Jumping motion is a power transmission mechanism that converts slowly stored elastic energy into a rapid, single-impulse launch of a payload off a surface. The core component is the energy-storage element — usually a compression spring, torsion spring, or pre-buckled flexure — that loads gradually under a low-power input then dumps its stored energy in milliseconds when a trigger releases. The purpose is to deliver a peak power output far beyond what the input motor can produce directly. You see it in jumping robots like Boston Dynamics' Sand Flea clearing 9 m walls and in mousetrap-car kits sold for school competitions.
The Jumping Motion in Action
Jumping motion works on a power-amplification principle. A small motor, typically a 6V brushed gearmotor pulling 200-400 mA, winds up a spring or cocks a flexure over 1-3 seconds. That stored elastic potential energy then releases through a latch in 5-30 milliseconds, producing peak mechanical power 100 to 1000 times greater than the input motor could ever deliver continuously. The Sand Flea robot stores roughly 60 J in a CO2-charged piston and releases it in under 100 ms — that's 600 W of peak power from a battery system that runs on far less.
The geometry matters because the launch impulse depends on both the energy stored and the time over which it transfers to the chassis. A stiffer spring with a shorter stroke gives a sharper, higher-acceleration jump but transfers less of its energy to forward motion. A softer spring with a longer stroke wastes less energy in heat and impact but needs more vertical travel before the foot leaves the ground. If the latch-and-release timing drifts even 2-3 ms, the launch angle scatters because the leg geometry is rotating during release. You will see this on hobbyist hopping robots — one jump goes 1.5 m, the next goes 0.8 m and 30° off-axis, and the cause is almost always either latch friction or a worn sear engagement.
Failure modes cluster around three things. Snap-through buckling flexures crack along the bend axis after 2000-5000 cycles in PLA, which is why competition jumpers use spring steel or carbon fibre. Compression springs take a permanent set if compressed beyond 80% of solid height, dropping launch energy by 15-20% with no visible damage. And the trigger sear wears fastest at the contact point — once that radius rounds past 0.3 mm, the release becomes unpredictable.
Key Components
- Energy Storage Element: The spring, flexure, or pressurised cylinder that holds elastic potential energy. Compression springs in jumping robots typically store 5-50 J at 30-60% of solid height. The element must be sized for at least 10,000 cycles at working stress, which for music-wire springs means peak shear stress under 700 MPa.
- Loading Drive: Usually a small DC gearmotor with a worm drive or ratcheting cam that compresses the spring slowly. The worm gear is non-back-drivable, so it holds the spring loaded without drawing current. Reduction ratios of 100:1 to 500:1 are common to multiply motor torque enough to load 50 N+ springs.
- Latch and Sear: The trigger that holds the spring at full compression and releases it on command. The sear contact must be hardened — at least 55 HRC — and engage with no more than 0.5 mm of overlap. Too much overlap and the release stutters; too little and the latch fails prematurely under vibration.
- Launch Linkage or Foot: The leg or piston that contacts the ground and converts spring extension into chassis acceleration. Foot contact patch and material set the friction coefficient — rubber at µ ≈ 0.8 grips most surfaces, but on smooth tile you need at least 0.6 to avoid slip-launch where the foot kicks back instead of pushing the chassis up.
- Reset Mechanism: Returns the system to the loading position after a jump. On simple toys this is just gravity. On robots like the EPFL Grillo it's a re-engagement cam timed off a Hall-effect sensor that detects when the chassis has settled.
Where the Jumping Motion Is Used
Jumping motion shows up wherever you need a brief, high-power impulse from a low-power input. The pattern repeats across robotics, packaging, defence, and toys — small motor, slow load, fast release. Designers reach for it specifically when continuous high-power actuation is impossible due to battery size, weight, or cost, and the duty cycle allows a long recovery between events.
- Mobile Robotics: Boston Dynamics Sand Flea throw-bot uses a CO2 piston jumping mechanism to clear obstacles up to 9 m vertical.
- Research Robotics: EPFL's Grillo and the Salto-1P from UC Berkeley use spring-loaded jumping linkages for terrain traversal research.
