# Vibration, waves and resonance in machines and structures: Forced vibration and resonance: when small pushes become large motion

Source page: https://www.materialinsights.org/learn/engineering/vibration-and-waves/p2/

[All courses](https://www.materialinsights.org/learn/) / [Understanding Engineering](https://www.materialinsights.org/learn/engineering/) / [Module 7: Vibration, waves and resonance in machines and structures](https://www.materialinsights.org/learn/engineering/vibration-and-waves/) Part 2 of 5

## Forced vibration and resonance: when small pushes become large motion

A car stuck in a sandy rut can often be freed by two people who push only as it rocks away from them, so that small shoves timed to its own rocking add up. That is resonance, and it explains why a ceiling fan wobbles at one speed, why a footbridge can sway under a crowd, and why a washing machine shudders as it spins up. This part shows how big the build-up can be, what limits it, and what happened in some real cases.

### 2.1 A push repeated at the natural frequency builds up motion until damping balances it, and the build-up is about 1 / (2 x zeta)

Forced vibration is vibration driven by something outside the system: a hand rocking a cradle, an unbalanced rotor, wind gusts, a footstep, ground shaking. The system then moves at the driving frequency, not its own. What matters is how the driving frequency compares with the natural frequency.

There are three regimes. When the driving is much slower than the natural frequency, the object just follows the push like a slow press on a spring, and the motion is the static deflection, force divided by stiffness. When the driving is close to the natural frequency, each push arrives just as the object is moving the right way and adds energy, so the motion builds cycle after cycle. This is resonance. Only damping limits it: at resonance the motion is about 1 / (2 x zeta) times the static deflection. A steel frame with zeta 0.01 magnifies by 50. A concrete floor with zeta 0.04 magnifies by 12. A car at 0.3 magnifies by under 2. When the driving is much faster than the natural frequency, the object cannot keep up: it barely moves, because the push reverses before the mass has gone anywhere. That third regime is the basis of vibration isolation in part 4.

Estimate 4 applies this to a ceiling fan, a case every Indian household knows. A small lump of dust or a slightly bent blade gives an off-centre mass, which swings round and pulls the motor sideways.

Data Worked estimate 4: a ceiling fan that wobbles at one speed

| Fan speed (rpm) | Driving frequency (Hz) | Out-of-balance pull = m x w^2 x r (N) | Ratio to natural frequency (4.5 Hz) | Magnification | Wobble at the hook (mm) |
|---|---|---|---|---|---|
| 150 | 2.5 | 3.0 | 0.56 | 1.4 | 1.1 |
| 200 | 3.3 | 5.3 | 0.74 | 2.2 | 2.9 |
| 270 | 4.5 | 9.6 | 1.00 | 10 | 24 |
| 350 | 5.8 | 16 | 1.30 | 1.45 | 5.8 |
| 400 | 6.7 | 21 | 1.48 | 0.83 | 4.4 |

Assumptions: motor and blades 5 kg hang from a rod and hook with sideways stiffness 4,000 N/m, so the natural frequency is square root of (4,000 / 5) / 2 x pi = 4.5 Hz; damping ratio 0.05; a 30 g lump of dust or bend equivalent at 0.4 m from the axis. Pull = 0.03 x (2 x pi x f)^2 x 0.4. Wobble = pull / 4,000 x magnification, with magnification = 1 / square root of ((1 - ratio^2)^2 + (2 x 0.05 x ratio)^2). Check: fans do wobble worst at one middle setting and calm down above it, and a few grams of balancing clip or a wipe of the blades cures it. A top-speed pull of only 21 N is about 2 kg of force, so it is the build-up, not the force, that does the damage.
 Source: Bloomfield, How Things Work (Wiley, 2015), ch. 9, 'Resonance and Mechanical Waves'  Common misreading

#### Resonance is not a bad thing in itself

Resonance only means a strong response when the driving matches the natural frequency. It can be harmful, as in the wobbling fan, or the whole purpose, as in a radio tuner, an ultrasonic cleaning bath or the vibrating-wire strain gauge of part 5. The engineer's question is whether the driving can ever match, and whether damping or design keeps the response acceptable if it does.

