Earthquake Shake Table Experiment You Can Build

Put three towers of different heights on the same shaking platform and something strange happens. Shake it slowly and the tallest one whips back and forth while the short one barely moves. Shake it faster and they swap: the short stiff one starts jumping and the tall one goes quiet. Same ground, same shaking, completely different responses. That is resonance, it is the single most important idea in earthquake engineering, and an earthquake shake table experiment built from card, marbles and rubber bands demonstrates it better than any animation.

Key Takeaways

• Build a shake table from two card panels, marbles as rollers and rubber bands as springs, then test towers on it.
• It teaches structural resilience: natural period, resonance, triangulated bracing and base isolation.
• Fits ages 8 to 15. Younger kids build and shake; older ones measure and compare.
• The usual failure: shaking by hand, differently every time. Pull the plate to a marked line and release, so the input is repeatable.
• Towers fall over. Build them below shoulder height, stand to the side, and use light blocks only.

What You’ll Need

Item	How much	Where to get it	If you don’t have it
Stiff card or thin plywood panels	2, about 30 × 30 cm	Cardboard box, craft shop	Two baking trays, or a tray plus the shallow box it sits in
Marbles	4–6	Toy box	Short pencils or dowels as rollers. Choking hazard under 3; also a slip hazard on the floor
Rubber bands	2–4 strong ones	Stationery drawer	Elastic hair ties, or a length of elastic cord
Tape	1 roll	Drawer	Masking tape works best; you’ll be re-taping
Building blocks	30–40	LEGO, wooden blocks, sugar cubes, dominoes	Dry spaghetti plus modelling clay makes better measurable frames
Modelling clay	1 small block	Craft shop, stationers	Needed for the spaghetti frames and for adding roof mass
Ruler and marker	1 each	School supplies	For marking the pull-back line, which is the key to repeatability
Small soft foam piece	1	Packaging	Only for the base-isolation test; a folded dishcloth works

Time: build the table 20 min, experiments 40 min, cleanup 10 min Cost: nothing if you have marbles and rubber bands Mess level: medium. Blocks will scatter

Do the Experiment: The Earthquake Shake Table Experiment

Step 1 — Build the table

Lay the bottom panel flat on the table. Place four to six marbles on it, spread out. Rest the top panel on the marbles. Now loop two rubber bands from the top panel to anchors taped to the table on either side, so the top panel is held in the middle and springs back when displaced. Tape small card walls around the bottom panel’s edge so the marbles can’t escape.

Step 2 — Make the input repeatable

Draw a line on the bottom panel 3 cm from the top panel’s resting edge. Every single test starts the same way: pull the top panel until its edge reaches the line, then release. That’s your standardised earthquake. Shaking by hand gives different input every time, which makes every comparison meaningless.

Step 3 — Predict, then test base width

Build two towers of identical height, say 25 cm, one with a narrow base and one with a base roughly a third of its height. Before testing, ask which survives longer and by how much. Then release the plate and count how many oscillations each tower survives, or time it to collapse. Repeat three times each.

Step 4 — Add bracing and test again

Build two spaghetti-and-clay frames the same height, one a plain rectangular grid and one with diagonal braces on every face. Same test, same pull-back line, three runs each. Record oscillations survived. The difference is usually dramatic and it is the same triangulation idea that holds up a crane.

Step 5 — Find the resonance

Build three spaghetti-and-clay towers of clearly different heights (about 10, 20 and 30 cm) each with a similar lump of clay on top as roof mass. Mount all three on the table at once. Now shake the plate by hand, slowly at first, then gradually faster, watching all three. Each one will sway most at its own shaking rate.

Step 6 — Test base isolation

Take the tower that performed worst and rest it on a small piece of soft foam, or on a mini platform sitting on three marbles, instead of taping it directly to the plate. Run the standard test again. Record the difference. This is a crude version of a technique used in real buildings and bridges.

What you should see: the wide-based tower outlasting the narrow one by a wide margin; the braced frame outlasting the unbraced one even more. The resonance demonstration is the one that gets gasps: as you speed up your hand-shaking, the tall tower’s sway grows, peaks, then shrinks again while the short tower takes over. Base isolation usually helps, but not always: if the foam is too stiff it does nothing, and saying so is part of the result.

What’s Actually Going On

An earthquake does not push a building sideways. It moves the ground, and the building’s own mass resists being moved: inertia. The structure between the moving foundation and the stationary-ish mass above has to deform, and that deformation generates the forces that damage buildings. The USGS explains the basic mechanics of why the ground moves at all: two blocks of crust slip past each other along a fault, radiating waves.

Every structure has a natural period — the time it takes to complete one full sway if you displace it and let go. It depends on mass and stiffness, roughly as T = 2π√(m/k). More mass lengthens the period; more stiffness shortens it. A tall flexible tower has a long period, maybe one or two seconds for a real mid-rise building. A short stiff one has a short period, a fraction of a second.

