How the two differ in kind
Two words used about the same member's vertical position, and they describe different things.
DEFLECTION is how far a member moves under load, and it is checked against a limit expressed as a fraction of the span. The limits protect things rather than strength: finishes that crack, partitions that distort, doors that bind, and the visible sag of a long member. Different limits apply to different load cases — live load alone and total load are commonly checked separately, because a temporary sag and a permanent one matter differently — and tighter limits apply where brittle finishes are attached.
CAMBER is a built-in upward curvature, fabricated into the member so that when the permanent DEAD load is applied the member flattens out and finishes level. A cambered truss or beam is manufactured bowed upward by a specified amount; the load takes the bow out.
The critical point is what camber is NOT. It is a fabrication instruction rather than a structural property: the member is no stiffer than an uncambered one, it deflects by exactly the same amount under the same load, and camber does nothing whatever about LIVE load — which arrives after the dead load has already flattened the member, and takes it down from there. Since most deflection limits apply to live load or to the load acting after finishes are installed, camber never substitutes for a deflection check.
The practical consequence, which causes most of the trouble on site, is that a cambered member ARRIVES BOWED. It is supposed to. Forcing it flat during installation — packing it down, or setting the finishes to the bowed profile as built — means the camber has nowhere to go when the dead load arrives, and the floor or roof ends up hollow in the middle exactly as if it had never been cambered.
The factors that actually differ
Show
| Camber | Deflection | |
|---|---|---|
| What it is | A manufactured upward curvature, specified as a dimension at mid-span. | The downward movement under load, checked against a fraction of the span. |
| Does it change stiffness | No. A cambered member deflects by exactly the same amount as an uncambered one. | Not applicable — it is the measured consequence of stiffness. |
| What load it addresses | Dead load only. The camber is sized to be taken out by the permanent load. | Live load, total load, and long-term effects — checked separately against their own limits. |
| Can one substitute for the other | Never. A cambered member still has to satisfy its deflection limits. | Not applicable. |
| What it fixes | Appearance and level — a long member that would otherwise read as sagging even while compliant. | Nothing by itself; it is the check that decides whether the member is adequate. |
| How it arrives on site | Bowed upward, visibly. That is correct and not a defect. | Not applicable. |
| The site error | Forcing it flat, or setting finishes to the as-delivered bowed profile. | Assuming a member that looks level is within its limits, which says nothing about the live-load case. |
| Long-term behaviour | In timber, camber can be lost over time as the member creeps under sustained load. | Creep is why timber and concrete are checked for long-term deflection as well as immediate. |
| Where it is used | Long-span steel beams, timber and steel trusses, and floor members where a visible sag would be unacceptable. | Every member, always. |
| Who specifies it | The designer, as a fabrication instruction with a stated amount and load basis. | The code, and the finishes attached to the member. |
Which one, and when
Choose camber when…
- A long-span member where the dead-load sag would be visible even though it is compliant.
- A floor or roof that has to finish level under its permanent load.
- Where a nominally flat roof's fall would be compromised by dead-load deflection at mid-span.
- Where the fabricator can build it in — which for trusses and steel beams is routine.
Choose deflection when…
- Always. Every member is checked against its deflection limits regardless of camber.
- Where brittle finishes are attached, which tightens the limit.
- For the live-load case specifically, which camber does not touch.
- For long-term deflection in timber and concrete, where creep under sustained load continues for years.
Now run your own numbers
This page holds no prices on purpose — a national average is wrong for almost every real project. Quantify both options with your dimensions and your local quotes.
Frequently asked questions
- Does camber make a beam stronger or stiffer?
- Neither. Camber changes the member's manufactured shape and nothing about its section, its material or its span, so its stiffness and its strength are identical to an uncambered member of the same specification — and under a given load it deflects by exactly the same amount. What changes is the starting point: a cambered member begins bowed upward and deflects from there, so after the dead load it sits level rather than sagging. That is an appearance and level benefit, and it is a real one on a long span. It is not a structural benefit, which is why a cambered beam still has to satisfy the same deflection limits and the same strength checks, and why specifying camber is never a response to a member that failed its deflection check.
- Why does camber do nothing for live load?
