Checking an Exposed Timber Beam

The Bonus That Belongs to the Shortest Load

Timber will carry more stress for a short time than it will carry forever, and the ANSI/AWC NDS puts that into design as a multiplier keyed to the cumulative duration a load acts. Table 2.3.2 sets the values: the lowest for permanent load, the highest for impact, and the snow entry sitting where a two-month accumulation belongs. It is not a discretionary uplift to be spent on a marginal result. It is an attribute of the load case being examined, and it belongs to that case only.

In a combination, the factor comes from the shortest-duration action present — dead plus snow is checked at snow's value, not at dead's. What gets skipped is the other half of that rule: the same member also has to be checked under permanent load alone, at the permanent factor, with no snow in the combination at all. Arithmetic decides which of the two governs, and permanent load wins whenever it is several times the snow figure. That is not exotic. A planted roof, a thick sand-cement topping, a heavy paver terrace on pedestals over a light snow region — the dead-only case governs on all of them, and the sheet that only ever shows dead plus snow has never tested it.

The Eurocode reaches the same physics down a different road, and mixing the two produces a number belonging to neither. EN 1995-1-1 modifies strength with a factor read off a grid of service class against load-duration class, so the answer depends on both how long the load acts and how wet the beam lives, and it divides by a material partial factor afterwards. Which duration class snow falls into is assigned by the National Annex rather than by the base standard, and it can differ by altitude within one country. Look it up for the jurisdiction the building is in rather than carrying it from the last project.

Two design values sit outside this multiplication entirely, and both are about to matter. Stiffness is one — the modulus used for deflection is unaffected by how long the load lasts, which is why the serviceability check gets none of this benefit. Crushing across the grain is the other, so the bearing at each end of this beam earns nothing from the fact that its governing load is snow. Beyond that, the duration factor is one link in a chain that also includes wet service, temperature, size or volume, lateral stability, flat use, incising and repetitive action. A figure with only the duration factor applied is a step in a calculation, not an allowable stress.

Write the combination next to the factor on the sheet. A reviewer who sees the snow multiplier printed with no named load case beside it cannot tell whether it was earned or assumed, and the reasonable response to that is a question rather than a signature.

The Ambient Sheet, on the Whole Section

With the adjusted value settled, the ambient check is short. The moment comes from the load and the span — the simple-span expression if the beam really is simply supported, which an entrance beam picking up a post from the storey above is not. The section modulus of the full 240 by 640 is just under 16.4 million cubic millimetres, and the bending stress is the moment divided by it. Everything interesting happens in the two adjustments that are specific to glued laminated timber rather than in that division.

The first is that glulam is graded as a manufactured beam, not as a piece of timber, so a large member is penalised for its volume while a poorly restrained one is penalised for its slenderness — and the NDS applies whichever of those two reductions is worse, not both stacked. Whether the second one bites depends on a site fact rather than a drawing: a CLT deck screwed down into the top face restrains the compression edge continuously, a deck merely bearing on it does not, and a beam with a rooflight slot running beside it may have long unrestrained stretches nobody meant to create. Settle that before choosing a stability factor, because it is the difference between a comfortable pass and a section change.

The second is orientation. A glulam beam is laid up with its best laminations at the extreme fibres, and unbalanced layups put the good material only in the tension face — which is why those beams carry a TOP mark from the plant, under EN 14080 in Europe and ANSI A190.1 in North America. A beam installed upside down is not the beam the design values describe. Neither is a beam run continuous over an intermediate support, where the tension fibre at the support is the top one; that condition needs a balanced layup specified at order stage, not discovered on the deck.

1. Fix the load combination and the duration class first, and name it on the sheet.
2. Take the moment from the real support condition, including any point load handed down from above.
3. Work the section modulus of the delivered section, not the nominal size on the architect's plan.
4. Apply the full adjustment chain, taking the worse of the volume and stability reductions rather than both.
5. Repeat the whole check for permanent load alone at its own factor.
6. Only then start the fire sheet, because it needs the section this one settled on.

Stiffness Earns Nothing From Duration

The deflection check gets no benefit from the snow being short. Stiffness is excluded from the duration adjustment, so the beam is exactly as flexible under a two-month load as under a permanent one, and the number that comes out of the elastic calculation is the number. What duration does change is creep, handled separately: the NDS multiplies the long-term component of deflection before adding the short-term part, and EN 1995-1-1 does the equivalent with a deformation factor tied to service class, producing a final deflection rather than an instantaneous one. Snow contributes little to that; the deck, the topping and the beam's own weight contribute all of it, for the life of the building.

