A raised floor is designed for a wheel, not for a crowd
An access floor's published ratings look like ordinary structural figures and are governed by a different case. The load that matters is CONCENTRATED — a single wheel, a castor, a jack or an equipment foot pressing on one point of one panel — rather than the distributed load that governs a slab.
The reason is that a panel is small and simply supported on pedestals at its corners, so a point load near the middle of a panel produces far more bending in it than the same total load spread over its area. A floor comfortably rated for a uniform load can fail under a trolley wheel carrying a fraction of it.
Position within the panel matters as much as magnitude, which is why ratings are quoted with the load applied at the worst position over a stated small contact area — and why a rolling load rating is a SEPARATE figure from a static concentrated one. Rolling loads are quoted for a number of passes, because the failure mode is progressive: repeated loading over the same track fatigues the panel and works the pedestals loose.
Contact area is the term most often ignored on site. A hard, small castor concentrates the same weight into a fraction of the area a pneumatic tyre or a spreader plate would, so the same equipment moved on different wheels is a different load case — which is why equipment routes across access floors are specified with protection boards rather than left to the mover.
- a_i
- area of each opening
- a_min
- the deduction threshold — a convention from a standard method, not a physical value
- A_net
- the measured quantity, which is legitimately larger than the geometric net area
A small opening is deliberately not deducted
Standard methods of measurement state which openings are deducted and above what size, and below that threshold nothing is taken out. That looks like a rounding convenience and it is an economic decision.
Work around a small opening COSTS MORE per unit than the plain work it interrupts. The material has to be cut, the offcut is usually unusable, the edges need trimming or making good, and the labour per square metre in that zone is a multiple of the labour in the open field. Deducting the opening would reduce the measured quantity for work that actually got more expensive.
So the threshold is set where those two effects roughly cancel, and it differs by trade and by standard — plasterwork, tiling, painting, cladding and flooring each carry their own rules, and two standard methods can give different, equally correct net areas for the same wall.
The practical consequence is that a net area has to state which convention produced it. Comparing a measured quantity against a geometric one taken off a model will show a difference, and the difference is usually the convention rather than an error in either.
A cutout removes area and adds perimeter
Taking a hole out of a surface reduces the material quantity and increases almost everything else, and the increase follows the PERIMETER of the hole rather than its area.
Every cutout adds edge trim or a frame, the cutting labour itself, and in most systems some means of supporting the cut edges. In a raised floor that support is explicit: panels cut for a service penetration lose corner support, so extra pedestals, stringers or a purpose-made frame go in around the opening — and the panels around a large cutout can need a higher rating than the field, because they are now carrying an edge.
That perimeter-driven cost is why many small cutouts are far worse than one large one of the same total area, and why a services layout that clusters penetrations is cheaper to build than one that distributes them evenly.
Cutouts also interrupt the floor's plan-level continuity. An access floor braced by its own panels loses some of that where a large opening is cut, and a run of openings along one line is the worst case — the same reason a run of holes in any diaphragm is worse than the same holes scattered.
A light well is a perimeter times a depth — until it splays
The internal surface of a STRAIGHT light well is its opening's perimeter multiplied by its depth: four walls of a box, and the quantity to be lined, plastered and painted. That is the case the calculator here answers, and it makes the perimeter — not the opening's area — the quantity that drives it.
The consequence is the one the perimeter paper sets out. Two wells admitting the same area of sky can need very different amounts of finish, because a long narrow opening has far more perimeter than a square one, and a deep well multiplies that difference by its depth.
A SPLAYED well breaks the box. Flared wider at the room than at the roof, its internal surface is a frustum, and its area exceeds both the perimeter of the top and that of the bottom times the depth — because the sloping faces are longer than the vertical distance they span. Measuring a splayed well as a straight shaft under-measures the lining, and the slope factor that corrects it is the one the pitched-roof paper uses.
The splay exists for optical reasons rather than for finishing ones: a flared well spreads the light across a wider area of ceiling and softens the edge of the pool below, and a splay on the side facing away from the sun's path does more than one facing it. The surface's REFLECTANCE does the rest, which is why a light well is finished white as a matter of performance rather than taste — a dark well can absorb most of what enters it before it reaches the room.
Levels: set out from a datum, never from the last piece
Pedestal heights, packing shims and any set of supports carrying a level or a fall are differences between two surfaces, and the way those differences are measured decides whether the finished plane is flat.
Every measurement carries an error. Taken from a single DATUM — one established level line or point that every measurement refers back to — those errors stay independent and none of them exceeds the accuracy of one reading. Taken cumulatively, each piece measured from the one before, the errors ADD, and a long run can drift by many times a single reading's error while every individual measurement was correct.
That is why levels are transferred by instrument to a datum line around a room rather than stepped along the floor, and why a fall across a paved area is set from the same datum at every point rather than by stacking a gradient piece by piece.
The substrate supplies the other half of the answer. A pedestal's height is the design level less the thickness of what sits on it less the level of the structure beneath — and that last term is a measured surface rather than a drawn one. A structural slab within its own construction tolerance can vary by more than the adjustment range of a fixed pedestal, which is why adjustable pedestals exist and why the survey of the slab happens before the pedestals are ordered.
Calculators that use this method
Basis
- PSA MOB PF2 PS/SPU and BS EN 12825 for access floor performance classes, including concentrated load at the worst position and separate rolling load ratings by number of passes.
- CISCA Recommended Test Procedures for Access Floors, for concentrated, rolling and ultimate load test methods and the stated indenter contact areas.
- RICS New Rules of Measurement 2 and the standard methods of measurement for trades, for opening deduction thresholds and the reason work around small openings is not deducted.
- BS 8204 and equivalent guidance on floor level tolerance and datum transfer, and manufacturers' adjustment ranges for fixed and adjustable pedestals.
- IES daylighting guidance on splayed light wells and the effect of splay geometry on the distribution of admitted light.