Floor Vibration And Footfall In Apartments

Complaints about a bouncy floor are rarely complaints about strength. The floor is not in danger. It is responding to footsteps in a way the occupant can feel, and that is a serviceability problem the deflection check does not catch.

Why deflection does not cover it

Deflection is a static check: apply the load, measure how far it sags, compare it to a limit.

Vibration is dynamic. What matters is how the floor responds to a repeated input at walking frequency, which depends on three properties the static check never asks about.

Natural frequency, meaning how fast the floor wants to vibrate. Driven by stiffness and mass.

Mass, because a heavier floor responds less to the same footfall energy.

Damping, how quickly the movement dies away. Driven mostly by the finishes, partitions and contents, and almost impossible to predict precisely.

A floor can be very strong, satisfy span over 500, and still have a natural frequency that walking excites.

The frequency problem

Walking produces a force at roughly 1.5 to 2.5 steps per second, plus harmonics at multiples of that.

If a floor's natural frequency sits near one of those harmonics, each footstep adds energy in time with the movement already there, and the response builds. That is resonance, and it is why the problem is a frequency match rather than a weakness.

Floors are usually grouped into two categories.

Low frequency floors, below about eight hertz, where resonance with a walking harmonic is the governing risk. Typical of long-span, light floors.

High frequency floors, above that, where resonance is not the issue and the response is a series of individual impulses. Typical of shorter-span concrete.

The practical consequence: pushing a floor's frequency up, rather than merely making it stronger, is what fixes the problem.

Where it shows up

Long-span lightweight floors, meaning timber or composite steel with a thin topping. The most common source of complaint in residential conversions and upper-level additions.

Open plan with few partitions, because partitions add both stiffness and damping, and an open floor plate has neither. A floor that was acceptable when fitted out can become unacceptable after a refit removes walls.

Cantilevered areas, including balconies, which have a lower frequency than the backspan.

Gyms, studios and plant rooms adjacent to apartments, where the input is rhythmic and far more energetic than walking. Group exercise is the worst case in a residential building and needs its own assessment, not a footfall one.

Timber floors on a concrete frame, where the topping and the structure have different responses.

What actually fixes it

In rough order of effectiveness.

More depth, meaning more stiffness. Raising the natural frequency is the reliable fix, and depth is how you get it. A deeper member, not a stronger one, which is the same conclusion as [deflection limits](/blog/deflection-limits-and-serviceability).

Shorter span, by adding a support. The most effective intervention available and often the least convenient.

More mass, which reduces the response amplitude. Helpful, and it lowers the frequency, so it has to be checked rather than assumed.

More damping, through finishes, partitions, false ceilings and in severe cases engineered dampers. Real, and hard to quantify in advance.

Continuity, because continuous members over several supports have higher frequencies than simply supported ones of the same span.

What does not fix it: adding reinforcement, or increasing the concrete grade. Both add strength. Neither adds meaningful stiffness.

Assessing it

For most ordinary residential floors, a frequency check against a target and a sensible span to depth ratio is enough.

For long spans, open plans, gyms or anything where the consequence of getting it wrong is a building full of complaints, it is a dynamic analysis, usually by [finite element modelling](/blog/finite-element-modelling-structural-design), reporting a response factor rather than a deflection.

On an existing floor it is measurable. An accelerometer and a walking test gives the actual frequency and response, which turns an argument about whether a floor is acceptable into a number. That is a different exercise from the [vibration monitoring](/blog/vibration-monitoring-on-construction-sites) used to protect neighbouring buildings during construction, though the instruments overlap.

The design-stage lesson

Vibration is cheap to design for and expensive to retrofit. Adding fifty millimetres of depth at the drawing stage costs almost nothing. Adding a beam and a column into a finished apartment costs a great deal, and on a long-span floor that is often the only remaining option.

ACSES provides structural engineering for residential and commercial buildings across Sydney. Talk to us about a project.