Over glazing the limit that matters is rarely the code one. A roof beam satisfying the usual ratio can still drop enough to close a movement gap onto a glazed head, and the glazing supplier publishes a head deflection allowance that is normally tighter than anything in the structural code — that allowance is the governing figure, and it is a contractual one. Beyond the arithmetic there is the line itself: a beam running above a long horizontal transom is compared directly against it by every visitor, and precamber is the only cheap answer. Camber is manufactured in and marked, so it has to be on the order rather than in the hope, and a cambered beam set the wrong way up is a defect that arrives on the lorry.

An Hour Off the Outside

The fire sheet starts by rebuilding the member. Under the reduced cross-section method of EN 1995-1-2, a notional charring rate for the product and its characteristic density is multiplied by the required period to give a char depth, and a further zero-strength allowance is deducted behind it to account for timber hot enough to have lost its stiffness without having burned. Both the rate and the allowance are read from the standard being designed to. The NDS gives the American equivalent in its Chapter 16, with a different charring model and a different treatment of the heated zone, and AWC Technical Report 10 sets that method out in full. A rate borrowed from one document and an allowance borrowed from the other produce a residual section that belongs to neither.

Then count faces, one dimension at a time. This beam sits under a deck, so the fire reaches its soffit and both cheeks but not its top: the width loses material from two sides, the depth from one. Take a rate of 0.7 mm per minute and a 7 mm allowance — illustrative values, not a licence to skip the table — and each exposed face costs 49 mm at sixty minutes. The 240 width becomes 142. The 640 depth becomes 591, because only the soffit is exposed.

The consequence is worse than those numbers look, because section modulus goes with the square of the depth. The residual 142 by 591 gives about 8.3 million cubic millimetres against the 16.4 million you started with — the beam has lost roughly half its bending capacity and a little under half its area in an hour, while looking, from the outside, like a slightly smaller beam. That is the single figure worth carrying out of this page.

The load side moves the other way, which is the only reason any of this works. Fire is an accidental situation, so the combination is not the strength one: companion actions come in at reduced factors under EN 1990 or under the extraordinary-event provisions of ASCE/SEI 7, and much of the snow that governed the ambient sheet is simply not in it. The material side moves too. The reduced cross-section method pairs the smaller section with a material factor of unity, carrying the whole strength reduction in the geometry rather than applying a second penalty on top; the NDS route converts reference design values toward average member strength before the fire check, which means the adjusted stress from your ambient sheet cannot be reused here. Fire design is not the ambient calculation with a smaller beam typed into it.

Two things belong on the sheet at the end. One is the residual dimension itself, with the face count it assumed, because a beam later exposed on a fourth side by a demolished partition is a different calculation. The other is the headroom between that residual and the smallest section the check still passes at, expressed as a thickness that may come off the visible face over the life of the building — the number a facilities manager needs before anyone sands a water stain out of the soffit. And the whole notional model rests on the member holding its geometry while it burns, which a beam whose adhesive line opens under heat does not do: where a lamination can fall away the exposed surface renews itself, the rate ceases to be the tabulated one, and the standards send that product down a different route entirely.

Where the Section Was Already Missing

Everything above assumed a rectangle. Exposed beams rarely stay rectangular: a notch at the bearing to drop the soffit level with a steel beam beside it, a birdsmouth where a rafter lands, a rebate for a blind box, a hole drilled for a sprinkler drop during second fix. A notch cut into the tension face at the end of a member does not merely reduce the depth in proportion; the NDS handles it at 3.4.3 by amplifying shear stress relative to the net depth by the square of the depth ratio, which is why a notch that removes a fifth of the depth does far more than a fifth of the damage. EN 1995-1-1 arrives somewhere similar through a notched-beam factor that falls away sharply with notch depth and rewards a tapered notch over a square one. Glulam gets less latitude than sawn timber here, and a tension-face notch in a glulam bending member is a question for the manufacturer before it is a calculation.

A hole through an exposed beam also gives the fire a fresh surface to work on. Char advances from every face the fire can reach, and a service penetration left open at both ends is two more faces in the middle of the span, in the region where the moment is highest. That is why penetrations through exposed members belong in the machine file and the fire strategy together, not in a coordination drawing issued after the beam was designed.

The rule that follows is simple and worth writing into the induction pack: nothing is cut into an exposed structural member after the fire case is signed without the engineer re-running it. Both sheets assumed a section. A saw changes the section on both.