Open-Graded Bedding: The Layer That Has to Stay Open

A Bay That Got Firmed Up First

Somebody runs the plate over the screeded bed before the first block goes down. It is not sabotage and it is not laziness — it is thirty years of muscle memory from sand-bedded work, where a loose bed under a wheelbarrow is a nuisance and a firmed one lays quicker and truer. The stone tightens, the screed marks disappear, the bay goes down beautifully, and the foreman notices nothing because there is nothing to notice.

Two months later the ring test at handover reads a fraction of what the bay beside it reads, and the two bays are indistinguishable to look at. Every joint is full of the right stone. The subbase below is the depth on the drawing. What changed is that a layer whose whole function on a permeable pavement is to pass water at a rate nobody ever specifies — because nobody expects it to be the constraint — was consolidated into the throttle for everything underneath it.

That is the difference worth holding on to. On a conventional pavement the laying course beds the unit and absorbs the compaction, and density in it is a virtue. Here the same fifty millimetres does that job and also sits in the flow path between the aperture and the reservoir, and the two requirements pull in opposite directions. Everything below is about the second one, on the assumption that you can already lay blocks.

Single-Sized Is a Purchase Order, Not a Description

Open-graded means a grading curve with the intermediate and fine sizes deliberately absent, so the particles bear on one another and the space between them stays space. ASTM D448, Standard Classification for Sizes of Aggregate for Road and Bridge Construction, is where the size numbers on your order come from, and the reference bedding material in ASCE/T&DI 68-18, Permeable Interlocking Concrete Pavement, is No. 8 at a nominal fifty millimetres screeded thickness. Some unit manufacturers call for No. 89 or No. 9 instead, usually because their aperture is too narrow for No. 8 to fill reliably and they would rather one stone served both the bed and the joint.

Washed is a separate word from the size number and it has to appear on the order in its own right, because a quarry whose day job is dense-graded road base will happily load you material that meets the No. 8 band and still carries the dust it was crushed in. The number that decides whether a load is fit is the material finer than the 75 micrometre sieve, measured by ASTM C117 alongside a sieve analysis to ASTM C136/C136M. Ask for both per delivery and read them at the gate, because a load rejected on the weighbridge costs an hour and a load rejected after screeding costs a bay.

Screeding It Without Consolidating It

Rails first. Steel tube or planed batten laid on the locked choker, spaced so the screed board spans them without any sag in the middle, and set to the underside of the bed — finished block level, less the unit thickness, less the nominal screeded depth, less nothing else. Every millimetre of rail error becomes surface error at a point where the surface can no longer be adjusted, so the rails get levelled off the same datum as the blockwork and checked at both ends and the middle of every run.

The pull is one pass. Tip generously ahead of the board so it is always working into surplus, draw it steadily, and take the surplus away rather than dragging it back over ground you have already screeded. A second pass across a hollow does not fill the hollow; it brings the board down onto the stone either side of it and starts kneading a material that is only doing its job while it is loose. What you are producing is loose stone at a correct thickness, not a smooth stone floor, and a bed that looks slightly unconvincing is usually the right one.

Lifting the rails leaves troughs, and those get filled by hand with the same stone, tipped a little proud and struck off with the edge of a trowel held vertical so it cuts across the surplus. What must not happen is the flat of the trowel going onto it. A troweled patch is a plate compactor with a small footprint, and it will read as a low block at exactly the spacing of your rails, which is the most diagnosable defect on the whole job and the most embarrassing.

Then nobody goes on it. Barrow runs stay on the laid blockwork with plywood down where the wheel turns, stone comes forward in buckets from the face, and the crew works off the pavement they have already built. Where a bay genuinely has to be crossed — a gully to be reached, a level to be checked — a board goes down and comes straight back up, and the print underneath it is reworked loose before the next unit lands on it. Kneeling boards are worse than boots, because they spread the load over exactly the area you least want densified evenly.

Screed only what will be covered before the crew leaves. An open bed overnight collects whatever the site's surface water is carrying, and on a windy afternoon next to a cut-and-fill operation it collects the site itself. A strip screeded ahead of laying is a working convenience that turns into an exposed filter the moment it rains, and the fines that arrive that way sit in the top of the bed where they are least reachable and most effective.

1. Lock the choker course until the surface stops moving under the plate, and no further.
2. Set rails on the choker to the underside of the bed, levelled from the same datum as the blockwork.
3. Tip ahead of the board and pull the bed in a single pass, carrying the surplus off rather than back.
4. Lift the rails and refill the troughs loose, striking off with a trowel on edge and never on its flat.
5. Lay from the face onto the loose bed, keeping all traffic on the units already set.
6. Fill the joints, compact, top up and compact again until the joints stop taking stone.
7. Screed no further ahead than the day's laying, and cover or close off any bed left open.

