The beam formula, worked all the way through
A shelf is a beam carrying a distributed load. For a uniformly loaded beam resting freely on supports at each end, the deflection at mid-span is:
δ = 5 w L⁴ / (384 E I)
where w is the load per unit length, L is the span, E is the modulus of elasticity of the material and I is the moment of inertia of the cross section. For a rectangular board, I = b h³ / 12, with b the depth of the shelf front to back and h the thickness.
Units have to be consistent or the answer is nonsense. Everything here is in pounds, inches and psi: w in pounds per inch, L in inches, E in pounds per square inch, I in inches to the fourth. If you enter the load in pounds per foot, divide by 12 before it goes into the formula. That single division is the most common arithmetic mistake people make with this equation, and it produces an answer twelve times too large.
The worked case
A red oak shelf, 36 inch span, 10 inches deep, three-quarters of an inch thick, carrying 20 pounds per running foot of hardback books, sitting on shelf pins at each end.
- E for red oak = 1,820,000 psi
- I = b h³ / 12 = 10 × 0.75³ / 12 = 10 × 0.421875 / 12 = 4.21875 / 12 = 0.3515625 in⁴
- w = 20 lb/ft ÷ 12 = 1.6667 lb/in
- L⁴ = 36⁴ = 36 × 36 × 36 × 36 = 1,679,616 in⁴
- Numerator: 5 × 1.6667 × 1,679,616 = 8.3333 × 1,679,616 = 13,996,800
- Denominator: 384 × 1,820,000 × 0.3515625. First 1,820,000 × 0.3515625 = 639,843.75, then × 384 = 245,700,000
- δ = 13,996,800 / 245,700,000 = 0.0570 inches
Just under a sixteenth of an inch on the day it is loaded. Against a span/360 criterion the limit is 36/360 = 0.100 inches, so it passes with room. Apply a creep factor of 1.5 for solid wood under permanent load and the long-term figure is 0.085 inches, which still passes, but not by nearly as much as the first number suggested.
Now change one thing. Swap the red oak for particleboard at E = 500,000 psi and everything else is identical, so the deflection scales by the ratio of the stiffnesses: 0.0570 × 1,820,000 / 500,000 = 0.207 inches on day one, already twice the span/360 limit, and with the creep factor of 3 that particleboard deserves it heads towards six-tenths of an inch. This is not a theoretical result. It is the sagging melamine bookshelf in everyone's spare room.
Creep, and why the day-one number lies
Wood and wood composites are viscoelastic. Under a load they hold continuously, they deflect immediately by the elastic amount the formula gives, and then keep deflecting slowly for years afterwards. The extra is called creep, and unlike the elastic part it does not come back when you take the books off.
| Material | Creep factor used | Behaviour |
|---|---|---|
| Seasoned solid wood | 1.5 | Standard long-term allowance for dry lumber under sustained load |
| Plywood | 2.0 | Glue lines and cross plies creep more than the wood does |
| MDF, particleboard, melamine | 3.0 | Poor, and worse in humidity; a working estimate, not a constant |
The 1.5 for seasoned solid wood is the long-established allowance used in timber design. The 2.0 and 3.0 figures for panel products are working estimates drawn from the general observation that composites creep considerably more, and they move with humidity in a way that no single multiplier captures. A particleboard shelf in a damp basement can exceed a factor of 3 comfortably.
The verdict this calculator gives is checked against the long-term figure rather than the day-one one, which is stricter than most published shelf calculators, and deliberately so. Nobody has ever complained about a shelf that did not sag.
Stiffness values, and how much to trust them
These are modulus of elasticity figures in psi for clear, straight-grained material at 12 percent moisture content, which is the basis published shrinkage and stiffness tables use.
| Material | E (psi) | Realistic range |
|---|---|---|
| Hickory | 2,160,000 | The stiffest common domestic hardwood |
| Yellow birch | 2,010,000 | 1.9-2.1 million |
| Douglas-fir | 1,950,000 | 1.6-2.0 million, grade dependent |
| Hard maple / red oak / white oak / ash | 1.74-1.83 million | Very close to each other in practice |
| Walnut, soft maple | 1.64-1.68 million | |
| Cherry, poplar, teak, mahogany | 1.49-1.58 million | Mahogany and teak vary widely by source |
| Pine, spruce, cedar | 1.11-1.29 million | Southern yellow pine is the outlier at 1.8 |
| Plywood, face grain along span | 1,300,000 | 1.1-1.6 million by species and ply layout |
| Plywood, face grain across span | 500,000 | A third of the stiffness — orientation matters enormously |
| MDF | 450,000 | 350,000-520,000 |
| Particleboard | 500,000 | 350,000-600,000, brand and grade dependent |
Two cautions. First, graded construction lumber is not clear wood: knots, sloping grain and checks reduce stiffness, and published design values for graded lumber are lower than the clear-wood figures above, sometimes by a third. Second, sheet goods vary between manufacturers far more than species vary between each other. Two sheets of three-quarter inch particleboard from different mills can differ by 40 percent in stiffness, and neither will say so on the label.