- Bloomfield, How Things Work (Wiley, 2015) · ch. 9, 'Resonance and Mechanical Waves'

### 2.2 Footbridges sway because people and bridge begin to move together

When you walk, you push down roughly twice a second, once for each step, and your body also sways from side to side once per two steps, so about once a second. A bridge that has a natural frequency near either number can be set moving by a column of people even though each footfall is gentle. The vertical case is simple resonance.

The Millennium Bridge in London, a shallow steel suspension footbridge across the Thames, opened on 10 June 2000 to a huge crowd. Within days it had to be closed, because under the crowd the deck swayed sideways by up to about 75 mm each way. Arup, its engineers, traced the cause to a feedback between walkers and bridge. As the deck moves, people instinctively widen their stance and step in time with it to keep their balance. That makes their sideways pushes line up with the motion, which makes the motion bigger, which pulls in more walkers. Below a certain crowd size nothing happens. Above it, the sway grows quickly. Bridges with lateral frequencies below about 1.3 Hz and light decks are susceptible. The fix, costing about 5 million pounds, was 37 viscous dampers plus some 50 tuned mass dampers, which lifted the damping above 20% of critical, and the bridge reopened in February 2002.

Estimate 5 shows why span matters so much. A simply supported beam vibrates at a frequency proportional to one over the length squared, so doubling the span cuts the frequency to a quarter. A short span is far above walking pace and safe. A long and light one falls into the walking band. The safe fixes are more depth (stiffness), more damping, or both, since reducing the mass makes the crowd's push more effective.

Data Worked estimate 5: first vertical frequency of a steel footbridge by span

| Span (m) | Working: f = (pi / 2) / L^2 x square root of (EI / m) | Frequency (Hz) | Against walking: steps about 1.6 to 2.4 per second |
|---|---|---|---|
| 20 | 1.571 / 400 x square root of (1e10 / 1,000) = 0.00393 x 3,162 | 12.4 | far above: safe |
| 40 | 1.571 / 1,600 x 3,162 | 3.1 | just below the second harmonic of footsteps (3.2 to 4.8 Hz): check |
| 60 | 1.571 / 3,600 x 3,162 | 1.4 | in the walking band: needs damping or more depth |
| 60, girder twice as deep | EI up about 4 times (module 4), so f x 2 | 2.8 | moved out of the band |

Assumed: EI = 200e9 x 0.05 m4 = 1e10 N m2 (a box girder about 1.2 m deep); mass per metre 1,000 kg/m including deck. Units: EI in N m2 over kg/m gives m4/s2, so the root gives m2/s and dividing by length squared gives per second. Cross-check with sag: the 60 m girder sags 5 x 9,810 x 60^4 / (384 x 1e10) = 0.17 m under its own weight (calculated), and 0.5 / square root of 0.17 = 1.2 Hz, close to the 1.4 Hz above (for a beam with spread-out mass the constant is nearer 0.56 than 0.5, which gives 1.37 Hz).
 Source: The Institution of Structural Engineers, Dallard et al., 'The London Millennium Footbridge', The Structural Engineer, vol. 79, issue 22 (2001), lateral instability below about 1.3 Hz; damping above 20% of critical after retrofit; https://www.istructe.org/journal/volumes/volume-79-(published-in-2001)/issue-22/the-london-millennium-footbridge/

- [The Institution of Structural Engineers, Dallard et al., 'The London Millennium Footbridge', The Structural Engineer, vol. 79, issue 22 (2001)](https://www.istructe.org/journal/volumes/volume-79-(published-in-2001)/issue-22/the-london-millennium-footbridge/) · web
- Building magazine, 'Arup: we have fixed the Millennium Bridge' (2001, read November 2026) · opened 10 June 2000; sway up to 75 mm each way; 37 viscous dampers and 54 tuned mass dampers; about 5 million pounds; https://building.co.uk/news/arup-we-have-fixed-the-millennium-bridge/1015137.article
- Klembczyk, 'Long Term Case Study on Metal Bellows Dampers', Taylor Devices (maker of the bridge's viscous dampers, read November 2026) · 37 viscous dampers fitted late 2001; reopened February 2002; https://www.taylordevices.com/wp-content/uploads/Long-Term-Case-Study-on-Metal-Bellows-Dampers.pdf