Everything That Can Close the Voids Between Screeding and Compaction

The most concentrated source of fines on the whole job is generated on the pavement, by your own crew, after the bed is down. Wet-saw slurry is rock flour in suspension, and it is produced within a metre of open joints and a screeded bed at the exact moment both are most vulnerable. Cut off the pavement — a bench, a board, a bunded tray, water taken away rather than allowed to find its own level — and never let cut water run back across laid units. A dry-cut station has the same problem in airborne form and needs to sit downwind of the bay, not on it.

Stockpiling is the quiet one. Washed stone tipped onto bare ground is washed stone plus the ground, and the last bucket out of a pile is mostly what was underneath it. Tip on a hardstanding, a sheet or a bay of clean boards, keep the last few hundred millimetres of every pile out of the job, and do not use a bucket that has been in topsoil that morning without washing it. The same applies to the loader ramp: mud tracked onto laid blockwork gets swept into the joints by the next person tidying up.

Neighbouring work is the third route, and it is the one nobody on the paving crew controls. Topsoil placed against the edge, render or brick washings running off a wall above, a wheel wash draining across the corner, an unstabilised slope upslope of the bay. One storm across bare soil delivers more fines than years of service traffic, and it puts them below the surface where sweeping cannot reach. If the pavement has to be built before the surrounding areas are stabilised, it needs a physical barrier or a sacrificial cover, and someone has to own that decision in writing.

Weather sits across all of it. A screeded bed in heavy rain is a filter running at full flow with nothing above it to intercept anything, and a bed that has taken a storm is not repaired by drying out. Plan the day so that stone placed is stone covered, and accept a shorter laying face in poor weather rather than an open bed and a longer one.

Compaction, and the Level You Cannot Get Back

The sequence is fixed and the reason for it is the bed. Units are set on the loose course, joints are filled, and only then does the plate go on — over the units, transmitting through them, closing the bed once with the load spread across the whole block face rather than into a bare stone surface. The bay comes down as that happens, which is why it is laid proud, and how proud is a number confirmed on a trial area against a string line rather than carried in from the last job. A neoprene or rubber-faced pad on the plate protects the block surface and the chamfers, and costs nothing next to a bay of scuffed units.

Work the whole bay including right up to the restraint, and keep back from the leading edge by a course or two so that nothing is compacted while it is still free to move sideways. Then top the joints up and go again, and keep going until the joints stop swallowing material and the surface stops settling under the machine. Blocks that were bedded on a course with a thick patch in it will show themselves during this, not before, because a thick spot closes further than a thin one and the difference only appears once the load has been applied.

None of it can be corrected from above. A block sitting low against a kerb, a threshold or a gully cannot be lifted and packed, because the packing under one unit will behave differently from the screeded bed under its neighbours and the difference reappears within a winter. The correction is to lift the affected area back to the bed, re-screed it and relay, and the argument about whether it needs doing belongs to the surface tolerance in the specification — which is worth reading before laying rather than after, since it is also what the ring test locations will be chosen around.

Proving the Bed Is Still Open

Test the finished pavement and write the result down. ASTM C1781/C1781M, Standard Test Method for Surface Infiltration Rate of Permeable Unit Pavement Systems, is the ring test for this surface, and a new pavement will return a number far above any design intensity, which is the point: the value is not the pass mark, it is the baseline everything later gets compared against. Take it at several locations, mark them on a plan, and include the bay that was laid in the worst weather rather than only the one that looks best.

That baseline is also the diagnostic. When a reading is low, the first question is whether the blockage is in the joint or below the units, and there is a cheap way to find out: replace the top of the joint fill in the test location and retest. If the rate comes back, the fines were surface-bound and a vacuum or regenerative-air sweeping regime will manage them. If it does not, the restriction is below the blocks — a bed that was compacted, contaminated, or laid without a choker under it — and no maintenance regime reaches it. That distinction is worth establishing while the installer is still on site.

So the paperwork is part of the work. Record the designation and the quarry, the washed specification and the delivery tickets that went with it, the screeded thickness and the allowance the order was built on, the baseline readings with their locations, and the prohibition on sand in the joints. A bay that was firmed up with a plate before laying looks exactly like the bay next to it for the rest of its life, and the only evidence that ever distinguishes them is the record you kept and the number you took on the day.