The Klann Linkage: A Machine That Walks on Legs

From one turning crank to a walking foot

Every walking gait needs two distinct phases at the foot: a stance phase, where the foot presses the ground and drives the body forward at roughly constant height, and a swing (recovery) phase, where the foot lifts clear and arcs back to the start. A rotating crank alone can't do this — a point on a spinning wheel traces a circle, and a circle scuffs the ground the entire way around. The Klann linkage solves the problem by borrowing the oldest trick in kinematics: the coupler curve.

In any four-bar linkage, a point rigidly attached to the floating coupler (the link that connects the two rockers, touching ground at neither pivot) traces a closed, non-circular path as the crank turns. These coupler curves can be teardrops, figure-eights, or D-shapes. Klann's insight was to cascade the geometry — a driving crank feeding a stacked pair of rocker-and-coupler bars — so the terminal coupler point (the foot) draws a path with a long, low, near-straight bottom and a tall recovery arch. One 360° revolution of the crank equals exactly one full step. Because the whole chain has a single degree of freedom, that foot path is deterministic: give the crank an angle θ and the foot position is fixed, with no ambiguity and nothing to control.

Counting the links: why six bars, one DOF

A Klann leg is a six-link planar mechanism. Grübler's (Kutzbach) criterion tells you the mobility directly. For a planar linkage:

• M = 3(n − 1) − 2j₁ − j₂, where n = number of links (including the fixed frame), j₁ = number of one-DOF joints (revolute pins), and j₂ = number of two-DOF joints (none here).
• The Klann leg has n = 6 links and, once every loop is closed, j₁ = 7 revolute pins (pin joints). So M = 3(6 − 1) − 2(7) = 15 − 14 = 1. (Miscounting the joints as 6 would wrongly give M = 15 − 12 = 3 — a reminder that shared pivots connecting three links must each be counted.)

The single result — M = 1 — is the whole point: one motor, one output, no redundant freedoms to fight each other. The links are, functionally: (1) the fixed frame, which carries the two grounded pivots; (2) the crank, the only link that makes full 360° rotations, driven by the motor; (3 and 4) two connecting/coupler bars that transmit the crank motion outward; and (5 and 6) a grounded rocker and the leg (foot) member whose far tip is the coupler point that touches the ground. The crank must be a full-rotation (Grashof) link; the rest oscillate through limited arcs. That crank-rotates-while-followers-rock behavior is a crank-rocker Grashof condition: the shortest link (crank) plus the longest must be ≤ the sum of the other two.

Tuning the geometry: the trade-offs that shape the gait

Every dimension in a Klann leg is a design variable, and they trade against each other the way suspension geometry does on a car. The gait engineer is really optimizing the coupler curve — its stance length, its lift height, its symmetry, and its velocity profile. The controlling levers:

• Crank radius (r): sets the overall amplitude of the step. Doubling the crank roughly doubles both stride length and foot lift, so it scales the gait but not its shape.
• Rocker and coupler lengths: set the shape — how flat the stance is, how tall the swing arch rises, and where the transitions fall. Small changes here can turn a smooth gait into one that plunges the foot into the ground or drags it.
• Ground-pivot spacing and height: tilts and stretches the curve. Moving the fixed pivots changes the ratio of stance-to-swing time, which sets the duty factor (fraction of the cycle a foot is on the ground).
• Transmission angle (μ): the angle between the coupler and the driven link, ideally kept in the range 40° ≤ μ ≤ 140°. Near 0° or 180° the mechanism approaches a toggle (dead) position where a small force gives huge joint loads and the leg can jam. Good Klann designs keep μ well away from these limits through the load-bearing stance.

The horizontal foot velocity during stance ideally matches the walker's forward speed so the foot doesn't slip. If the stance segment isn't a straight, constant-velocity line, the foot scuffs — scrubbing energy, wearing the tip, and jittering the chassis. That single requirement — a straight-line, constant-speed stance — is the hardest thing to synthesize and the reason walking linkages are famous.

Phasing multiple legs: from one leg to a stable walker

One leg can't walk — at some point in its cycle the foot is in the air and there's nothing holding the body up. A walker phases several legs on a common crankshaft so that a supporting subset is always in stance. The rule is simple statics: the vertical projection of the center of mass must stay inside the support polygon formed by the feet currently on the ground.

• Two legs per side, 180° out of phase: the minimum. When one leg lifts, its partner is planted. A four-leg (two-per-side) Klann walker is the canonical desktop demo.
• Six or eight legs: distributes weight, shortens the airborne fraction, and gives a smooth, insect-like tripod-style gait with a low center of mass. More legs also spread the peak crank torque over more of the cycle, flattening the load on the motor.
• Duty factor: with a duty factor above 0.5 per leg, overlapping stance phases guarantee at least static stability at any instant — no dynamic balancing needed, which is exactly why these machines can be pure mechanism with a hobby gearmotor.

Because all legs derive from one crankshaft, the machine steers not by re-timing legs but by driving left and right crankshafts at different speeds — a skid-steer, tank-style turn — or by a differential. There is no per-leg actuation to coordinate; the mechanism is the gait controller.