### 2.3 Wind can feed a structure: vortex shedding on chimneys and the fall of the Tacoma Narrows Bridge

A steady breeze flowing past a round chimney does not leave it smoothly. Whirls peel off alternately from each side, and each one gives the chimney a small sideways kick. The rate of kicks is about 0.2 x wind speed / diameter. If that rate matches the natural frequency of the chimney, it begins to sway across the wind. Estimate 6 finds that a 2 m chimney is excited by a wind of only 36 km/h. The cure is a spiral of fins (strakes) wound round the top, which stops the whirls forming in step along the height.

The Tacoma Narrows Bridge, in Washington State, USA, is the best known case. It opened in 1940 with a main span of about 850 m, and a deck stiffened only by shallow plate girders about 2.4 m deep. On 7 November 1940, in a wind of about 68 km/h, its deck first rocked up and down, then switched to a twisting motion, tilting by about 45 degrees at its quarter points, and the central span broke away a little after 11 o'clock.

The shedding rate for that wind did not match the twisting seen, so simple resonance was ruled out. What happened was aerodynamic flutter: the wind's force depended on the angle of the deck, so a small twist changed the airflow in a way that fed energy into the twist on every cycle. Damping works by removing energy each cycle; here the wind added more than damping removed, an effect called negative damping, and so the motion grew until failure.

Before a long deck is built, whether it hangs from towers as in a 'Cable-stayed bridge' or from main cables like an 'Iron suspension bridge', a scale model goes into a wind tunnel and engineers find the wind speed at which flutter would start, then make sure it lies well above any storm on the site. The price is a deeper truss or a streamlined box section, with more steel or more shaping than strength alone would need.

Data Worked estimate 6: wind speeds that make a 2 m chimney shed whirls at its own frequency

| Wind speed (m/s) | Wind speed (km/h) | Shedding rate = 0.2 x speed / 2 m (Hz) | Against a chimney natural frequency of 1.0 Hz |
|---|---|---|---|
| 3 | 11 | 0.3 | well below |
| 5 | 18 | 0.5 | below |
| 10 | 36 | 1.0 | matches: lock-in, strong sway across the wind |
| 15 | 54 | 1.5 | above |
| 20 | 72 | 2.0 | above |

The constant 0.2 is the Strouhal number for a round cylinder over a wide range of flows. A taller and thicker chimney has a lower natural frequency, so the matching wind is lower and more common. Check: tall steel stacks are the usual victims, and many have helical strakes in the top third, which agrees with the cure above.
 Source: STRUCTURE magazine (structuremag.org), article on the 1940 Tacoma Narrows Bridge failure (read November 2026), wind about 42 mph, main span 2,800 ft, girders about 8 ft deep, vertical then torsional motion, flutter as the accepted cause; https://www.structuremag.org/?p=13425  Common misreading

#### Tacoma was not just a push at the natural frequency

Popular accounts often show it as resonance, as if the wind were a steady rhythm of footsteps. The accepted explanation is self-excited flutter. In resonance, the driving is independent of the structure's motion. In flutter, the motion itself changes the wind force, so the structure drives itself. The difference matters because more stiffness alone does not cure it; the deck's shape and twisting stiffness do.

- Petroski, To Engineer Is Human (Macmillan, 1985) · ch. 13, 'The Ups and Downs of Bridges'
- [STRUCTURE magazine (structuremag.org), article on the 1940 Tacoma Narrows Bridge failure (read November 2026)](https://www.structuremag.org/?p=13425) · web

### 2.4 Machines shake when a spinning part is off-centre: the washing machine that walks and the engine on its mounts

Any spinning part with its mass slightly off the axis is a vibration source. The off-centre mass swings round and pulls the shaft sideways with a force equal to mass x (angular speed)^2 x radius, so the force grows with the square of the speed. Double the rpm and the pull rises fourfold.

