Cam Mechanism Automata Kids Can Build From Cardboard

Here is the rule that makes this a measurement rather than a craft: a round cam glued off-centre by 1 cm lifts its follower exactly 2 cm, and the size of the disc makes no difference at all. Build one with a 1 cm offset and one with a 2 cm offset, mark a ruler on the side of the box, and your kid can predict the lift before turning the handle. Cam mechanism automata kids build in an afternoon usually get judged on whether the duck looks like a duck. This one gets judged on whether the follower moved 20 mm, and the answer is a number.

Key Takeaways

• You’ll build a crank-driven box automaton with interchangeable cardboard cams, then measure how far the figure rises for each one.
• The concept is that a cam is a shape that stores a motion. Change the shape and you change the movement without changing anything else.
• The rule to test: for a circular cam, lift equals twice the offset from the axle, regardless of disc diameter.
• Best for ages 8 to 14. An adult cuts the discs and the box holes.
• The usual failure is a follower that jams in its guide rather than sliding. Alignment matters more than anything else in this build.

What You’ll Need for Cam Mechanism Automata Kids Can Build

Item	How much	Where to get it	If you don’t have it
Shoebox or similar card box	1	Cupboard	Any box about 20 × 15 × 12 cm
Corrugated cardboard	2 sheets A4 size	Recycling	Cereal-box card, doubled up
Wooden dowel, 6 mm, 25 cm	1	Hardware shop	A bamboo skewer, though it flexes more
Bamboo skewers	3	Supermarket	Wooden kebab sticks
Drinking straws	3	Kitchen	Rolled paper tubes, taped
White glue	Small bottle	Stationery	Low-temperature glue gun, adult only
Ruler and pencil	1 each	Stationery	Any straight edge
Masking tape	One roll	Hardware drawer	Any tape
Craft knife (adult only)	1	Hardware shop	Strong scissors plus a skewer to pierce holes
Cork or bottle cap	2	Kitchen	A blob of tape to cap sharp points

Time: prep 30 min, building 40 min, experimenting 20 min Cost: free if you have a box and a skewer Mess level: low to medium, mostly cardboard offcuts

Do the Experiment

Step 1 — Make the frame and the axle holes

An adult pierces a hole in each long side of the box, 5 cm up from the base and centred, just big enough for the dowel to spin freely. Push the dowel through. It must turn without binding, so enlarge a hole slightly rather than forcing it.

Step 2 — Build the follower guide

Pierce a hole in the box lid directly above the axle. Push a 4 cm length of straw through it and tape it so it stands perfectly vertical. Slide a skewer down the straw until it touches the axle. This straw is the single most important part of the build: if it leans, the follower will stick.

Step 3 — Give the follower a foot

Glue a 2 cm cardboard disc flat onto the bottom end of the skewer, so the follower rests on the cam with a flat face rather than a point. A point digs into the cam and judders. Cap the top end of the skewer with a cork or a ball of tape.

Step 4 — Cut two circular cams and mark their offsets

Cut two cardboard discs, 6 cm across. On the first, mark a point 1 cm from the centre; on the second, 2 cm from the centre. Pierce at those marks and glue each disc onto the dowel so it spins with the axle. These are eccentric cams, and the marked distance is the offset.

Step 5 — Measure the lift for each cam

Tape a paper ruler to the straw guide and mark where the top of the follower sits. Turn the handle one slow revolution and mark the highest and lowest positions. Measure the difference. Predict first: 1 cm offset should give 2 cm of lift. Then fit the second cam and predict again.

Step 6 — Change the shape, change the story

Cut a snail cam: a spiral that grows from 2 cm radius to 4 cm over most of a turn, then drops straight back. Fit it and turn the handle slowly. Then make two cams and mount them half a turn apart on the same axle, each with its own follower, so two figures alternate.

What you should see: The lift comes out at twice the offset, usually within 2 to 3 mm, and the error is nearly always a slightly loose follower rather than wrong geometry. The circular cam gives a smooth, even rise and fall that feels gentle at the top and bottom of the stroke. The snail cam feels completely different: a slow climb, then a sudden drop with an audible clack. Two cams set half a turn apart make the figures take turns, which is the moment the project stops being a mechanism and starts being a story.

What Cam Mechanism Automata Kids Build Actually Measure

Cam	Offset or profile	Measured lift	Character of the motion
Circular, small offset	1 cm	2 cm	Smooth, even, gentle at both ends
Circular, large offset	2 cm	4 cm	Same character, double the travel
Snail	2 to 4 cm radius	2 cm	Slow rise, sudden drop, clack
Pear	Flat section on a circle	Varies	Rises, pauses at the top, falls
Two circular, 180° apart	1 cm each	2 cm each	Alternating, one up while one is down

What’s Actually Going On

A cam converts rotation into a specific pattern of back-and-forth motion. The clever part is that the pattern is not programmed or controlled; it is stored in the outline of a piece of cardboard. Whatever shape the follower traces as it rides over the cam is the movement you get, which makes a cam a piece of physical memory.