Resonance is what happens when the ground shaking contains a lot of energy at the structure’s natural period. Each push arrives in time with the building’s own sway and adds to it, so the motion grows far beyond what the ground itself is doing. Your three-tower demonstration shows this directly: at each hand-shaking rate, one tower is near resonance and the others are not.

This is not an abstraction. Mexico City in 1985 is the textbook case, and it is textbook because the physics is so legible. The city centre sits on soft former lake-bed sediments, and those sediments amplified the ground motion enormously at periods around two seconds. Anderson and colleagues’ analysis of the strong-motion records in Science documented ground motion in the lake-bed zone far larger than on firmer ground nearby, despite the earthquake’s source being hundreds of kilometres away. Buildings with natural periods near two seconds (broadly, mid-rise structures of roughly six to twenty storeys) were damaged out of all proportion, while both shorter and much taller buildings fared better. The hazard was not the shaking alone. It was the match between the ground’s period and the buildings’ period.

Which brings us to base isolation, the idea your foam pad crudely models. If you cannot change the ground, change the connection. Kelly’s review of aseismic base isolation describes inserting flexible bearings between a structure and its foundation, which lengthens the structure’s period, moves it away from the dominant ground period, and dissipates energy in the bearings rather than in the building. Real isolators are engineered laminated rubber or sliding bearings, and they are used in hospitals, museums and bridges worldwide.

Now the honest correction, because the shake table misleads if you stop here. On a home shake table, the stiffest tower usually wins, and children conclude that earthquake-safe means rigid. Real seismic design does not work that way. Modern codes expect a building to deform inelastically in a controlled fashion during a severe earthquake, absorbing energy through ductility rather than resisting everything elastically. A structure stiff enough to resist a big earthquake without yielding would be absurdly expensive, and brittle failure is catastrophic while ductile deformation is survivable. The other thing the table can show you is the soft storey: build one tower with its ground-floor columns longer and unbraced while the floors above are tightly braced, and watch the deformation concentrate at the bottom. That exact configuration, open ground floor for parking or shops, has caused a large share of real collapses.

How to Explain It to Your Kid

Ages 5–8: The ground runs away and the building has to catch up

Say the earthquake doesn’t push the building. It pulls the floor out from under it, and the top of the building is heavy and wants to stay still, so the building gets bent. Then show them with their own body: stand them on a towel and tug it gently sideways. Their feet move, their head lags.

Then let them build and shake, and shake, and shake. At this age the goal is one idea only: wide bottoms and triangles survive; tall skinny towers don’t. Let them discover it by knocking things over, which they will do with enthusiasm.

Ages 9–12: Natural period, and counting swings

Introduce natural period. Measure it: displace a tower by 2 cm, let go, and count how many sways happen in ten seconds. Divide to get sways per second. Do this for all three towers. The tall one gives the lowest number.

Then connect it: shake the table at roughly each tower’s own rate and watch which one responds. That’s resonance, measured and predicted rather than just described. Then run the standardised test and record oscillations-to-collapse for braced and unbraced versions, and express the improvement as a ratio.

Ages 13+: Period matching, ductility and the soft storey

With a teenager, go at the real design problem. Build a soft-storey tower and a uniform one of the same height and mass, and compare. Then discuss why real codes favour ductility over brute stiffness, and what that means about how a building is supposed to behave in a severe earthquake — damaged but standing, with people alive inside.

The extension that makes it an investigation: add mass to the top of one tower, in measured increments of modelling clay, and re-measure its natural period each time. More mass should lengthen the period. Plot period against added mass. If your data follows a square-root trend, you have experimentally recovered the relationship in T = 2π√(m/k) with plasticine.

The question to ask: “If a tall building and a short building are on the same street during the same earthquake, could the tall one be fine and the short one be wrecked? What would have to be true about the ground for that?”

Why Teaching This Early Pays Off

Resonance is one of a handful of ideas that keeps reappearing: musical instruments, radio tuning, microwave ovens, MRI machines, the wobbling of bridges under footsteps, and the destruction of mid-rise buildings in soft-soil cities. A child who has physically felt a tower respond to one shaking rate and ignore another has an anchor for all of it.

For families in earthquake country this is also practical knowledge, and the practical and the conceptual reinforce each other. A child who understands why a soft ground floor is dangerous understands why building codes exist, which is a different kind of civic education than being told to stand in a doorway. Our piece on Mexico City’s earthquake engineering, and how to explain it to kids covers that history properly.

Keep the claim honest, though. An afternoon of shaking towers does not raise anyone’s grades and no evidence suggests otherwise. What it reliably does is produce a correct intuition about a hard physical idea, plus the experience of standardising an input so comparisons mean something — which is the actual skill. The structural ideas connect directly to our spaghetti tower buckling experiment and the bridge-building challenge.

