A Stephenson topology six-bar linkage is a planar mechanism built from six rigid links and seven revolute joints arranged so the two ternary links (the links carrying three pivots each) are not directly connected to each other. The topology was catalogued in the 1870s by British engineer William Kennedy following work tracing back to Robert Stephenson's valve-gear era. By separating the ternaries with binary links, the mechanism delivers complex coupler curves, programmable dwells, and exact-position synthesis that a four-bar can't produce — which is why you find it driving folding blades, seat recliners, and intermittent feed mechanisms.
How the Stephenson Topology Six-bar Linkage Works
Six-bar linkages come in two topologies: Watt and Stephenson. Both have six links and seven revolute joints, but they differ in how the two ternary links sit in the chain. A ternary link is simply a link with three pivot points on it. In a Watt chain the two ternaries share a joint. In a Stephenson chain they don't — they're always separated by binary links. That single topological difference changes everything about what the mechanism can do.
The Stephenson chain has three sub-variants — Stephenson I, II, and III — depending on which link you ground and which one you drive. Stephenson III is the workhorse. You ground a binary link, drive another binary link as the input crank, and read motion off a ternary link as the output. The two-loop structure means the output point traces a coupler curve with up to 6th-order behaviour, so you can specify five or more precision points along a path, or build in a programmable dwell where the output sits nearly stationary while the input keeps rotating. A four-bar can hit at most 5 precision points and can't dwell properly. A Stephenson III can hit 8 or 9 with proper synthesis.
Get the link lengths wrong and the mechanism either locks up at a singularity or branches — meaning it flips to the wrong assembly configuration mid-cycle. Tolerance stack matters. On a typical industrial Stephenson III with 100 mm links, a cumulative pivot-to-pivot error above ±0.1 mm shifts the dwell window by 2-3° of crank rotation, which on a packaging line running at 80 cycles per minute is enough to mistime the downstream tool. Bushing wear is the slow killer — once radial play exceeds about 0.15 mm at any pivot, the coupler curve drifts and the precision points you spent hours synthesising stop landing where they should.
Key Components
- Ground link (frame): The fixed binary link that anchors the mechanism. In a Stephenson III this carries two of the seven revolute pivots. Frame stiffness matters — flex above 0.05 mm at the pivot mounts under load directly corrupts the output coupler curve.
- Input crank: A binary link driven by a motor or actuator, typically rotating continuously at 10-300 RPM in industrial use. Length is usually the shortest in the chain so it can fully rotate without crossing a singularity.
- First coupler (binary): Connects the input crank to the first ternary link. Carries no ground pivots. Its length sets the phase relationship between the input rotation and the first loop's transmission angle, which should stay between 40° and 140° to avoid force amplification problems.
- First ternary link: One of the two three-pivot links that define the Stephenson topology. It belongs to both kinematic loops simultaneously, which is what gives the mechanism its higher-order coupler curve. Pivot-to-pivot tolerances on this link must hold to ±0.05 mm or the dwell timing drifts.
- Second coupler (binary): The binary link that separates the two ternaries — this is the link that distinguishes Stephenson from Watt topology. It transmits motion between the first ternary and the second ternary without sharing a pivot between them.
- Second ternary link / output: The output-bearing ternary link. The point of interest is usually a pin or tool-mount on the third pivot of this link, and that's where you read the coupler curve. In a Stephenson III this link is grounded; in a Stephenson II it's a floating output.
- Revolute pivots (×7): Standard revolute joints, typically needle bearings or bronze bushings rated for 5-50 kN depending on duty. Radial play above 0.15 mm at any single pivot is enough to noticeably distort the coupler curve and shift precision points.
Real-World Applications of the Stephenson Topology Six-bar Linkage
Stephenson topology shows up wherever a four-bar runs out of capability — when you need a long dwell, a complex non-circular path, or exact placement at five or more precision points along a single cycle. Designers reach for it in packaging machinery, automotive seat mechanisms, prosthetic knees, and intermittent feed drives in printing and textiles. The two-loop structure costs more pivots and more synthesis effort than a four-bar, but it earns its keep when the motion profile is too demanding for simpler chains.
- Packaging machinery: Bosch and IMA cartoning machines use Stephenson III six-bars to drive flap-folding blades that need a sustained dwell while glue is applied, then a fast retract — a profile a single four-bar cannot generate cleanly.
- Automotive seating: Lear Corporation and Faurecia recliner mechanisms use Stephenson II topology to translate a short handle pull into a long, controlled seatback rotation with a defined locking dwell at each detent.
- Prosthetics: The Otto Bock 3R60 polycentric knee uses a six-bar Stephenson-style chain to shift the instantaneous centre of rotation during the gait cycle, providing toe clearance during swing and stability during stance.
- Printing and textile feeds: Heidelberg sheet-fed press grippers use Stephenson III synthesis to dwell at the sheet pickup point for 30-40° of crank rotation, transfer cleanly, then dwell again at the release.
- Aerospace deployment: Landing-gear retraction mechanisms on regional aircraft like the Bombardier Q400 use six-bar Stephenson chains to fold the gear into a confined bay along a path no four-bar could trace.
