Laying a Paver Patio or Drive

The pavers are not the pavement

Strip the sentiment out of a segmental pavement and you are left with a flexible pavement structure that happens to have a decorative top. A concrete paver on a drive carries roughly the same job as an asphalt wearing course: it resists abrasion, sheds most of the water, and spreads a wheel load a little. Everything structural happens underneath it. The subgrade takes the load, the aggregate base spreads it, and the bedding layer does nothing more than let each unit seat true. When a drive fails, the paver is almost never the thing that failed — it is the layer that reported the failure.

That reframing changes how a crew spends its day. If the pavers are a wearing surface, then the hours that matter are the ones spent on excavation depth, on moisture in the subgrade, on lift thickness and compaction passes, on drainage. The hours spent laying units are production hours: satisfying, visible, and largely irrelevant to whether the job survives ten winters. Crews that get this backwards produce beautiful patios that ripple by the third spring, because the whole crew's attention went to pattern and cut quality while the base was built by whoever was free.

Sort every component on the job into one of two buckets before you start. Load-carrying: subgrade, geotextile separation, aggregate base, and — on a drive — the edge restraint, because unrestrained edges let the whole structure creep sideways and lose interlock. Non-structural: bedding sand, the pavers, joint sand, sealer. The second list is where the money looks like it is going and the first list is where the pavement actually lives. Interlock is the one bridge between them: units only share load with their neighbours if the joints are tight, the courses are true, and the perimeter is held. Break interlock and you have loose blocks sitting on gravel.

Interlocking concrete pavement design is covered in ASTM C936, Standard Specification for Solid Concrete Interlocking Paving Units, for the units themselves, and by the aggregate base and compaction requirements your local highway or building authority adopts for flexible pavement. Neither the paver spec nor a manufacturer's leaflet tells you how thick your base must be — that comes from subgrade strength and traffic, and it is a local determination.

Reading the subgrade before you order anything

Excavation is the first structural decision and the only one you cannot revise cheaply. Dig to the design depth, then look at what is at the bottom. Firm, uniform, granular soil that a plate compactor rings off is one pavement. Wet clay that pumps under the machine, or a bottom that varies from gravel to topsoil across ten feet, is a different pavement and needs a different base thickness — or needs undercutting and replacing before any base goes in.

Proof-roll the bottom rather than trusting the shovel. Run a loaded wheelbarrow or the plate over it and watch for deflection and for water rising to the surface. Soft spots that move under foot traffic will move under a car. Digging out a pocket of organic soil and backfilling with compacted aggregate is a morning's work at excavation stage and a full teardown after the pavers are down.

Uniformity matters as much as strength. A pavement built half over undisturbed subgrade and half over a backfilled service trench will settle differentially even if both halves are competent, because the trench consolidates on a different schedule. Trenches crossing the footprint want compacting in thin lifts and, where they are recent, want the base thickened or bridged across the trench line.

Separation geotextile between subgrade and base is not optional on fine-grained or wet soils. Without it, the base aggregate migrates into the subgrade and the subgrade fines pump up into the base, and the base thins by an inch over years without anyone touching it. On clean granular subgrade it earns less, but it never hurts. Lap the rolls generously and turn them up the excavation walls.

Drainage is set here too. Water that reaches the bedding layer and cannot leave will soften the base and, in freezing climates, will lift the whole assembly. Design a slope into the subgrade that matches the finished surface slope, so the base is a constant thickness rather than a wedge, and give the low edge somewhere to discharge — daylight, a drain, or a permeable edge. Trapped water inside a curbed patio with no outlet is a slow failure with no visible cause.

Building the layer that carries the load

Base aggregate wants to be a dense-graded crushed material with fines that lock, not a rounded washed stone that rolls. The gradation matters more than the rock type: a well-graded crushed aggregate compacts into an interlocked mass that behaves like a slab, while single-size stone stays a pile of marbles no matter how many passes you make.

