finish joint over it is on the architectural ones; the fire and waterproofing continuity through it is on neither. That interface is where most joint failures we are called to actually originate — not in the product.
Name each joint for what it does, on the drawing, before the BOQ is written. It is the cheapest quality control available on the whole package.
A short checklist for a drawing set
- Every line on the drawing labelled with which of the five it is.
- Movement joints given a width and a movement, not just a width.
- The finish joint shown coincident with the structural joint, not near it.
- Wall, ceiling and facade continuations of each floor joint shown — joints do not stop at the skirting.
- Fire rating and waterproofing continuity through the joint stated, and assigned to somebody.