- Industrial robotics: Pick-and-place arms on FANUC and Epson SCARA-adjacent machines occasionally use a Stephenson III to generate approximate-straight-line motion at the end-effector while keeping the actuator off the moving structure.
The Formula Behind the Stephenson Topology Six-bar Linkage
The single most useful equation for a Stephenson chain is the Kutzbach mobility equation, which tells you whether your linkage actually has the 1 degree of freedom you think it does. Get this wrong during synthesis and you'll build a structure that won't move, or an over-constrained chain that binds. The interesting range is narrow — for a planar six-bar with revolute joints only, you need exactly 6 links and 7 joints to land on M=1. Add an eighth joint and you've got a structure (M=0). Drop to 6 joints and you've got 2 degrees of freedom, which means an unconstrained floppy chain that needs a second actuator. The sweet spot is the textbook Stephenson configuration; deviation from it isn't a tuning knob, it's a wrong design.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| M | Mobility (degrees of freedom of the mechanism) | dimensionless | dimensionless |
| n | Total number of links including the ground | count | count |
| j1 | Number of lower pairs (1-DOF joints, e.g. revolute or prismatic) | count | count |
| j2 | Number of higher pairs (2-DOF joints, e.g. cam-follower contact) | count | count |
Stephenson Topology Six-bar Linkage Interactive Calculator
Vary link length, pivot error, cycle rate, and bushing play to estimate dwell timing shift in a Stephenson III six-bar linkage.
Equation Used
This calculator estimates how cumulative pivot-to-pivot error shifts the dwell window of a Stephenson III six-bar. It uses the article reference case: a 100 mm industrial linkage with 0.10 mm cumulative error produces about a 2-3 degree dwell shift. The timing error converts that crank-angle shift into milliseconds at the selected cycle rate.
- Uses the midpoint 2.5 deg of the article's stated 2-3 deg dwell shift.
- Linear scaling is used near the 100 mm link and 0.10 mm tolerance reference case.
- One mechanism cycle equals one input crank revolution.
- Radial play is compared with the article's 0.15 mm warning threshold.
Worked Example: Stephenson Topology Six-bar Linkage in a marine winch drum payout linkage
A commercial fishing gear builder in Bergen is designing a Stephenson III six-bar to drive the level-wind guide on a 12 kW hydraulic trawl winch. The guide must dwell briefly at each end of the drum to lay the wire flush against the flange, then sweep across the 600 mm drum face. The crank turns at a nominal 45 RPM, and the team needs to verify the linkage has exactly 1 DOF before committing to machining the ternary plates from 20 mm 6082-T6 aluminium.
Given
- n = 6 links
- j1 = 7 revolute joints
- j2 = 0 higher pairs
- Crank speed (nominal) = 45 RPM
Solution
Step 1 — apply the Kutzbach mobility equation at the nominal Stephenson III configuration with 6 links and 7 revolute joints:
One degree of freedom. The linkage moves under a single input — exactly what you want for a motor-driven crank. At 45 RPM nominal, the level-wind guide completes one full sweep-and-dwell cycle every 1.33 seconds, which matches the wire payout rate the trawl needs.
Step 2 — at the low end of the typical operating range (15 RPM, slow recovery mode), one cycle takes 4.0 seconds. The dwell at each drum end stretches to roughly 0.6 seconds, which is plenty of time for the wire to seat cleanly against the flange. You'd run this mode when retrieving a partially fouled net.
Step 3 — at the high end (90 RPM, fast deck recovery), cycle time drops to 0.67 seconds and dwell time at each end falls to about 0.10 seconds. That's where things get tight: the wire mass plus rope friction means inertia overshoots the dwell window, and you'll see uneven layering on the drum above roughly 75 RPM in this build.
Step 4 — sanity-check what would happen if a designer mistakenly added an extra brace pin between the two ternary links (a common synthesis mistake when someone tries to stiffen the chain):
Negative mobility means the chain is over-constrained and will either bind, flex, or shear a pin on the first cycle. This is why the Stephenson topology must be respected exactly as catalogued — you cannot add a joint to make it stiffer.
Result
The linkage has M = 1 — a single degree of freedom — confirming the design is kinematically correct and ready for machining. At 45 RPM nominal, the level-wind sweeps the 600 mm drum face cleanly with crisp dwells at each flange. At 15 RPM the dwells become luxuriously long and wire seating is excellent; at 90 RPM dwell time collapses below 0.10 seconds and you'll see uneven wire layering, so the practical operating ceiling for this build sits around 75 RPM. If your built linkage refuses to move smoothly through one full cycle, the most likely causes are: (1) a transmission angle dropping below 30° at a pivot pair — check the second coupler-to-ternary angle at top and bottom dead centre; (2) branch defect, where the linkage was assembled in the mirror configuration and locks at the singularity — disassemble and flip the second ternary; or (3) accumulated pivot-to-pivot length error above ±0.1 mm on the ternary plates causing the two loops to fight each other and bind under load.
