Stepwise.

Shelf Sag Calculator

Work out how far a shelf will bend under load from its span, thickness and material, and what a front lip or a shorter span would do about it.

How to use this calculator

  1. 1Enter the clear span between supports rather than the total shelf length - it is the unsupported distance that bends.
  2. 2Pick the real material. The difference between Baltic birch and melamine-faced particleboard is a factor of four in stiffness, and they look similar on the rack.
  3. 3Choose a load preset, or weigh a representative foot of what is going on the shelf and enter the total.
  4. 4If the answer fails, try shortening the span first and adding a front lip second. Thicker material is the expensive way to solve it.

How the calculation works

Deflection = 5 W L^3 / (384 E I) I (plain shelf) = b t^3 / 12 I (with front lip) = sum over both parts of [ I_own + A (y_bar - y)^2 ], the lip transformed by E_lip / E_shelf
W
Total load on the shelf in pounds, spread evenly along it - contents plus the shelf itself
L
Clear span between supports, in inches
E
Modulus of elasticity of the shelf material, in psi. Hard maple is about 1,830,000; MDF about 450,000
I
Second moment of area of the cross-section, in in^4
b, t
Depth front-to-back and thickness of the shelf, in inches
y_bar
Distance from the top face to the neutral axis of the composite section

The load appears as the total rather than as a load per unit length, which is why the span is cubed here rather than raised to the fourth power. Hold the load per foot constant instead and the total rises with the span, restoring the fourth power - that is the version that matters when you are choosing where to put the supports.

The front lip is handled by the transformed-section method: the lip is converted into an equivalent width of shelf material in proportion to the ratio of the two moduli, then both parts are taken about the combined neutral axis. This is exact for a rigid glue line, which a properly clamped wood joint is.

A centre point load of the same total weight deflects 1.6 times as much as the same weight spread evenly, since the point-load case is P L^3 / 48 E I. If the shelf carries one heavy object, treat this page as optimistic.

Worked example

The classic failure: 36 in of melamine holding books

  1. 1.A 36 in span of 5/8 in melamine-faced particleboard, 11-1/4 in deep, carrying mixed books at about 22 lb per foot - so 66 lb of books plus the shelf itself.
  2. 2.Second moment of area is 11.25 x 0.625^3 / 12 = 0.229 in^4, and the modulus is only 350,000 psi.
  3. 3.Deflection is 5 x 72.9 x 36^3 / (384 x 350,000 x 0.229) = 0.553 in, against a threshold of 0.06 in.
  4. 4.Nine times over. And creep roughly doubles it, which is why these shelves end up visibly bowed rather than merely soft.

Result: About 0.55 in of sag - roughly nine times the visible threshold

The same span in Baltic birch with a front lip

  1. 1.Same span and same books, but 3/4 in Baltic birch at 1,600,000 psi, with a 3/4 in by 1-1/2 in hard maple lip glued along the front edge.
  2. 2.The lip transforms to 1,830,000 / 1,600,000 = 1.14 times its own width in birch, and sits well below the shelf, so it drags the neutral axis down and adds an A x d^2 term far larger than its own bending stiffness.
  3. 3.Second moment of area rises from 0.396 in^4 for the bare shelf to 2.05 in^4 - more than five times.
  4. 4.The deflection lands at 0.014 in - a quarter of the threshold, on the same span that failed by a factor of nine.

Result: 0.014 in - the lip does more than the material change

How far can 3/4 in oak span?

  1. 1.Solid red oak, 3/4 in thick and 9-1/4 in deep, loaded with hardcovers at 28 lb per foot over a 42 in span.
  2. 2.That is 98 lb of books plus about 7 lb of shelf.
  3. 3.The "longest span that stays under it" figure is the useful output here - it solves the same equation backwards for the span at which deflection just reaches 0.02 in per foot.
  4. 4.It comes out near 31 in, so 42 in is well past what this shelf will carry without showing it.

Result: Around 31 in is the limit - 42 in sags visibly

Why span matters more than anything else

Deflection under a distributed load goes as the fourth power of the span once the load per foot is held constant. That is a brutally steep relationship, and it is not intuitive. A shelf 20% longer sags about twice as much. A shelf twice as long sags sixteen times as much.

Everything else is gentler. Thickness is cubed, so doubling it divides sag by eight - real, but you rarely have the option to double a shelf. Depth front-to-back is only linear, which is why making a shelf deeper barely helps: you add load capacity and stiffness in roughly equal measure and end up close to where you started. Material stiffness is linear too, and the whole range from particleboard to hickory is only a factor of six.

The practical consequence is that the cheapest fix is nearly always another support. A bookcase 6 ft wide with one centre divider is not twice as stiff as one without; it is around sixteen times as stiff. Cabinet makers who space their verticals at 30 to 32 in are not following a style convention, they are staying on the flat part of that curve.

The front lip, and why it works so well

Bending stiffness comes overwhelmingly from material far from the neutral axis, because each element contributes in proportion to the square of its distance from it. In a plain rectangular shelf the neutral axis sits at mid-thickness, so the wood near it - most of the shelf - is barely earning its keep.