Thickness follows traffic and subgrade, and there is no universal number. A pedestrian patio on good ground is a fraction of what a drive carrying a delivery truck needs on soft clay, and freeze-susceptible regions carry frost-depth requirements that dwarf both. What governs is your local pavement design guidance and building department — ask, and write the answer on the job sheet rather than carrying a remembered figure from a different climate.

Compact in lifts, not in one dump. Each lift should be within the reach of the machine you have — a small plate compacts far less depth than a reversible plate or a roller — and each lift gets full coverage in overlapping passes, both directions. The failure signature of a base placed too thick is a surface that tests firm and a lower third that never densified, which then consolidates under traffic. Uniform moisture during compaction does most of the work; bone-dry aggregate will not lock and saturated aggregate will pump.

Grade the base to the finished profile, not flat with a wedge of sand for correction. Bedding sand exists to seat the units and nothing else. A base that is an inch low in a hollow, filled with an inch of extra sand, will show as a depression once traffic consolidates the sand — sand does not compact into a stable thickness the way aggregate does. Screed rails and a straightedge over the compacted base catch this before it is buried.

Quantities for excavation, base, and bedding follow directly from the area, the base depth you settled on, and the compaction allowance for the aggregate, and getting the depths wrong is the expensive error rather than getting the paver count wrong.

Edge restraint as structure, not trim

Everything above the base survives on horizontal confinement. A field of pavers under a turning wheel wants to spread; the only thing stopping it is the perimeter. Treat edge restraint as part of the load-carrying list and it gets installed like structure — spiked into a base that extends past the last course, so the restraint is bearing on compacted aggregate rather than on the soil beyond the excavation.

The commonest edging failure is not the edging breaking. It is the base stopping at the paver line, so the restraint and its spikes sit in loose backfill, and the whole perimeter walks outward a fraction each season until the joints open and interlock is lost. Extend the compacted base beyond the finished paver edge — more on a drive than a patio — and set the restraint on it.

Vehicle edges, especially where wheels track close to the perimeter or where a drive meets a lawn, take a different order of force than a patio edge against a house wall. Restraint type follows that: what suffices on a walkway may not survive a turning truck tyre. Where a soldier course is used, remember it is being asked to hold the field, so the restraint behind it has to be up to the job.

Curves and inside corners concentrate the outward thrust. Spike spacing tightens on curves, and the aggregate behind the restraint gets compacted deliberately rather than shovelled back loose.

The bedding layer and the discipline of not fixing grade with sand

Bedding is a thin, uniform, screeded layer of coarse washed sand — consistent thickness across the whole job, screeded and never walked on before the pavers go down. Its single function is to let each unit find its seat and to accept a small amount of consolidation under the initial compaction. Every additional job people give it — correcting grade, bridging a low spot, making up for a base that came in shallow — degrades the pavement.

Two numbers govern here: uncompacted screed thickness and the amount it will consolidate. Keep the layer thin and consistent, and the consolidation is uniform and invisible. Let it vary from a half inch in one place to two inches in another and the surface will telegraph that variation permanently, because thicker sand consolidates more.

Screed rails set on the compacted base, pulled and backfilled as you go, are the reliable method. Rails resting on sand rather than base reproduce the base's errors and add their own. Screed only as far ahead as you will lay that session — rain on open bedding sand means rescreeding, and overnight sand collects footprints and debris.

Never compact bedding sand before laying. Compacting it produces a hard, dense skin that the units cannot seat into, and the whole point of the layer is that it deforms slightly to accommodate the dimensional variation between units. The compaction happens after the pavers are down, through the pavers, which seats them and drives sand up into the lower joints.

Laying: pattern, cuts, and keeping courses honest

Establish the layout from the most visible constraint, not from the most convenient corner. On a patio that is usually the house wall or the door threshold; on a drive it is often the garage apron or the street edge. Set a string line and a starting course that is genuinely square and genuinely straight, then check every few courses with the string rather than trusting the pattern to stay honest — small errors compound and are only correctable while the field is small.