- Because it has already been used up by the time the live load arrives. Camber is sized so that the permanent dead load flattens it — that is the design intent — so by the time the building is finished and occupied, the member is level and has no reserve of upward curvature left. Live load then deflects it downward from level exactly as it would an uncambered member. Since most serviceability limits are written for live load, or for the load acting after finishes are installed, the check that matters is unaffected by camber. This is also why a cambered member that still looks level under occupancy is not evidence of anything: it says the dead-load camber was sized correctly and nothing about the live-load performance.
- The trusses arrived bowed upward — is that a defect?
- Almost certainly not; it is what a cambered truss looks like before it is loaded. Manufacturers build camber into trusses precisely so they finish level, and a delivery of trusses with a visible upward bow at mid-span is the expected condition rather than a fault. Two things follow. It should not be forced out during installation — packing the truss down, or fixing it to a stringline set flat, removes the camber's ability to be taken out by the load and leaves the member deflecting below level when the dead load arrives. And the finishes should be set to the DESIGN level rather than to the as-installed bowed profile, since the profile will change. Where the bow looks larger than expected, the figure to check is the manufacturer's specified camber rather than a visual judgement.
- What deflection limits apply?
- Several, and a member has to satisfy all of them rather than the one that is easiest. The pattern across codes is consistent even where the numbers differ: a tighter limit where brittle finishes such as plaster are attached and a looser one where they are not; live load alone and total load checked separately, since a sag that comes and goes affects doors and partitions differently from a permanent one; and long-term deflection allowed for in timber and concrete, where creep under sustained load means the sag continues to grow for years after construction. Cantilevers carry their own, usually tighter relative to their projection. Which limit governs depends on the finishes, which is why the finish specification is an input to the structural check rather than a downstream decision.
- Is camber lost over time?
- In timber, it can be, and it is worth knowing when specifying it. Timber creeps under sustained load — it continues to deflect slowly under a constant dead load for years — so a timber truss or beam cambered to be level under its dead load will, over time, deflect further and end up slightly below level. Codes address this by applying a creep factor to the long-term deflection calculation, and designers sometimes specify a larger camber to allow for it. Steel does not creep at ordinary temperatures, so a steel beam's camber is stable. Concrete creeps substantially, which is why long-term deflection in concrete members is a separate calculation with its own factors, and why a concrete member's eventual sag can be a multiple of its immediate one.
- Does camber matter on a flat roof?
- It matters more than on a floor, because of ponding. A nominally flat roof is designed with a fall to its outlets, and dead-load deflection at mid-span can consume or reverse a minimal fall exactly where water would collect. Water standing there is load, which increases the deflection, which collects more water — a progressive mechanism that has caused collapses. Camber addresses the deflection half directly: a cambered member flattens under dead load, so the designed fall survives rather than being cancelled. It is one reason guidance recommends designing roof falls at a gradient well above the minimum, and it is why a roof structure's deflection behaviour is checked alongside the tapered insulation layout rather than separately from it.
- How is the camber amount decided?
- By calculating the deflection under the load the camber is meant to offset and building in that amount as an upward curvature — so the specification is a dimension at mid-span together with a statement of the load basis. What that load basis is matters: camber sized to offset the full dead load leaves the member level once everything is on, while camber sized to offset only the structure's self-weight leaves it slightly low once the finishes are added. Steel fabrication conventions typically camber for a proportion of dead load rather than the whole, on the basis that over-cambering is more visible and more troublesome than a small residual sag, and because achieved camber has a tolerance. The specification should therefore state the amount and the basis rather than simply calling for camber.
- Can camber be added to an existing member?
- Not practically, which is why it is a fabrication decision rather than a site one. Camber in steel is produced by heat or by mechanical means during fabrication, and in a timber truss it is built into the geometry of the assembly at manufacture — neither is something to introduce to a member already in place. Where an existing member has deflected more than is acceptable, the remedies are structural rather than geometric: adding support to reduce the span, which is by far the most effective; strengthening the member with a flitch plate or a composite addition; or replacing it. Jacking a deflected member up and packing it is occasionally done as part of a strengthening scheme, and on its own it simply transfers the problem rather than solving it.