"L/250" is quoted so often it has become a reflex. Ask any engineer the deflection limit for a beam and the answer comes back instantly. But the number on its own is meaningless without two qualifiers it usually arrives without: deflection of which part, under which load combination. Get either wrong and a beam that "passes L/250" can still be a serviceability failure.
Three deflections, not one
The first trap is treating deflection as a single quantity. The codes distinguish at least three, and each has its own limit:
- Total deflection - the full sag of the member under all sustained and variable load, measured from the unloaded position. This is the one L/250 usually refers to, and it is about appearance and general function.
- Deflection after construction - the additional sag that occurs after the finishes, partitions and cladding are installed. The structure has already deflected under its own weight by then; what matters to a brittle partition is only what happens afterwards.
- The active or variable part - the deflection due to the imposed and other variable actions, the part that moves up and down in service and that things like glazing and partitions feel directly.
A floor that passes total L/250 can still fail the tighter L/500 limit on the part of the deflection that affects partitions. The limits are different because they protect different things.
The limit that governs depends on what the deflection is threatening - overall appearance, a brittle finish, a sensitive piece of equipment - so you have to know which deflection you are checking before the number means anything.
The combination matters
The second trap is the load combination. Strength checks use the ULS combinations; serviceability uses its own SLS combinations, and for long-term deflection the relevant one is the quasi-permanent combination, in which variable actions are reduced by their ψ2 factor to represent the load that is, on average, present most of the time.
This matters enormously for concrete and composite floors, because of creep. Under sustained load concrete goes on deforming for months and years, and the effective stiffness drops accordingly:
For a composite floor, creep under the quasi-permanent load can roughly double the apparent deflection over a year compared with the short-term elastic value.
An instantaneous deflection check that looks comfortable can quietly fail once long-term creep under the quasi-permanent combination is included. The two are not the same calculation.
Getting it right
Three habits keep serviceability honest:
- State which deflection you mean. Total, after-construction, and variable are different quantities with different limits; pick the one that matches the thing you are protecting.
- Use the right SLS combination. Long-term deflection is a quasi-permanent check with creep, not a short-term elastic one.
- Do not stop at the strength check. On longer-span concrete and composite floors, serviceability - not strength - is frequently what actually governs the member size.
The related composite-beam article covers how partial shear connection makes these deflections worse, and the EC serviceability provisions are in the standards reference. The headline: L/250 is not an answer until you say which deflection and which combination.