In a washing machine's spin cycle, the wet clothes bunch up on one side of the drum. The drum sits in a tub which hangs on springs, so the whole assembly has a natural frequency of around 3 Hz, or about 180 rpm. During spin-up the drum must pass through that speed. There the motion is largest, because the drive frequency matches the natural one. Above it, the drum is spinning much faster than the tub can respond, and the tub settles into a smaller, steady wobble. If the load is badly lopsided, the shake at resonance is violent enough to shift the whole machine across the floor. This is the 'walking' that Estimate 7 sizes up. Many modern machines sense the imbalance during a slow turn and redistribute the clothes, or refuse to spin fast.

An engine on rubber mounts shows the opposite design choice. The engine of a four-cylinder car fires twice per revolution, so at an idle of 800 rpm (13 Hz) the engine pushes at about 27 Hz. The mounts are chosen so the engine bounces on them at roughly 8 to 12 Hz, well below that. The trade-off is that soft mounts isolate the cabin well but let the engine rock when torque changes, as in a quick start, so many cars use hydraulic mounts that are soft for small motion and stiff for large. See 'Washing machine', 'Internal combustion engine' and 'Heavy-duty diesel engine'.

Data Worked estimate 7: a washing machine tub with a 1 kg lump of wet clothes at 0.2 m from the axis

| Spin speed (rpm) | Out-of-balance pull = 1 x w^2 x 0.2 (N) | Magnification (damping 0.15) | Tub motion (mm) |
|---|---|---|---|
| 120 | 32 | 1.7 | 2.2 |
| 180 (resonance at 3 Hz) | 71 | 3.3 | 9.5 |
| 400 | 350 | 0.25 | 3.5 |
| 1,200 | 3,160 | 0.02 | 2.9 |

Assumptions: moving mass 70 kg; suspension stiffness = 70 x (2 x pi x 3)^2 = 24,900 N/m. Tub motion = pull / stiffness x magnification. Far above resonance the motion tends to lump mass x radius / total mass = 1 x 0.2 / 70 = 2.9 mm, independent of speed. Check: the pull at 1,200 rpm is large, about 320 kg of force, yet the tub moves only a few millimetres, because the springs isolate the rest of the machine; the worst motion for the floor is the crossing at 180 rpm. A machine on a floor with friction coefficient 0.5 holds against about 340 N of sideways push, so lighter loads or soft feet matter.
 Source: Bloomfield, How Things Work (Wiley, 2015), ch. 9, 'Resonance and Mechanical Waves'

- Bloomfield, How Things Work (Wiley, 2015) · ch. 9, 'Resonance and Mechanical Waves'

### See it in the table
[St Steel](https://www.materialinsights.org/commodities/steel/)

### Check your understanding

A structure has a damping ratio of 0.02. Roughly how much larger than the static deflection is its motion at resonance?
 At resonance the magnification is about 1 / (2 x zeta) = 1 / 0.04 = 25.

On the Millennium Bridge, why did sway grow only once enough people were on it?
 Walkers adjust their stance to the moving deck, which makes their sideways pushes line up. Below a threshold crowd size the pushes are random and cancel.

Why is the Tacoma Narrows collapse not best described as plain resonance?
 The accepted explanation is torsional flutter: the twist altered the airflow so as to add energy each cycle, a self-excited motion with negative damping, rather than an independent push at the natural frequency.

[← Part 1: Why things vibrate: mass, springiness and damping](https://www.materialinsights.org/learn/engineering/vibration-and-waves/p1/)[Mark read · Part 3: Balancing, waves and sound →](https://www.materialinsights.org/learn/engineering/vibration-and-waves/p3/)

## Other sections

- [Module overview](https://www.materialinsights.org/learn/engineering/vibration-and-waves/index.md)
- [Why things vibrate: mass, springiness and damping](https://www.materialinsights.org/learn/engineering/vibration-and-waves/p1/index.md)
- [Balancing, waves and sound](https://www.materialinsights.org/learn/engineering/vibration-and-waves/p3/index.md)
- [Controlling vibration: isolation, tuned masses and noise](https://www.materialinsights.org/learn/engineering/vibration-and-waves/p4/index.md)
- [Earthquakes, measuring vibration and avoiding resonance](https://www.materialinsights.org/learn/engineering/vibration-and-waves/p5/index.md)
- [Key terms, quiz and sources](https://www.materialinsights.org/learn/engineering/vibration-and-waves/review/index.md)