- Toy Manufacturing: Hexbug Battle Spider and the classic Tomy Skip-It use compression-spring jumping mechanisms triggered by a cam-driven sear.
- Packaging Machinery: Jet-pulse ejectors on Heuft X-ray inspection lines use a compressed-air jumping mechanism to kick rejected bottles off a conveyor in under 50 ms.
- Defence and Aerospace: iRobot 110 FirstLook throw-and-recon robot uses a stored-spring jumping mechanism to right itself after being thrown into a building.
- Educational Kits: Mousetrap-car STEM kits and the Tamiya Jumping Insect kit teach elastic potential energy through a wound torsion-spring jumping mechanism.
The Formula Behind the Jumping Motion
The launch height of a jumping mechanism comes from the energy balance between stored elastic energy and gravitational potential energy at apex. What matters in practice is how the equation behaves across the operating range. At the low end of stored energy, parasitic losses — friction in the latch, foot slip, internal damping in the spring — eat 30-50% of input. At the high end, you start running into structural limits where the chassis itself flexes and absorbs energy that should have gone into the launch. The sweet spot is usually 60-80% of the spring's rated maximum, where efficiency peaks around 65-75%.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| h | Maximum jump height of the chassis centre of mass | m | ft |
| η | Energy transfer efficiency from spring to chassis (typically 0.4-0.75) | dimensionless | dimensionless |
| k | Spring rate | N/m | lbf/in |
| x | Spring compression at launch | m | in |
| m | Total chassis mass | kg | lb |
| g | Gravitational acceleration (9.81) | m/s<sup>2</sup> | ft/s<sup>2</sup> |
Jumping Motion Interactive Calculator
Vary stored energy and release time to see peak launch power and the spring-release impulse.
Equation Used
Peak release power is the stored energy divided by the time it takes to dump that energy into the jumping mechanism. Shorter release time gives a much higher power burst for the same stored energy.
FIRGELLI Automations - Interactive Mechanism Calculators.
- Stored elastic energy is released over the selected time interval.
- Losses in latch, spring, linkage, and ground contact are neglected.
- Release time in milliseconds is converted to seconds for power.
Worked Example: Jumping Motion in a competition micro-jumper for a robotics contest
You are sizing the compression spring on a 250 g micro-jumping robot for the Hebocon contest in Tokyo. The robot uses a stainless steel compression spring with a rate of 2000 N/m, latched by a 3D-printed PETG sear engaged by a 6V N20 gearmotor through a 1000:1 worm reducer. You want to know the achievable jump height at the low, nominal, and high end of the spring's usable compression range so you can decide which compression target to set in the cam profile.
Given
- k = 2000 N/m
- m = 0.250 kg
- g = 9.81 m/s<sup>2</sup>
- η = 0.65 dimensionless
- xnom = 0.025 m
Solution
Step 1 — at the nominal compression of 25 mm, calculate the stored elastic energy:
Step 2 — apply the 65% efficiency factor and divide by m × g to get the apex height at the nominal point:
Step 3 — at the low end of the typical operating range, 15 mm of compression, the spring stores far less energy because the relationship is quadratic in x:
That's 60 mm of jump — the robot barely clears its own chassis height. You'd see it twitch off the ground and flop, not jump. Useful only for tuning the latch reliability without burning out the gearmotor.
Step 4 — at the high end of the usable range, 35 mm compression (close to but not at solid height for a typical 50 mm free-length spring):
In theory 325 mm. In practice you'll measure 250-280 mm because at this compression the chassis flex starts absorbing energy, and a PETG sear at 35 mm load (70 N) begins to creep — release timing scatters by 5-10 ms and launch angle drifts off vertical.
Result
Nominal jump height comes out to 0. 166 m, or about 17 cm — enough to clear a typical Hebocon obstacle and land cleanly. At the 15 mm low end the robot barely lifts 60 mm and looks broken; at the 35 mm high end it theoretically reaches 325 mm but realistically 250-280 mm because chassis flex and sear creep eat the extra energy. The sweet spot sits at 25-30 mm compression where energy transfer stays clean. If your measured height comes in below predicted by 30% or more, check three things: spring set (measure free length — if it's dropped more than 5%, the spring has yielded), foot slip at launch (look for skid marks on the launch surface and check the rubber durometer is below 70A), and sear hardness (a PETG sear at over 5N latch load deforms within 50 cycles and changes the release timing).
