Building a Pedestal Paver Terrace

Survey the Fall You Have, Not the One on the Section

Drawn falls are intentions. Tapered insulation gets cut to a schedule, boards get swapped at the eleventh hour, a sump gets set 15 mm proud, and the deck itself deflects under everything you are about to put on it. On a refurbishment the drawn fall may not exist at all. Take levels on the finished waterproofing, on the grid you intend to build to, and treat the result as the only fall that matters.

Shoot the whole area from one instrument position if you can, because moving the level halfway across is how a systematic error gets built into a height schedule that then looks entirely self-consistent. Record the deck level at every pedestal grid line, not at the corners of the terrace — five spot heights across twelve metres will miss a local flat spot completely, and a flat spot under a pedestal terrace is a pond you have built a lid over. Where the readings disagree with the drawing by more than the pedestal's adjustment can absorb, that is an RFI before it is a purchase order.

Two readings deserve separate attention. The high point governs the shortest pedestal, and the shortest pedestal in most ranges is a fixed or barely-adjustable component — go below its height and the geometry has nowhere to go. The low point governs the tallest, and tall pedestals stop being a stock item and start being a base plus extenders, with a different lead time and a different price. Both ends of the survey therefore land straight on the order.

Timing matters as much as method. Survey after the waterproofing is complete and after any ballast or protection layer is down, because both change the surface the pedestals will actually stand on, and survey before the pedestals are ordered rather than before the insulation is cut. On a long-span deck, take the readings with the deck loaded as far as the programme allows — a bare deck reads flatter than the same deck under the finished weight of the paving, and the fall you sign off is the one the pedestals get wound to.

1. Set the instrument once, in a position that sees the whole terrace, and note where it stood.
2. Level the finished waterproofing at every pedestal grid line, working across the fall then along it.
3. Mark the high point and the low point on the plan and write their readings on it.
4. Check the run between them with a tape on the deck, not scaled off the drawing.
5. Flood-test or hose the area and record where water stands, then re-level those spots.

Count the Grid Before You Price the Heights

Pedestal count is a grid problem and it is genuinely simple: a line of pedestals every module across the terrace in each direction, plus one to close each run. Twelve metres against a 600 mm module gives twenty bays and twenty-one lines; seven metres gives eleven full bays, a 400 mm remainder and thirteen lines. Twenty-one by thirteen is 273 heads, and that is the figure the material order starts from.

It starts from it rather than ends at it. The remainder strip along the seven-metre dimension is a row of cut pavers, and cut pavers need support along a sawn edge that the corner-bearing grid does not provide; add heads there, at planter kerbs, at drain surrounds where a slab is trimmed around an access frame, at any step or level change, and along the parapet where the last course lands on a support rather than a shared corner. Module comes from the paver and the loading rather than from preference, so if the specification changes across the terrace — heavier units at the exposed corner, a lighter unit in the sheltered pocket — the module can change with it, and the count becomes two grids added together instead of one grid over the whole area. On a terrace carrying much furniture those extras stop being a rounding allowance and become a line of their own on the take-off.

One Location at a Time, Then Bought in Bands

A pedestal height is calculated for one place on the roof. Finish height above the deck at that point, less the paver thickness, less what the correction at that location consumes, and the answer applies to that pedestal and its immediate neighbours only. Read the first term carefully: it is the height above the deck under that pedestal, not above a single datum for the whole terrace. Feed a roof-wide datum into a location-specific formula and the fall comes out with the wrong sign, which is precisely the mistake that produced 273 identical pedestals in the first place.

Once the survey exists the schedule is arithmetic. On the terrace above, with the shortest pedestal at the high point set to 60 mm and the deck dropping 10 mm per 600 mm grid line, the heights run 60, 70, 80 and on up to 260 mm at the outlet corner — twenty-one nominal heights across the fall, thirteen pedestals wide at each of them.

Nobody orders twenty-one nominal heights, because a threaded pedestal covers a range rather than a value. You group the grid lines into the ranges the manufacturer publishes, count the heads in each group, and order bands. Three bands is typical for a fall of this size; a shallower fall over the same run may collapse into two, and a terrace stepping over a downstand beam will need a band of its own for the step.