Note also that plywood with the face grain running across the span is a different material from the same sheet turned ninety degrees. Cutting shelves the wrong way out of a sheet costs you roughly two-thirds of the stiffness, and it is an easy mistake to make when you are optimising a cut list for yield.
What to do when it fails
In order of how much difference they make:
- Reduce the span. Deflection goes with the fourth power of span. Cutting a 36 inch span to 30 inches — one extra divider — cuts the sag to 48 percent of what it was. Nothing else comes close.
- Add thickness. Stiffness goes with the cube of thickness. Three-quarter inch to one inch is 2.4 times stiffer. Three-quarter to an inch and a half is eight times.
- Add a lip. A solid strip glued to the front edge, standing proud below the shelf, moves material away from the neutral axis and is dramatically effective for the wood it uses. A three-quarter by one and a half inch lip on the front of a plywood shelf can more than double its stiffness.
- Change material. Going from particleboard to hardwood is a factor of about 3.6 in E, and a factor of 2 in creep on top of that. Going from cherry to hard maple is a factor of 1.2 and rarely worth it on its own.
- Increase depth. Linear only. A 12 inch shelf is 20 percent stiffer than a 10 inch one, and it will also carry 20 percent more books, so this one largely cancels itself out.
The end condition is the one lever people over-claim. A shelf glued into stopped dados in a rigid case is genuinely stiffer than one on pins, and the textbook fixed-fixed case is five times better. Real casework does not reach that, because the case sides flex and rotate rather than holding the shelf ends rigidly. If the calculator's free-end answer passes, build it and stop worrying. If only the fixed-end answer passes, you are relying on a boundary condition your joinery probably will not deliver.
Questions people ask
How much do books actually weigh?
More than people expect, and it depends on the format. Hardbacks on a 10 inch deep shelf run about 20 to 25 pounds per running foot. Mass-market paperbacks are lighter at 12 to 15. Large-format art books, textbooks and bound journals push 30 to 35. Vinyl records are the heaviest thing most people put on a domestic shelf at roughly 45 to 55 pounds per foot, which is why record shelves that were built like bookshelves fail. Canned goods in a pantry are 35 to 40. If the shelf is a display shelf for objects, weigh a representative foot of them rather than guessing.
Why does my shelf keep sagging even though it passed the calculation?
Almost certainly creep. The standard beam formula gives the elastic deflection that happens the moment the load goes on, and that part is recoverable. Under a load carried continuously for months and years, wood and especially wood composites keep deforming, and that part is not recoverable. A shelf that measured a sixteenth of an inch on day one can measure three-sixteenths after two years, and once bowed it stays bowed even when emptied. Humidity accelerates it, and manufactured panels are far worse than solid wood. This calculator applies a creep multiplier and checks the verdict against the long-term number for exactly this reason.
Is plywood or solid wood stiffer for a shelf?
Solid hardwood is stiffer than ordinary plywood of the same thickness — around 1.8 million psi for oak against about 1.3 million for good plywood with the face grain running the right way. But plywood is more dimensionally stable, does not need a wide glue-up, and holds its shape better across a wide shelf, and Baltic birch closes much of the stiffness gap. The orientation is not optional: plywood with the face grain running across the span rather than along it loses roughly two-thirds of its stiffness, so a shelf cut the wrong way from a sheet performs like a completely different material.
What sag is actually noticeable?
The eye picks up deflection at somewhere around span divided by 180 to 200 on a horizontal line, sooner if there is a straight reference nearby like a case rail or a door edge below it. Span over 360 is the usual furniture standard and reads as flat. Span over 240 is visible if you look for it and invisible if you do not. A useful alternative rule is a maximum of 1/32 inch of sag per foot of span, which works out to about span over 384 and is the fussiest of the common criteria. Pick the standard by where the shelf is: a garage shelf can sag visibly and nobody minds, a shelf at eye level in a living room cannot.