Choosing the Jumping Motion: Pros and Cons
Jumping motion competes against other ways to generate brief, high-power impulses. The choice depends on payload mass, repetition rate, and how much you care about acoustic signature and reset time.
| Property | Spring-Based Jumping Motion | Pneumatic Impulse | Direct High-Torque Motor |
|---|---|---|---|
| Peak power output | 100-1000× input motor power | 500-2000× compressor power | 1× motor rated power |
| Reset time between events | 1-3 seconds | 0.2-1 second (with charged tank) | Instant |
| Energy efficiency (input to launch) | 40-75% | 15-30% | 60-85% |
| Cycle life before key part replacement | 5,000-50,000 (spring or sear) | 100,000+ (seals limit) | 10,000+ hours bearing life |
| Cost for a 250 g jumper build | $15-40 | $80-200 | $150-400 |
| Acoustic signature | Mechanical click + impact | Loud air discharge | Quiet whine |
| Best application fit | Low-power mobile robots, toys | Industrial reject ejectors | Continuous-jump platforms with grid power |
Frequently Asked Questions About Jumping Motion
Launch angle scatter almost always traces back to release timing variation, not spring inconsistency. A 2-3 ms drift in the sear release means the leg linkage is at a different angle when energy transfer peaks, and that maps directly to launch direction.
Check the sear contact surface under magnification. If the engagement edge has rounded past about 0.3 mm radius, replace it with a hardened steel sear. Also check that the trigger pull force is consistent — a servo-driven trigger with PWM jitter will release at slightly different latch loads each cycle.
It comes down to packaging and stroke length. Compression springs are the easiest to load and meter but need vertical clearance equal to roughly twice the launch stroke. Torsion springs pack flat and are ideal for low-profile jumpers like the Salto-1P, but loading them requires more complex cam geometry. Snap-through flexures store the most energy per gram but only work for one specific stroke and crack predictably in fatigue.
Rule of thumb — if your chassis height budget is below 30 mm, use torsion. Below 15 mm, use a flexure and accept the cycle-life penalty. Otherwise compression springs win on simplicity.
You're seeing spring set. A music-wire compression spring loaded above 80% of solid height takes a permanent length reduction over the first few hundred cycles. The free length shortens by 2-5%, which drops stored energy by 4-10% because energy scales with x squared.
Measure free length when new and again after 500 cycles. If it has shortened more than 3%, either reduce your working compression or upgrade to chrome-silicon spring wire which has roughly 3× the fatigue resistance of music wire at the same stress.
That's a launch-angle problem, not an energy problem. A purely vertical launch wastes all energy on height. The optimum trajectory angle for distance over flat ground is 45°, but jumping robots usually run 60-75° because they need clearance over obstacles.
Tilt the launch leg so the foot leaves the ground at the angle you want. Salto-1P tilts its body before each jump using a reaction wheel — for a simpler build, just mount the foot at a fixed offset angle and aim the chassis. Don't try to redirect the impulse mid-flight, you don't have the power budget.
Start with the ideal m × g × h, then divide by your expected efficiency. For a typical spring-driven jumper with rubber feet on a hard surface, plan for 50-65% efficiency. For a flexure-based jumper or anything jumping on grass or carpet, plan for 30-40%.
Example — a 500 g robot jumping 0.5 m needs ideal energy of 0.5 × 9.81 × 0.5 = 2.45 J, but at 55% efficiency you actually need to store 4.5 J. Size the spring for that, not the ideal value, or you'll undershoot every time.
You've crossed the slip threshold. The foot needs static friction high enough to hold against the horizontal component of the spring force during release. On polished tile or glass with a hard plastic foot, µ can drop below 0.3, and the foot squirts out from under the chassis instead of pushing it up.
Switch to a soft rubber foot pad with a Shore A durometer of 40-60. That gets you µ ≈ 0.7-0.9 on most indoor surfaces. If the launch is still slipping, increase foot contact area — minimum 50 mm2 for a 250 g jumper.
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
