Clearances, Layout and Accessibility Dimensions

These are accepted values, not derived ones

Almost every figure on this page came from measurement or from consensus rather than from a calculation. A turning space is the swept circle of a wheelchair measured across a population of users; a manoeuvring clearance is the floor area a person needs to approach a door, reach the handle and pull it past themselves; a parking stall width is a vehicle plus the door opening beside it.

That has two consequences for how they are used. First, they cannot be interpolated or argued down: a dimension that is ninety per cent of the required one is not ninety per cent compliant, it is non-compliant, and the person it excludes is excluded completely.

Second, they vary between jurisdictions in ways that look arbitrary and are not — different standards were fitted to different survey populations at different dates, and several have been revised upwards as the equipment they describe has grown. Powered wheelchairs and mobility scooters are substantially larger than the manual chairs the earliest dimensions were based on, and the standards have moved.

So the honest output of every page here is a figure with its source named. The calculators state which standard's value they have applied, because the same question genuinely has different correct answers in different places.

A door needs floor space on the side nobody draws

A door's clear opening width is the dimension everyone checks, and it is the easy half. The harder half is the MANOEUVRING CLEARANCE: the area of level, unobstructed floor a person needs on each side to operate it.

The requirement is asymmetric and it depends on the approach. Pulling a door requires clear floor on the latch side, because a wheelchair user has to position beside the leaf, pull it past themselves and then move through — and the space needed on that side is substantial. Pushing needs less. Approaching from the front needs less than approaching along the wall.

This is why a door that satisfies its width can still be unusable, and why the classic failure is a door at the end of a corridor, or with a radiator, a boot bench or a fire extinguisher in the latch-side space. Nothing on a plan makes that space look occupied, because it is drawn as floor.

Door hardware and closing force belong to the same set. A door with a compliant clearance and a closer adjusted too heavily is closed to some of the people the clearance was provided for, and closing force is a maintenance quantity that drifts rather than a fixed one.

Revolving doors are the sharpest case, because they are never an accessible route in themselves. Their footprint calculation sizes the enclosure; the compliant route is the adjacent swing or sliding door, and the revolving door's own dimensions say nothing about whether the entrance works.

Parking is sized by the vehicle that turns worst

A parking layout is a geometry problem with one dominant input: the DESIGN VEHICLE. Aisle width, stall depth, corner radii and the whole efficiency of a layout follow from the turning envelope of the largest vehicle that must use it, and that is rarely a car.

The vehicle that governs is usually a service one — a refuse lorry, a delivery vehicle, a fire appliance — and its swept path is much wider than its body, because the rear wheels track INSIDE the front ones through a turn. That difference between the front outer and rear inner paths is the swept width, and it is what clips kerbs and bollards that were set out from the vehicle's dimensions rather than from its path.

Angle changes the arithmetic in a way that is easy to get backwards. Angled stalls need a narrower aisle and are easier to enter, but they consume more depth per stall and usually force one-way circulation. Ninety-degree stalls are the most efficient per unit of area and demand the widest aisle. Which wins depends on the site's shape rather than on either being better.

Accessible stalls are not a percentage applied to the total either. The requirement is a stepped schedule — so many accessible stalls up to a given total, then a further one per band — with a proportion of them van-accessible at greater width, and with a requirement that they be on the SHORTEST accessible route to the entrance. A layout that provides the right count in the far corner has met the number and failed the requirement.

A queue is not linear, and an average is the wrong input

The length of a driveway throat, the stacking space before a barrier, and the reservoir at a car park entrance are queueing problems, and queues do not behave proportionally.

As the arrival rate approaches the rate at which vehicles can be served, the expected queue grows toward infinity rather than toward some comfortable multiple. The practical consequence is that utilisation around eighty per cent is roughly where queue length starts to rise steeply, and a facility designed at ninety-five per cent utilisation has been designed at the point where small variations produce large queues.

Which means an AVERAGE arrival rate is the wrong input. Arrivals are not evenly spaced, and the peak fifteen minutes of a school run or a shift change can be several times the hourly average. A throat sized on the average overflows on to the public road during every peak, which is exactly the outcome the requirement exists to prevent.

The consequence of getting it wrong falls outside the site. A queue that exceeds the throat backs into the carriageway, so the failure mode is a highway safety problem rather than an inconvenience — which is why highway authorities specify minimum stacking distances rather than leaving it to the developer's judgement.