Bison Innovative Products and Buzon Pedestal International both publish range tables for their bases, extenders and slope-correcting heads, and the bands below are grouped against a range set of that shape rather than against any one product. Take the actual break points from the literature for the system specified, because a band boundary in the wrong place turns into a site full of pedestals wound to the last thread — which is where they creep, and where they rock.

Then add the components the bands do not contain. Slope correctors for wherever the fall exceeds what a flat head can sit on. Extenders as a separate line, not folded into the base count. Spare heads at a sensible percentage, because damaged threads on a windy roof are ordinary. And a small stock of the shortest component in the range, since the high corner is always tighter than the survey suggested.

What the Pedestal Is Actually Standing On

A pedestal gathers the load of a module of terrace and puts it through a base plate. On a 600 mm grid that is 0.36 m² of tributary area — the pavers, the people, the planters and whatever the code assigns as imposed load for an accessible terrace — arriving on a base plate of a few hundred square centimetres. Three and a half thousand square centimetres of terrace resolving onto two or three hundred is the number to hold on to. Below the plate sit the membrane, the cover board and the insulation, and it is almost never the membrane that governs. It is the insulation's compressive behaviour.

That behaviour is specified, not assumed. ASTM C578, Standard Specification for Rigid, Cellular Polystyrene Thermal Insulation, and ASTM C1289, Standard Specification for Faced Rigid Cellular Polyisocyanurate Thermal Insulation Board, both grade product by compressive resistance, and the underlying measurement is ASTM D1621, Standard Test Method for Compressive Properties of Rigid Cellular Plastics. Two cautions come with those figures. The compressive resistance is quoted at a stated deformation from a short-term test, and the load under a terrace is permanent — manufacturers publish a separate long-term or creep design stress for exactly this reason, and it is a good deal lower. Second, the grade that is adequate under a membrane in the open field is frequently a grade or two short of what a paved area needs, so the insulation specification often has to change under the terrace and stay changed at the boundary.

The remedies are all about area. A cover board over the insulation spreads the plate load into something the board below can carry; a bearing pad or a proprietary load-spreading disc under the base does the same job at the other end; a slip sheet keeps the pedestal foot off the membrane so that three decades of small movements do not wear a hole in it. Get the sequence right and the loads resolve. Get it wrong and the evidence is a dimple pattern in the membrane on a 600 mm grid, found the first time anybody lifts a paver.

Wind Gets Under a Raised Paver Before It Gets Over It

A ballasted membrane holds down because stone sits on it. A pedestal paver has open air beneath it and an open joint at every edge, so pressure equalises under the slab and the aerodynamics are not the same thing at all. ANSI/SPRI RP-4, Wind Design Standard for Ballasted Single-ply Roofing Systems, addresses ballast; a raised paver deck is assessed through the components-and-cladding route of ASCE 7, Minimum Design Loads and Associated Criteria for Buildings and Other Structures, together with the paver and pedestal manufacturer's own wind testing. FM Global Property Loss Prevention Data Sheet 1-29, Roof Deck Securement and Above-Deck Roof Components, governs on FM-insured work. Which of those applies is a project question with a written answer, and it is not one to settle with a rule of thumb about slab weight.

What every one of them agrees on is that a roof is not one wind zone. The corners see the worst of it, the perimeter strip sees more than the field, and the width of those zones is a calculated dimension driven by the building's least horizontal plan dimension and its mean roof height. On a terrace that sits in the corner of a building — which is where terraces tend to sit, because that is where the views are — a large fraction of the paved area can fall inside the corner and edge zones. That changes the paver specification, not just a note on the drawing.

Responses run in a familiar order: heavier units in the high-pressure zones, mechanical clips or interlocking heads, a restraint at the perimeter, or a solid perimeter course laid directly on a support rather than on adjustable heads. Each of those alters the take-off — heavier pavers change the module, clips are a fixing count, a restrained perimeter is a linear item — so the wind assessment belongs before the material order, not after the drawings are stamped.

Wind also arrives during the build, and the part-finished deck is the weakest the assembly will ever be. Loose pavers stacked on temporary heads have none of the restraint the finished terrace relies on, and a corner course set but not yet clipped is exactly the unit the design identified as most exposed. Weight down or strap anything left overnight in the zones the calculation flagged, and stop laying in the corner when the forecast turns rather than working to the edge and hoping.