Installing a Steel Beam

What this calculation does not cover

• The number returned is the allowance, not the deflection. How far the member actually moves depends on the load, the modulus of elasticity and the second moment of area, none of which this page asks for, so a 6 m span at L/360 gives 16.7 mm of permitted sag with no indication of whether the joist under it moves 8 mm or 40 mm.
• Codes normally impose two limits on the same member at once, one on live load alone and a looser one on dead plus live, and this page returns a single ratio at a time, so an L/360 allowance is the right yardstick only for the live-load half of that pair. The three options offered are the common generic set; brittle finishes, supported masonry and glazing attract tighter fractions in many codes. Some codes also pair the ratio with an absolute cap in millimetres that a long span reaches before the fraction does, and this page only ever divides.
• Span is used exactly as entered, which is the wrong input for a cantilever. Several codes define L for a cantilevered member as twice the projection, so a 1.2 m (4 ft) balcony joist entered as 1.2 m (4 ft) returns roughly half the allowance intended; how L is defined at a cantilever, and whether it is measured clear or centre to centre of bearings, is jurisdictional.
• The allowance has no time in it. Timber and concrete keep deflecting under sustained load, and codes deal with creep by multiplying the long-term component of the calculated deflection before it is compared with a limit like this one, commonly by a factor of the order of 1.5 for dry timber and 2.0 for green material, with the exact figure set by the governing code. A member that clears the limit on the day it is installed can exceed it a year later under the same dead load.
• Clearing a deflection limit is a serviceability result, not a structural pass: bending stress, shear near the supports, bearing crushing on the wall plate and buckling of an unrestrained compression edge are separate checks a member can fail while sitting comfortably inside its L/240 allowance. On a low-slope roof there is also ponding, where water collects in the deflected shape and the added weight deepens it further, which is a stiffness and drainage question no span fraction answers.

What this calculation does not cover

• Both flanges are built from the single width and the single thickness you enter, doubled, so the figure describes a doubly-symmetric shape; a fabricated girder with a wider or thicker bottom flange has to be worked out as two separate flange areas and added, not averaged into one entry.
• Nothing here asks how long the beam is. The answer is the mass of one metre of section, so the piece weight that sets the crane class and the needle sizing comes from multiplying it by the actual stick length, and it reads identically for a 1 m (3.3 ft) stub and a 12 m (39 ft) span.
• The section is taken as constant from end to end. Copes, notches, web penetrations and castellated cells remove steel from a real member, while end plates, stiffeners, cleats, bolts, weld metal, primer, galvanizing and an intumescent film all add it back, and none of the four dimension fields can see any of that.
• The entry fields stop at 60 mm of flange thickness, 40 mm of web and 1.2 m of depth, and anything beyond is brought back to the limit with a notice. The heaviest jumbo column sections and deep fabricated plate girders sit outside that range, so what comes back for them is the weight of a smaller section than the one on the drawing; at the light end a web under 3 mm is lifted to 3 mm and reads heavy.
• Depth, flange and web go in and only a cross-sectional area comes out — those same dimensions are never assembled into a second moment of area. So nothing on this page states what the beam can span, carry or deflect, and two sections with the same weight per metre can differ widely in stiffness depending on how much of the steel sits in the flanges rather than the web.

What this calculation does not cover

• Multiplies one deflection by one fraction and never asks how long or how deep the beam is, which is what decides whether the camber can be induced at all. Fabricators work to a practical minimum around 20 mm (3/4 in), and shallow or short members — commonly anything under about 7.5 m (25 ft) — cannot be cambered reliably by heat or press. Specify 6 mm and what arrives is a beam carrying whatever natural sweep the mill left in it, laid crown up.
• The number returned is a target, not what gets built. Camber is ordered in increments, typically 6 mm (1/4 in), and the tolerance on it is one-sided: more than specified is acceptable, less is not, so the beam usually lands high. On a composite floor that is paid for in concrete — a slab screeded to a level top surface runs thin over an over-cambered mid-span, where cover and capacity are, and thick at the ends of an under-cambered one, where the extra volume is.
• For a composite beam the deflection to enter is the BARE steel section's deflection under wet concrete and its own weight, before the slab cures and the section acts compositely — not the much smaller dead-load deflection of the finished composite section. Camber taken from the composite figure leaves the floor sagging during the pour, which is the moment the level is set.

What this calculation does not cover

• The three minimums are imperial figures, and the checker converts whatever you enter into inches before comparing, so the result is always reported in inches even in metric mode. A 100 mm bearing, the usual minimum in British and European lintel practice, converts to 3.94 in and is reported as a failure against the 4 in general minimum, a shortfall of 1.6 mm. Treat a marginal fail of that size as a units artefact and check the minimum your own jurisdiction actually specifies, which may be expressed in millimetres and may scale with span.
• This is a comparison of one length against a fixed minimum. It never asks for the load above the opening, the span, or the width of the lintel where it sits on the wall, so it cannot say whether the masonry under the bearing is being crushed. Bearing stress is the end reaction divided by the bearing area, length multiplied by width, and on a long span or a heavily loaded opening that check can demand more bearing than the code minimum, so a pass here is a geometric pass and not a stress pass.
• The figure you measure is how far the lintel overlaps the wall, not how much of that overlap is solidly supported. Bearing over ungrouted hollow block cells, an unfilled perp joint, a chase, a cavity tray or a compressible damp-proof course delivers less effective bearing than the tape shows, and hollow units at the bearing normally have to be filled solid or swapped for a solid-top or bond beam unit before the measured length counts.
• Nothing here examines the masonry beyond the bearing. Load spreads downwards and outwards from the bearing into the wall below, so a narrow pier between two openings, a bearing close to a corner, a movement joint or an unrestrained return can be the real weak point even when the bearing length itself is generous. The required end distance or minimum pier width is a separate check set by your masonry code.
• A passing bearing says nothing about the lintel itself. Whether the section can carry the load over the span, and whether it deflects enough to crack the masonry above or bind the door or window below, are separate questions handled by the lintel load capacity calculator. Deflection over brittle finishes is normally limited to a fraction of the span, commonly somewhere between span over 360 and span over 600 depending on the finish and the jurisdiction, and that limit often governs before strength does.

What this calculation does not cover

• The solids figure has to be solids BY VOLUME, and data sheets print solids by weight right beside it. On a heavily filled intumescent the weight figure runs well above the volume figure — a product quoted at 78% by weight can be near 65% by volume — and entering the higher one understates the wet material by that same margin, on a coating whose whole job is to reach a listed dry thickness.
• Theoretical volume, with nothing lost between the pot and the steel. Spraying open sections — web, both faces of every flange, edges — throws a large share of the material straight past the member, and applicators deliberately overapply on top of that because gauge readings have to meet a MINIMUM everywhere rather than an average. Both push the quantity actually ordered well above this number.
• Covers the intumescent film alone. The system that carries the fire test is a specified primer, the intumescent, and usually a sealer or decorative topcoat, each with its own coverage rate and its own compatibility restrictions — a rated build-up is not the intumescent on its own, and a topcoat chosen for colour rather than from the listing can stop the film expanding.