Detectable warnings and tactile surfaces are information, not texture

A tactile paving surface is a message, and the different patterns mean different things: blister surfaces warn of a crossing point where the footway meets a carriageway, corduroy warns of a hazard such as a stair, and directional surfaces guide a route.

Because they are a language, using the wrong one is worse than using none. A blister surface laid where there is no crossing tells a blind pedestrian that it is safe to step into a road, which is precisely the hazard the system exists to prevent.

Their geometry is prescriptive for the same reason. The depth of the field, its alignment with the crossing, the colour contrast against the surrounding paving and the offset from the kerb are all specified, because the surface has to be recognisable under a foot and a cane at walking pace by someone who is not looking for it.

The area calculation here is therefore a quantity for a specified layout rather than a design of that layout. The pattern, its extent and its position come from the relevant standard and from the crossing's geometry, and the page returns how much of it to order.

Shafts, wells and the space around the machine

A lift is not a car in a hole. The hoistway has to accommodate the car and its frame, the counterweight and its guides, the rails and their brackets, the running clearances on every side, the OVERRUN above the top landing and the PIT below the bottom one — and the pit and overrun are safety spaces that exist to stop the car or the counterweight striking anything if it overtravels.

Those two are the dimensions most often squeezed in a refurbishment, and they are the two that cannot be. Machine-room-less designs removed the separate machine room but did not remove the overrun; a building with insufficient headroom above the top floor cannot take a conventional traction lift however the rest of the shaft is arranged.

An escalator's wellway opening is the same class of problem in a different shape. The opening is set by the truss geometry, which follows from the rise and the angle, plus the flat steps at each end and the balustrade — so the structural opening is considerably longer than the visible run, and it is fixed before anything else about the floor can be laid out.

Both are cases where the calculator returns a planning envelope rather than a specification. The manufacturer's actual dimensions govern, they differ between products, and a shaft built to a generic figure can fail to accept the unit eventually procured — which is why the envelope is agreed with a supplier before the structure is built rather than after.

Devices on a grid: listed spacing is not a radius

Smoke detectors and electrical outlets are both spaced by rule rather than by calculation, and both rules are more subtle than the headline number.

A smoke detector's listed spacing describes a SQUARE grid, not a circle. The common thirty-foot figure means detectors thirty feet apart, and the point farthest from any of them — the corner of a grid square — is about twenty-one feet away, 0.7 times the spacing, which is how NFPA 72 writes the same rule as a radius. BS 5839-1 writes it as a radius directly: 7.5 m (24 ft 7 in) to the nearest point smoke detector, 5.3 m (17 ft 5 in) to a heat detector.

Ceiling shape then modifies it. Beams, joists, sloped ceilings and high spaces all change how smoke travels and reduce the permitted spacing, sometimes substantially, and a flat-ceiling figure applied to a beamed or vaulted room is a real under-provision.

Receptacle spacing follows a different form: the requirement is that no point along a wall line is more than a stated distance from an outlet, which produces a maximum SEPARATION of twice that distance between adjacent outlets and requires one within half of it from each doorway or wall end. Counting outlets by dividing a wall length misses both ends, and the ends are where furniture and extension leads accumulate.

In both cases the arithmetic is a count over a run — the fastener paper's plus-one problem — and the value that governs comes from a listing or a code rather than from the geometry.

The count follows from the corner, not from an area. Circles do not tile, so a floor divided by a detector's circle buys too few; the right count is the smallest grid whose cells keep their corners inside the radius — (sx ÷ 2)² + (sy ÷ 2)² no more than R² — and the grid need not be square, which saves detectors on a long, narrow floor. A corridor under 2 m (6 ft 7 in) wide is covered along its length alone under BS 5839-1.

A tree's protected ground is set from its stem, not its crown

BS 5837 draws a tree's root protection area as a circle whose radius is twelve times the stem diameter measured at 1.5 m (4 ft 11 in), capped at 707 m² (7,610 sq ft); several stems combine as the square root of the sum of their squares. AS 4970 starts from the same twelve times, measured at 1.4 m (4 ft 7 in) and held between 2 m and 15 m (6 ft 7 in and 49 ft), and adds a structural root zone for the tree's stability.

Like every figure in this paper it is an accepted value, not a derived one: the standards' proxy for the roots a tree cannot afford to lose. The arboriculturist may reshape the circle to follow the roots actually found, keeping its area, and the fencing follows the report rather than the arithmetic.