Pattern choice has structural weight on a drive. Herringbone patterns resist the twisting and braking forces of vehicles far better than running bond or stack bond, because no continuous joint runs in the direction of travel. Running bond with joints aligned along the wheel path invites creep. On a foot-traffic patio the choice is aesthetic; on a drive it is a load decision.

Draw units from several bundles at once rather than working one bundle to exhaustion. Manufacturing produces slight colour and dimension variation between production runs, and blending across bundles distributes it invisibly. Working bundle-by-bundle produces visible blocks of shade that no amount of later cleaning corrects.

Cuts belong at the perimeter, never at the field edges of a heavy-traffic area, and never smaller than about a third of a unit — small slivers work loose, especially at edges and under vehicle loads. Where a cut would come out tiny, adjust the layout to split the difference across two courses instead. Cut dry with dust control or wet, but keep slurry off laid pavers; it stains.

Paver quantity with waste allowance for cuts and pattern is a number best fixed before the first bundle is opened, because ordering short mid-job risks a different production run and a visible colour break.

Compaction of the laid field runs after all full units and cuts are in and before joint sand is fully swept — a plate with a protective mat, in overlapping passes, working the whole field. This seats the units, consolidates the bedding, and reveals any unit that was riding high on a lump of sand.

Joints, jointing sand, and why interlock is fragile

Full joints are what turn a field of separate units into a pavement. Sand filling the joint transfers shear between adjacent pavers; an empty or half-filled joint transfers nothing, and the unit under a wheel deflects alone. Sweeping sand once and calling it done leaves joints that appear full but settle over the first weeks of use.

Work joint sand in with the surface dry, sweep, compact, sweep again, and expect to top up. Compaction after the first sweep drives sand down and exposes how much more the joints will take. Two or three cycles is normal, and a joint that keeps swallowing sand is telling you the bedding below is still consolidating.

Polymeric jointing behaves differently and is unforgiving of process. It requires a genuinely dry surface, a specific fill depth below the paver chamfer, careful removal of every grain from the paver faces before wetting, and a watering method that activates without washing the joint out. Residue left on the surface cures as a haze that is difficult to remove. Ambient temperature and forecast rain both govern whether it can be installed that day, and the product's own instructions carry the limits.

Joint width is a spec, not a preference. Units are manufactured with spacer bars or a nominal joint, and forcing units tighter than designed removes the sand's ability to transfer load; letting them drift wider than designed loses interlock and lets sand escape. Consistency across the field matters more than any particular width.

Sealing is a maintenance and appearance decision, not a structural one. It resists staining and slows joint sand loss, and it needs a fully cured, dry, efflorescence-free surface — which usually means waiting a season on new work. Sealing over a pavement whose base is failing hides nothing and fixes nothing.

Failure signatures and what they say about the base

Read a failed pavement backwards through the layers. Rutting along wheel paths on a drive means the base or subgrade is deforming under load — either the base is thin for the traffic, was compacted in lifts too thick, or the subgrade was never adequate. No amount of relifting pavers fixes it; the base has to come out and be rebuilt to the right thickness.

A localised dip, sharp-edged, usually points at a soft spot in the subgrade or a service trench that consolidated. Broad, gentle waviness across a large area more often means bedding sand of inconsistent thickness, or a base graded by eye rather than by straightedge.

Edges that have spread, with joints opening progressively toward the perimeter, are a restraint failure — and almost always a base that stopped short of the restraint. Look at whether the base extends beyond the last course before blaming the edging product.

Heave in freezing climates points at water and frost depth: either the base is above frost-susceptible material with no drainage path, or the excavation never reached the depth local practice requires. Standing water on the surface after rain says the finished slope is too flat or was lost during laying; a segmental pavement wants a positive fall in a direction that has somewhere to go.

Repair is where segmental pavement earns its keep. Lifting units, correcting the layer beneath, and relaying the same units is genuinely possible — which is why marking and stockpiling a small number of spare units from the original run is worth the pallet space. What is not repairable at the surface is a base built wrong, and that is the argument for spending the front half of the job below the pavers.