A strip glued along the front edge and standing below the shelf changes this twice over. It adds material at a distance, and it pulls the neutral axis down towards itself, which lengthens the lever arm for the shelf above. A 3/4 in by 1-1/2 in strip on a 3/4 in shelf can more than triple the second moment of area while adding about a tenth to the weight.

The joint has to be a real glue line for this to hold - the transformed-section calculation assumes the two parts act as one piece and do not slide against each other. A glued, clamped edge joint does that. A strip pinned on with brads does not, and behaves closer to two independent beams.

  • Front lip — Cheapest and most effective per dollar. Reads as a thicker shelf from the front, which most people prefer anyway.
  • Torsion box — Two thin skins over a light grid. Enormous stiffness for the weight, and the reason a hollow-core door will span further than a solid shelf twice its mass.
  • Rear cleat — Helps far less than a front lip, because the back edge of a loaded shelf is usually already close to a support.
  • Metal insert — A steel bar let into a groove in the front edge. Common in commercial casework where a wooden lip would be too deep.

Creep: the sag that arrives later

Wood and wood-based panels are viscoelastic. Load one and it deflects immediately, and then it goes on deflecting slowly for years, taking a permanent set that does not come back when the load is removed. This is creep, and it is the reason a shelf that looked acceptable when it was built is bowed a decade later.

Panel products creep hardest. MDF and particleboard under sustained load can reach roughly twice their initial deflection; solid wood is better behaved at around one and a half times. Humidity makes it worse - a shelf that cycles between damp and dry summers and winters creeps considerably faster than one held at steady conditions, because each moisture cycle lets the fibres slip a little further.

The design response is simply to include it. A shelf sized so that its initial deflection is at the visible threshold has already failed; the number to keep under the threshold is the long-term one. That is why this page reports both, and why a marginal pass on the first figure should be read as a fail.

What this assumes, and where it stops

Assumptions

  • The load is spread evenly along the shelf and the shelf is simply supported at both ends - free to rotate at the supports, not clamped. Real shelves in dadoes are slightly stiffer than this, so the result errs safe.
  • Stiffness figures are published class averages at about 12% moisture content. Individual boards and individual sheets vary widely around them.
  • Any front lip is taken as glued to the shelf along its full length so that the two act as a single section.
  • Creep factors are approximate multipliers on the initial deflection - roughly 2 for panel products and 1.5 for solid wood - and depend heavily on humidity cycling.

Limitations

  • A single heavy object in the middle deflects about 1.6 times as much as the same weight spread out. This page assumes an even load, so a shelf carrying one aquarium or one amplifier is worse than the number here.
  • The calculation is for stiffness, not strength. A shelf can be well inside its deflection limit and still be nowhere near breaking, and a badly overloaded one may fail at the supports rather than in bending.
  • It does not model shelves fixed rigidly at both ends, cantilevers, or shelves supported along the back edge as well as at the ends. All three behave differently.
  • Adjustable shelf pins concentrate load at four points and can crush into a particleboard carcase long before the shelf itself becomes the problem.

Common questions

How much sag is acceptable?

About 0.02 in per foot of span is the woodworking threshold - below it the eye reads the shelf as straight, above it people see the curve. On a 32 in span that is a little over 1/16 in. The structural convention of span over 360 is stricter on short spans and looser on long ones; for shelving the per-foot rule matches what actually looks acceptable. Whichever you use, apply it to the long-term figure rather than the initial one.

Is thicker material or a shorter span the better fix?

Shorter span, almost always. Thickness is cubed but you can rarely change it by much - going from 3/4 in to 1 in is a 30% change that buys a factor of 2.4. Span is to the fourth power and you can often change it by a third simply by adding a divider, which buys a factor of five or more. A front lip beats both on cost.

Why do flat-pack bookshelves sag so badly?

Because melamine-faced particleboard is around 350,000 psi against 1,800,000 for hard maple, it is usually 5/8 in rather than 3/4 in, the spans are set by the width of the box rather than by what the shelf can carry, and the material creeps hard. Every one of those choices is a cost decision and all four push the same way. The shelves are not badly made; they are made to a price that the physics does not respect.

Does the finish or the edge banding stiffen a shelf?

Iron-on veneer edge banding does essentially nothing - it is a few thousandths of an inch of material with no depth, so it adds no meaningful second moment of area. Finish does nothing structurally either, though sealing all faces helps by slowing the moisture cycling that accelerates creep. A solid wood lip is a different thing entirely, and this page models it.

What about a shelf supported at the back as well as the ends?

It helps, but far less than it looks like it should. A cleat along the back edge carries load close to where the carcase already supports it, and the front edge - which is where the deflection is - is unaffected. If you can only add support in one place, add it at the front.

Sources

Formula and content last reviewed on .

Results are estimates for information only, not professional advice.

Report an error

Related calculators

Tools people commonly use alongside the shelf sag calculator.

See all science & engineering calculators →