Shelf span, thickness and sag

Every shelf in a house is a simply supported beam carrying a uniform load, which means the amount it droops is not a matter of taste or luck. It is four variables, and they behave the way every other number in storage and shelving does, which is predictably: how long the span is, how thick the board is, how stiff the material is, and how heavy the load is. Only one of those four is under real control after the shelf is built, and it is the one people reach for last.

What "span" means, and the two ways people measure it wrong#

Span is the clear distance between the supports, face to face, not the length of the board, and it is worth taking the measurement the way how to measure a room properly describes: once, written down, from a named datum. A 36 in (91 cm) board sitting in a bookcase with 3/4 in (18 mm) sides has a span of 34.5 in (88 cm), and that is the number the arithmetic uses.

Two mistakes recur. The first is measuring the outside of the case, which overstates the span by the thickness of two sides and makes a marginal shelf look worse than it is. The second, much more expensive, is counting a support that is not one. A shelf resting on five spoon-shaped pin supports along its back edge is supported at the ends only: the pins carry load but do not resist rotation, and a back cleat screwed to the wall behind a shelf does nothing for sag unless the shelf is fixed to it along its whole length. If the shelf can lift off the support, that support is holding weight, not stiffness.

Where the numbers come from#

The deflection of a uniformly loaded shelf is

sag = 5 x w x L^4 / (384 x E x I)

with w the load per unit length, L the span, E the modulus of elasticity of the material, and I the second moment of area of the section, which for a rectangle is b x d^3 / 12 where d is the thickness.

Two exponents do all the work. Span appears to the fourth power, and thickness appears to the third. Everything else is linear.

That has consequences worth stating in plain numbers:

• Increasing the span from 30 to 36 in, a 20 per cent change, increases sag by 107 per cent. It roughly doubles.
• Increasing the span from 30 to 48 in, a 60 per cent change, increases sag by 555 per cent. It goes up more than six times.
• Increasing thickness from 3/4 in to 1 in, a 33 per cent change, reduces sag by 58 per cent. It more than halves it.
• Doubling the depth of the shelf halves the sag for the same total load, but a shelf twice as deep normally holds twice as much, so in practice depth is close to neutral.

This is why span is the variable to fix and depth is the variable to ignore. It is also why the standard advice to "use a thicker board" is right for the wrong reason: the gain is not from strength, it is from the cube.

Stiffness by material#

E is a property of the material, and the spread across the boards sold for shelving is larger than most people expect.

Chipboard is around a quarter as stiff as oak, which is the whole explanation for why flat-pack shelving droops and a joiner's shelf does not. Note the last row: glass is enormously stiff but brittle, so its span is set by breaking strength rather than by sag, and the working limits come from the supplier rather than from this table.

The plywood figure assumes the face grain runs along the span. Turned the other way, a plywood shelf loses roughly half its stiffness, and it is a common error in site-cut shelving because it is the orientation that wastes less sheet.

Load: what is actually going on the shelf#

Sag is linear in load, so a record shelf carrying two and a half times the load of a book shelf sags two and a half times as much. That single line means every span in the table above halves for vinyl once the fourth-power relationship is worked back through: about 24 in (61 cm) in plywood, and under 20 in (51 cm) in melamine. Record storage is the one domestic case where the usual shelving rules are simply not conservative enough, which is covered further in the guide to book, vinyl and media storage dimensions.

The four ways to buy back a span#

In order of how much they gain per unit of effort.

1. Add a support. Sag scales with the fourth power of span, so halving the span cuts the sag to one sixteenth. A single central divider turns a hopeless 48 in (122 cm) melamine shelf into two comfortable 23.5 in (60 cm) ones. In a bookcase this is a vertical divider; on a wall it is a third bracket.
2. Add a front edge stiffener. A 1.5 to 2 in (38 to 51 mm) strip glued to the front edge on edge, so its depth is vertical, raises the effective thickness of the section dramatically because that dimension is cubed. A 3/4 in shelf with a 2 in front rail behaves closer to a 2 in shelf across most of its width, and it typically doubles the workable span. It also reads as a thicker, more expensive shelf, which is why cabinetmakers have done it for centuries.
3. Go thicker. Moving 3/4 in to 1 in (18 to 25 mm) removes 58 per cent of the sag. Moving to 1.25 in (32 mm) removes 78 per cent.
4. Change material. Swapping melamine chipboard for birch plywood at the same thickness cuts sag by roughly two thirds, and it is often the cheapest change of the four if the shelf is being made rather than bought.

The shelf span and sag calculator runs the same arithmetic for any combination of the four variables, including the creep factor below.

Turning the board on edge, adding a back stiffener, or screwing the shelf to a wall cleat along the back all help far less than people hope, because they do not add depth to the section in the direction that is being bent.

Fixed shelves, adjustable shelves and brackets#

Adjustable shelves on pins are the weakest common arrangement, because the shelf simply rests on four points and can lift. Fixed shelves housed into a dado groove in the sides gain a small amount of end restraint, which reduces mid-span sag by roughly 20 per cent relative to a simply supported shelf, and they stiffen the whole case against racking. If a run is at the edge of its span, housing the shelves is worth more than it looks.

Wall brackets change the arithmetic in a useful way, because a bracket set in from the end creates a cantilever that partly balances the sag between them. Setting brackets in from each end by about one fifth of the total board length roughly halves the mid-span deflection compared with brackets at the very ends, and it is free. The overhang beyond the bracket should not exceed about 8 in (20 cm) for a 3/4 in board, or the end of the shelf becomes the part that droops.

Bracket spacing has its own rule that is independent of sag: no bracket should be more than 32 in (81 cm) from the next in a stud-framed wall, because that is two 16 in (406 mm) stud bays, and a bracket that misses a stud is carrying nothing but plasterboard. Stud positions and the depths involved are covered in TV wall mount dimensions, VESA and stud spacing.

Depth, spacing and the rest of the shelf#

Depth barely affects sag, but it decides whether the shelf is usable at all: 9 to 12 in (23 to 30 cm) for general books, 14 in (36 cm) for art books, and 7 in (18 cm) for paperbacks alone. Vertical spacing follows the same logic, and both are set out in full in shelf depth and vertical spacing by content. The heights at which a shelf is reachable at all are a separate constraint again, covered in reach zones and useful storage heights, and the standard carcass dimensions that most shelving is built from appear in built-in cabinetry dimensions and bookcase dimensions.

One number ties the two together. A bay 30 in (76 cm) wide and 12 in (30 cm) deep, with shelves at 12 in (30 cm) vertical centres, holds about 25 linear feet (7.6 m) of books per 8 ft (2.4 m) of height, which is roughly 300 hardbacks and about 500 lb (227 kg) delivered into the floor at two points. That is worth knowing before a tall bookcase goes on an upper floor or on a floating floor, and it is the sort of load figure the standards and codes reference exists to pin down.