Take the circular cam first, because its maths is exact. The disc is a circle whose centre sits a distance e from the axle. As the axle turns, the highest point of the disc’s surface above the axle varies between the disc radius plus e and the disc radius minus e. Subtract one from the other and the radius cancels: the total travel is 2e. That is why a bigger disc with the same offset gives exactly the same lift, which almost nobody predicts correctly the first time.

There’s more hiding in that geometry. For a flat-faced follower, the height follows a cosine of the rotation angle, which means the follower performs precise simple harmonic motion, the same mathematics as a pendulum’s swing or a mass bouncing on a spring. A kid turning a crank at a steady rate is generating a sine wave with cardboard. That connection is worth pointing out to a teenager who has met sine curves only on graph paper.

The snail cam shows why real cam design is harder than it looks. Its sudden drop means the follower is in free fall for part of the cycle and then stops abruptly, which produces a jolt. Engineers care about this intensely, because the forces in a cam-follower system depend on the follower’s acceleration, not its position. Modern design work is therefore framed around acceleration profiles: Flocker’s 2012 paper in the Journal of Mechanical Design presents an acceleration-based cam profile precisely to control clearance and impact in single-dwell applications. In an engine, a badly shaped cam makes valves bounce; in an automaton, it makes the duck clack.

Cams and the mechanisms around them have a long design tradition. Cornell’s Kinematic Models for Design Digital Library holds the Reuleaux collection of nineteenth-century teaching models of exactly these linkages, and mechanism design remains standard content in courses such as MIT’s electromechanical robotic systems. The reason cams persist in a world of software is simple: a shape never needs reprogramming, never crashes and costs nothing to copy.

One honest caution about the build. Cardboard cams wear. After a few hundred turns, the edge compresses where the follower presses hardest and your measured lift drops by a millimetre or two. That is not a mistake, it is wear, and noticing it is a genuine engineering observation.

How to Explain It to Your Kid

Ages 5–8: the shape is the dance

Let them turn the handle and watch the figure bob. Then swap in the snail cam without explaining anything and let them describe the difference in their own words: “it goes up slow and falls down fast”. That sentence is a motion profile. Then let them draw a shape they think will make a particular dance and cut it with an adult’s help.

Ages 9–12: twice the offset

Measure the offset with a ruler, predict the lift, measure the lift. Do it for both cams. When the prediction works twice, ask the question that tests whether they understood: would a disc twice as big with the same offset lift more? Most will say yes. Then test it, and the fact that the radius cancels becomes memorable because they were wrong first.

Ages 13+: profiles, phase and acceleration

Introduce phase as an angle. Two cams 90° apart give a different story from two cams 180° apart, and the axle can be marked in degrees to set it deliberately. Then introduce the idea that the follower’s acceleration, not its height, determines the forces, and have them design a cam that rises and falls without any sudden corner in its outline. That is the actual problem engine designers solve.

The question to ask: “I want the figure to rise, stay up for a quarter of the turn, then come down gently. Sketch me the cam that does it.”

Why Teaching This Early Pays Off

Automata sit in an unusual place: they are mechanical, visual and narrative at once, which gives children who do not think of themselves as engineering-minded a legitimate way in. The underlying idea, that a shape can encode a behaviour, is also one of the most transferable in design. It reappears in gear profiles, in the grooves of a vinyl record, in the shape of a key, and in the lookup tables that still run inside embedded software.

More immediately, this is a clean introduction to periodic motion. A child who has turned a crank and watched a follower trace the same path every revolution has a physical anchor for the word cycle, for frequency, and eventually for the sine function. That anchor is worth having before the graphs arrive.

The structure to use is guided rather than free. Alfieri and colleagues’ 2011 meta-analysis found that unassisted discovery did not improve learning, while guided discovery with explanation alongside the activity did. In practice, that means stating the twice-the-offset rule and then testing it, rather than hoping it emerges from play. The National Research Council’s framework lists developing and using models as a core practice, and a cam is literally a physical model of a motion. One afternoon will not teach kinematics, but it does put a name and a number on something a child built.