Sizing, forces, and scaling the mechanism

Sizing a Klann walker starts from the payload and works back through the linkage to the motor. Suppose a four-leg desktop machine of mass 0.5 kg (weight ≈ 4.9 N). With two legs sharing stance, each planted foot carries on the order of 2.5 N vertical. That force is reacted through the leg member as an axial/bending load and appears at the crank as a torque that varies through the cycle — highest when the mechanical advantage (the ratio of crank motion to foot motion) is poorest, typically near stance-to-swing transitions.

• Crank torque: from virtual work, τ_crank = F_foot · (∂s_foot/∂θ_crank). The bracketed term is the instantaneous velocity ratio; where the foot moves fast per degree of crank, torque demand spikes. For a small walker this peaks in the range of tens of mN·m — comfortably within a 6 V hobby gearmotor delivering perhaps 50–150 mN·m at 60 rpm.
• Pin (joint) reactions: each revolute pin sees a reaction that can exceed the foot load several-fold near toggle positions, because the linkage amplifies forces there. This is what sizes the pins and bushings.
• Scaling law: geometrically scale a walker by a factor k and its mass (∝ k³) grows faster than the load-bearing cross-sections (∝ k²). Stresses rise ∝ k — the classic square-cube problem. A meter-scale Klann walker in steel needs proportionally beefier joints, ball or needle bearings instead of plain bushings, and a real speed-reducing gearbox, whereas the toy version runs on friction pins in laser-cut acrylic.

Material choice follows the scale: acrylic or 3D-printed PLA for toys and demos; aluminum or steel bars with bronze bushings or rolling-element bearings for robots that carry sensors and batteries.

Where the Klann linkage actually shows up

The Klann mechanism lives in a specific niche: places that want leg-like obstacle clearance without the cost and complexity of actuated legs. Real hardware includes:

• Educational and hobby walkers: the "Spider" and "Mechanical Spider" kits — laser-cut acrylic or wood, a single DC gearmotor, four to eight legs — are the most common Klann implementations in the world. They're a staple of mechanism-design courses precisely because you can watch the coupler curve draw a step.
• Small mobile robots: hobby and research platforms use Klann legs where wheels bog down — loose sand, gravel, stairs, debris — and where the weight and power budget can't support servo-per-joint legs. The mechanism's tall foot lift lets a modest walker step over obstacles several times what an equivalent wheel could climb.
• Toys and animatronics: the deterministic, self-timing gait makes it ideal for battery toys and window displays that must "walk" reliably for thousands of cycles with zero control electronics.
• Rough-terrain concepts: because a legged foot only touches at discrete points, a Klann walker can bridge gaps and voids that would swallow a wheel, and it leaves a lighter ground footprint — attractive for delicate or uneven surfaces.

What it is not used for: high-speed travel, load-hauling, or terrain that demands adaptive foot placement. Those belong to wheels, tracks, or actuated legs. The Klann linkage owns the middle ground — legged motion you can build with a saw, a motor, and no software.

Limits and failure modes

The Klann linkage is elegant, but its constraints are real and worth naming:

• Fixed gait — no adaptation. The foot path is baked into the bar lengths. It cannot sense a rock, lengthen a stride, or place a foot precisely. On terrain that doesn't match the tuned curve, the foot lands early or late, scuffing and wasting energy. Every walking linkage shares this: they are open-loop by nature.
• Toggle (dead-center) positions. If the geometry lets the transmission angle μ approach 0° or 180° under load, the leg can lock, buckle a bar, or overload a pin. Careful synthesis keeps μ inside the working band, but a poorly copied or rescaled design can drift into a toggle and jam.
• Joint wear and slop. Seven pins per leg, times four-plus legs, means dozens of revolute pairs each accumulating clearance. Backlash at the pins blurs the coupler curve, dropping the foot short of its ideal line — the gait degrades before anything breaks. This is a fatigue and wear problem: pins see millions of oscillating cycles, so undersized or unlubricated joints fail first.
• Stress concentration at pin holes. Each hole is a stress raiser in the bar; combined with the amplified pin reactions near toggle, this is the classic crack-initiation site. Generous fillets, hardened bushings, and keeping μ healthy are the defenses.
• Efficiency. Any foot scuffing, plus friction in every pin, means a Klann walker is less efficient than a good wheel on flat, hard ground — often markedly so. It earns its keep only when the terrain punishes wheels.

Klann linkage versus other ways to move a chassis over ground

Drive concept	Links / parts	DOF	Obstacle it clears	Key limitation
Wheel	1 rotating body	1	≈ ¼ of wheel radius	Stalls on steps taller than hub
Klann leg	6 bars, 7 pins	1	≈ foot-lift height (large)	Foot slips/scuffs off ideal line
Jansen leg	8 bars, 10 pins	1	Moderate (flatter arc)	Lower lift, more joints to wear
Theo-style tracked belt	Many links + rollers	1	Step ≈ belt height	Heavy, high friction losses
Servo-actuated leg	3–4 links + motors	3–4	Terrain-adaptive, unlimited	Needs power, sensors, control