When to Use a Stephenson Topology Six-bar Linkage and When Not To
The choice between Stephenson topology, Watt topology, and a simpler four-bar comes down to motion complexity vs. cost. If a four-bar can do the job, use a four-bar — it has fewer pivots, lower cost, and longer service life. Step up to six-bar only when the motion profile genuinely demands it.
| Property | Stephenson Six-Bar | Watt Six-Bar | Four-Bar Linkage |
|---|---|---|---|
| Maximum precision points per cycle | 8-9 with full synthesis | 6-7 with full synthesis | 5 maximum |
| Dwell capability | Programmable dwell up to 60° of crank rotation | Limited dwell, typically under 30° | No true dwell, only momentary slowdown at dead centre |
| Typical pivot count | 7 revolute joints | 7 revolute joints | 4 revolute joints |
| Synthesis complexity | High — requires loop-closure equations and branch checking | High — similar effort to Stephenson | Low — closed-form Freudenstein equation |
| Cost (mid-volume industrial build) | 1.0× baseline | 1.0× baseline | 0.4-0.5× baseline |
| Common operating speed range | 10-300 RPM | 10-300 RPM | 10-1000+ RPM |
| Best application fit | Long dwells, complex coupler curves, exact path synthesis | Approximate straight-line motion (Watt's original use) | Simple oscillation, rotation, basic crank-rocker tasks |
| Failure mode with worn pivots | Coupler curve drift, dwell timing shift | Path drift, especially in straight-line section | Increased backlash, easier to detect and tune out |
Frequently Asked Questions About Stephenson Topology Six-bar Linkage
The deciding factor is whether your output needs to be grounded or floating. Stephenson III grounds one of the ternary links, so the output pivot moves on a fixed-radius arc plus a coupler-curve offset — ideal for a swinging tool head where you want a defined output axis. Stephenson II leaves the output ternary floating, so the output point traces a fully free coupler curve in 2D space — better for path generation where the output isn't constrained to rotate about a fixed axis.
Rule of thumb: if you can describe the output motion as 'an arm that swings', use Stephenson III. If you describe it as 'a point that traces a shape', use Stephenson II.
Kutzbach only counts degrees of freedom — it tells you nothing about singularities. A Stephenson chain has dead-centre positions where the transmission angle at one of the internal pivots passes through 0° or 180°, and at that instant the input crank has no leverage to drive the output. The mechanism doesn't fail topologically, it just loses force transmission momentarily.
Check your transmission angles at the second coupler — that's the binary link separating the two ternaries, and it's where Stephenson chains most commonly hit singularities. Keep all transmission angles between 40° and 140° throughout the cycle. If you can't, you need to re-synthesise with different link ratios.
Synthesis assumes perfectly rigid links and zero-clearance pivots. Reality has neither. The two most common culprits are pivot clearance and link flex. A typical 8 mm bronze bushing with 0.05 mm radial clearance contributes roughly 0.03° of angular slop per pivot, and with 7 pivots in series that's 0.2° just from clearance — small, but it compounds with link flex.
The bigger issue is usually ternary-link bending. Ternary plates carry asymmetric loads because they have three pivots, and if you've cut them from sheet stock under 8 mm thick they flex enough under peak load to shift the dwell window by several degrees. Either thicken the plates or rib them.
Yes — replace the input crank with a slider, and one revolute joint at the input becomes a prismatic joint. Mobility stays at M=1 because Kutzbach treats prismatic and revolute as equivalent lower pairs. This variant is sometimes called a Stephenson slider-six-bar.
Works well when your prime mover is naturally linear — pneumatic cylinder, hydraulic ram, or an electric Linear Actuator. Watch out for the input dead-centre: at full extension and full retraction the slider has zero velocity, so if you need a smooth output velocity through that region you'll want to design the linkage so the output's high-velocity phase aligns with the slider's mid-stroke.
Classic branch defect. A Stephenson chain has two valid assembly configurations (mirror images of each other across the loop-closure constraint), and during synthesis you might solve precision-point equations that span both branches. The mechanism assembles fine in the first branch, drives through part of the cycle, then hits the singularity that separates the branches and either locks or jumps.
Diagnostic: rotate the crank slowly by hand and watch the second ternary. If at some angle it tries to flip through the line connecting its two end pivots, you've got a branch defect. Fix is to re-synthesise with branch-checking constraints, or accept that the linkage operates only over a partial crank rotation between the two singularities.
If your motion profile is highly arbitrary — for example, a custom velocity curve with multiple accelerations and decelerations — a cam will out-perform any linkage. Cams give you point-by-point freedom over the output profile, where a six-bar is constrained by what loop closure mathematically allows.
The crossover point is roughly 9 precision points. Below that, a properly synthesised Stephenson III is cheaper, lighter, and far more reliable than a cam (no follower wear, no spring return, no contact stress). Above 9 points, cam wins. Also pick the cam if your output needs to dwell for more than 90° of crank rotation — that exceeds what a Stephenson chain can produce without a structural singularity.
References & Further Reading
- Wikipedia contributors. Six-bar linkage. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