Before You Start: What Parents Should Know

Best age	8–14. Under 8, an adult builds the mechanism and the child designs the cams and the figures
Supervision	Direct for all cutting and for a glue gun. An adult pierces the box and cuts the discs
Real hazards	Craft knives cut deeply, especially on curved cuts, so an adult does them on a cutting mat. Skewer points are a puncture and eye hazard, so cap every exposed end with a cork or tape before the first turn of the handle. Hot glue guns run at about 120 to 190 °C and blister skin on contact; use a low-temperature gun or white glue instead
Don’t	Don’t substitute a metal skewer or a nail for the wooden follower. It concentrates force, cuts the cam and is far more dangerous if it slips out of the guide

When It Doesn’t Work

• The follower sticks instead of sliding. The straw guide is not vertical, or the skewer is bowed. Re-tape the straw against a set square, and swap a bent skewer for a straight one.
• The figure rises but doesn’t come back down. Friction exceeds the weight pulling it down. Add a small weight to the follower, such as a blob of modelling clay below the lid, or widen the guide slightly.
• The cam slips on the axle. It needs gluing on both faces, not just one. Glue a second small disc behind the cam as a collar so it cannot wander along the dowel.
• The lift is less than twice the offset. Usually the follower foot is riding off the edge of the cam at the extremes. Make the cam disc larger, or reduce the offset so the follower stays on the surface throughout.
• Turning the handle twists the whole box. The axle holes are too tight. Enlarge them a little and add a straw sleeve in each hole as a bearing.

Take It Further

1. Profile library. Cut four cams, measure each one’s lift and describe each one’s character, then graph height against rotation angle in 45° steps. You end up with four motion profiles drawn from your own measurements, which is how cam designers actually work.
2. Phase choreography. Mount three cams on one axle at 0°, 120° and 240° with three followers, and the figures move in sequence. Mark the dowel in degrees so the sequence is designed rather than discovered.
3. Drive it with gears. Add a gear pair so one turn of the handle gives two turns of the cam axle, which doubles the cycle rate without changing the lift. The cardboard gear train covers that build, and for more cardboard technique there’s cardboard engineering with low-cost materials.

What to Watch For Over the Next 3 Months

• Week 4: Spotting cams in the house. A door latch, a sewing machine, a windscreen wiper linkage, the lifter in a music box.
• Month 2: Talking about motion in terms of shape. Describing a movement as “slow up, fast down” and reaching for a drawing rather than a word.
• Month 3: Designing backwards from the motion. If your child sketches the cam they need before cutting it, rather than cutting and seeing, the design habit has arrived. Simple machines are a good companion, and levers, pulleys and ramps use the same build materials.

Frequently Asked Questions

Why doesn’t a bigger cam disc give a bigger lift?

Because the lift is the difference between the highest and lowest surface heights, and the disc’s radius appears in both. Subtract one from the other and the radius cancels, leaving twice the offset. Test it with two discs of different sizes at the same offset; the result is identical and it convinces far better than the algebra.

Can we use a bamboo skewer for the main axle?

Yes, but it flexes, which makes the lift slightly inconsistent under load. A 6 mm hardwood dowel from a hardware shop costs very little and makes a noticeably better machine. If you use a skewer, keep the cams close to the box sides to reduce bending.

How do I make a figure pause at the top?

Flatten part of the cam’s outline so that for a range of rotation angles the surface height does not change. That flat is called a dwell, and it is one of the most useful features in cam design. In engines, dwells hold valves open for a set fraction of the cycle.

Our cam makes a loud clack. Is it broken?

No, you have built a sudden drop into the profile, which is what makes the clack. If you want it gone, round the corner where the radius falls away sharply, so the follower descends gradually. If you like it, keep it and note that you designed a sound, which is a legitimate thing for a mechanism to do.

---

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. Flocker, F. W. (2012). “A Versatile Acceleration-Based Cam Profile for Single-Dwell Applications Requiring Cam-Follower Clearance During Dwell.” Journal of Mechanical Design, 134(7). https://doi.org/10.1115/1.4007002
2. Cornell University Library. KMODDL: Kinematic Models for Design Digital Library. https://kmoddl.library.cornell.edu/
3. Techet, A., & Hover, F. (2009). Design of Electromechanical Robotic Systems (2.017J). MIT OpenCourseWare. https://ocw.mit.edu/courses/2-017j-design-of-electromechanical-robotic-systems-fall-2009/
4. Alfieri, L., Brooks, P. J., Aldrich, N. J., & Tenenbaum, H. R. (2011). “Does discovery-based instruction enhance learning?” Journal of Educational Psychology, 103(1), 1–18. https://doi.org/10.1037/a0021017
5. National Research Council. (2012). A Framework for K-12 Science Education. National Academies Press. https://doi.org/10.17226/13165
6. White, R., & Gunstone, R. (1992). Probing Understanding. Falmer Press.