Before You Start: What Parents Should Know

Best age	8 to 15. Under eight, the build is too fiddly and the repeatable-input discipline is beyond reach, though shaking towers over is still fun.
Supervision	Adult builds the table and runs the shaking. Children build towers and record results.
Real hazards	Falling blocks and towers. Marbles underfoot on a hard floor. Marbles as a choking hazard for under-threes. Snapped spaghetti fragments.
Don’t	Don’t build towers above shoulder height, don’t use glass, tins, bricks or heavy metal objects as blocks or roof mass, and don’t put the shake table on a high shelf or worktop edge.

Two safety points that matter in practice. Marbles on a hard floor are a genuine fall risk for everyone in the house, so count them at the start, contain them with taped card walls, and count them again at the end. And keep the towers short: a 25 cm tower that collapses is entertaining, while a 70 cm tower of wooden blocks collapsing at table height can land on a foot or a face. If there is a toddler in the house, do this behind a closed door.

When It Doesn’t Work

• The top plate slides off instead of oscillating. Your rubber bands are too loose or anchored badly. Anchor them to something taped firmly to the table, and keep the pull-back distance small — 3 cm is plenty.
• Every tower survives every test. The input is too gentle. Increase the pull-back distance to 5 cm, add roof mass to the towers, or build them taller. Then re-run the earlier tests, because changing the input invalidates old comparisons.
• Results are inconsistent run to run. Almost always inconsistent input. Use the marked line religiously, release rather than push, and do three runs per configuration.
• The resonance demo doesn’t show anything. The three towers are too similar. Make them clearly different: 10 cm, 20 cm and 30 cm with the same roof mass, and shake slowly at first — most people start far too fast.
• The marbles stop rolling and the plate sticks. Grit or bent card. Wipe the bottom panel, make sure it’s flat, and spread the marbles wider apart.

Take It Further

Measure natural period against mass. Take one tower, measure its period by counting sways in 10 seconds, then add 5 g of clay to the roof and measure again. Repeat to 25 g. Plot period against mass. The relationship should curve, and if you plot period against the square root of mass you should get something closer to a straight line — which is the formula appearing out of plasticine.

Model the soft storey and then fix it. Build two identical towers except that one has an unbraced, taller ground storey. Test both and record where the deformation concentrates. Then retrofit the soft-storey tower by adding diagonal braces to the ground floor only, and test again. Record all three results. Retrofitting a soft storey is a real engineering intervention carried out on thousands of actual buildings, and your child will have done the scale-model version.

What to Watch For Over the Next 3 Months

• Week 4: They look for cross-bracing on buildings and bridges, and comment on open ground floors.
• Month 2: The word “resonance” shows up somewhere else — a swing, a guitar string, a rattling window, the hum in a car at one particular speed.
• Month 3: They ask a harder question, like why some buildings in an earthquake survive while their neighbours don’t, or what a building’s period has to do with the soil underneath it. That question is the whole topic.

Frequently Asked Questions

How do I make the shaking repeatable?

Mark a line and pull the top plate to it every time, then release rather than push. Hand-shaking produces different amplitudes and frequencies on every run, which makes comparisons between towers meaningless. The marked-line release is the single change that turns this from a toy into an experiment.

Why did mid-rise buildings suffer most in Mexico City in 1985?

Because the soft lake-bed sediments under the city centre amplified ground motion strongly at periods near two seconds, and mid-rise buildings have natural periods near that value. The shaking and the buildings resonated together. Analysis of the strong-motion records published in Science documented the amplification.

Does the stiffest tower really win?

On a shake table, usually yes, and that’s the demo’s biggest blind spot. Real seismic design relies on controlled inelastic deformation, ductility, rather than resisting everything rigidly, because a structure stiff enough to stay elastic in a major earthquake would be prohibitively expensive, and brittle failure is far more dangerous than bending.

What is base isolation, in simple terms?

Flexible bearings placed between a building and its foundation so the ground can move without dragging the structure with it. This lengthens the building’s natural period, moving it away from the ground’s dominant period, and absorbs energy in the bearings. Hospitals, museums and bridges use it. A piece of soft foam under a model tower is a crude version.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Anderson, J. G., Bodin, P., Brune, J. N., Prince, J., Singh, S. K., Quaas, R., & Oñate, M. (1986). “Strong Ground Motion from the Michoacan, Mexico, Earthquake.” Science, 233, 1043–1049. https://doi.org/10.1126/science.233.4768.1043
2. Kelly, J. M. (1986). “Aseismic base isolation: review and bibliography.” Soil Dynamics and Earthquake Engineering, 5, 202–216. https://doi.org/10.1016/0267-7261(86)90006-0
3. US Geological Survey. “The Science of Earthquakes.” https://www.usgs.gov/programs/earthquake-hazards/science-earthquakes
4. Gordon, J. E. (1978). Structures: Or Why Things Don’t Fall Down. Penguin.
5. Ready.gov. “Earthquakes.” US Department of Homeland Security. https://www.ready.gov/earthquakes
6. National Research Council. (2012). A Framework for K-12 Science Education. National Academies Press. https://doi.org/10.17226/13165