Control & Expansion Joint Spacing
Control joint spacing for a slab, with the panel-shape rule that governs narrow slabs — and why you probably need no expansion joints.
Which joint is which
Four different things get called joints and they do different jobs. Most slabs need three of them and not the fourth.
On a narrow slab, thickness is the wrong rule
The familiar guidance is 24–36× the slab thickness — 8 to 12 ft for a 4″ slab. The rule nobody quotes alongside it is that panels have to be roughly square, no worse than 1.5:1. On anything narrow that limit bites first. A 4 ft path in a 4″ slab is allowed 12 ft by thickness and only 6 ft by shape. Joint it at 12 and it cracks across the middle of every panel, which is exactly what you see on badly built sidewalks everywhere.
The two rules cross at a width of 8 ft for a 4″ slab. Below that, stop thinking about thickness. It's also why the old trade habit of jointing a path at intervals equal to its width works — it's the aspect rule, arrived at by eye a long time before anyone wrote it down.
And you probably don't need expansion joints at all. Concrete shrinks far more over its life than it ever expands, so a slab on ground is managing shrinkage, not growth. An expansion joint in a shrinking slab simply opens wider, and that gap destroys the aggregate interlock that was transferring load across it. What you do need is isolation joints wherever the slab meets something that won't move with it — columns, walls, footings, posts, drains.
- Cut a quarter of the depth. This is the part that gets skimped, and a joint that only scores the surface controls nothing — it has to create a weakened plane deep enough that a crack finds it easier than the concrete beside it.
- The timing window is narrow and temperature-dependent. Too early and the edges ravel; too late and the slab has already chosen its own cracks, at which point the joints are decoration.
- Control joints don't prevent cracking. Nothing does. They decide where.
- Reinforcement isn't an alternative to joints. It holds cracks tightly closed so they still transfer load — and steel running continuously through a control joint defeats it.
- Line joints up with column lines and re-entrant corners. An inside corner is where a crack will start whatever the grid says.
How to use
- Enter the slab size and thickness.
- Check which rule governs — on a narrow slab it is panel shape, not thickness.
- Cut a quarter of the slab depth; anything shallower controls nothing.
- Cut within the window for your saw, which is narrower than it sounds.
Frequently asked questions
How far apart should control joints be?
Twenty-four to thirty-six times the slab thickness, in the same units — which for a four inch slab gives eight to twelve feet. But that is only half the rule. Panels also have to be roughly square, no worse than one and a half to one, and on a narrow slab that limit governs instead. The correct spacing is the smaller of the two.
Why did my sidewalk crack between the joints?
Almost certainly because the panels were too long and thin. A four foot path in a four inch slab is allowed twelve feet by the thickness rule but only six by the shape rule, and jointed at twelve it cracks across the middle of every panel. The two rules cross at eight feet of width for a four inch slab — below that, thickness is simply the wrong rule to use.
Why do sidewalks have joints so close together?
Because the old trade habit of jointing a path at intervals roughly equal to its width is the panel-shape rule, arrived at by eye long before anyone wrote it down. A four foot path jointed every four or five feet gives panels close to square, which is exactly what the shape rule asks for. It looks conservative against the thickness rule and it is not.
How deep should a control joint be cut?
A quarter of the slab thickness — an inch in a four inch slab. This is the part that most often gets skimped and the reason joints fail to work. The cut has to create a weakened plane deep enough that a crack forming below it finds the joint easier to follow than the concrete beside it. A groove that merely scores the surface controls nothing.
When should I cut the joints?
Between four and twelve hours after placing with a conventional wet saw, or one to four hours with an early-entry saw. The window is temperature dependent and genuinely narrow: too early and the edges ravel behind the blade, too late and the slab has already chosen its own cracks, at which point the joints you cut are decoration.
Do I need expansion joints in a slab?
Usually not. Concrete shrinks considerably more over its life than it ever expands, so a slab on ground is managing shrinkage rather than growth. An expansion joint in a shrinking slab simply opens wider, and that gap destroys the aggregate interlock that was transferring load across it. What you do need is isolation joints wherever the slab meets something that will not move with it.
What is the difference between control and isolation joints?
A control joint is a deliberate weak line within the slab that decides where it cracks. An isolation joint separates the slab entirely from something else — a column, a wall, a footing, a drain — so neither restrains the other. They look similar and do quite different jobs, and the isolation joint is the one most slabs genuinely need and most often lack.
Do control joints stop concrete cracking?
No, and nothing does. Concrete shrinks as it dries, the shrinkage is restrained by the ground beneath it, and something has to give. A control joint decides where. A slab with well-placed joints has cracks in straight lines at the bottom of the cuts where nobody sees them; a slab without has the same amount of cracking distributed at random across the surface.
How much does a slab actually shrink?
Roughly four to eight hundred millionths of strain, which over a hundred foot slab is about half an inch to an inch. A twelve foot panel opens something like a sixteenth of an inch at each joint. That opening is what any sealant has to accommodate, and it is why joints in exterior slabs need a sealant that stays flexible rather than a rigid filler.
Does reinforcement replace joints?
No. Steel does not stop cracking — it holds the cracks that form tightly closed so they still transfer load and do not spall at the edges. It is a complement to joints rather than an alternative. Steel running continuously through a control joint actually defeats the joint, which is why mesh is cut at joint lines or the joint is placed where the steel stops.
Where else should joints go?
Lined up with column lines, and at every re-entrant corner — an inside corner is where a crack will start whatever the grid says, so the joint should run out of it. Joints should also pass through the full width of the slab rather than stopping short, since a joint that ends part way simply moves the crack to the end of it.
Is this a design?
No, it is layout guidance. Subgrade preparation, slab thickness, reinforcement, load transfer across joints and the curing regime all interact, and a slab carrying vehicles or racking is an engineered problem. On an ordinary slab the two things that matter most are cutting the joints on time and cutting them deep enough — both of which cost nothing and are both routinely got wrong.
🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.