The Schatz Linkage is a six-bar spatial mechanism with three pairs of revolute joints whose axes are skewed in space, producing a rhythmic 3D inversion motion as the input shaft rotates. Unlike a planar four-bar that traces a flat curve, the Schatz traces a closed three-dimensional figure-eight path. Engineers use it to drive container blenders and oloid-based mixers because the motion folds material on itself in all three axes at once. The Willy A. Bachofen Turbula shaker-mixer, in production since 1956, is the best-known industrial application.
Schatz Linkage Interactive Calculator
Vary input revolutions, RPM, link length, and length error to see figure-eight cycle timing and closure sensitivity.
Equation Used
The Schatz linkage closes one output figure-eight inversion for each full input revolution, so cycle count equals input revolutions. At constant RPM, one cycle takes 60/rpm seconds. The length-error percentage shows how a small bar-length mismatch compares with nominal link length.
FIRGELLI Automations - Interactive Mechanism Calculators.
- Ideal Schatz geometry closes one complete figure-eight for each input revolution.
- Input speed is constant.
- Length error is evaluated as a simple percentage of nominal bar length.
- This timing calculator does not solve the full spatial Bennett-Schatz closure geometry.
Inside the Schatz Linkage
Paul Schatz discovered the linkage in 1929 while dissecting a cube into three congruent parts — the centre part, a so-called cube belt, can invert through itself if you connect its end-faces with the right joint geometry. The Schatz Linkage is the kinematic skeleton of that inversion. Three rigid links connect through three revolute joints whose axes meet at very specific skew angles. When you spin the input link at constant RPM, the output link does not rotate uniformly — it rocks, swings, and rolls through a 3D figure-eight path that closes perfectly after one full input revolution.
The geometry is unforgiving. The skew angle between adjacent joint axes must be set so that the three axes form a closed spatial chain consistent with the cube-belt construction — typically the perpendicular common-normal lengths and twist angles follow the Bennett-Schatz overconstrained condition. If you build the linkage with arbitrary link lengths it will simply jam, because a generic six-bar spatial linkage has zero degrees of freedom. The Schatz only moves because the link parameters satisfy a special closure condition discovered by Schatz and later generalised by Bennett and Goldberg. Get the link lengths off by even 0.5 mm on a 100 mm bar and the joints bind through part of the cycle, the motor draws spike current, and the bearings start brinelling within hours.
Failure modes are mostly about precision. The most common one is over-tight joint preload — builders see the linkage seize at one specific input angle and assume the geometry is wrong, when really the bearings are fighting the slight elastic deformation of the bars. The second most common failure is twist mismatch on welded frames: a 1° error in the twist angle between adjacent joint axes converts smooth inversion into a stuttering motion that loads the input motor with 3-5× the nominal torque at the worst point in the cycle.
Key Components
- Input Link (Driver Bar): The driven member that rotates at constant input speed, typically 30-100 RPM in industrial mixers. It carries the first revolute joint axis. Bar straightness must hold within 0.1 mm over a 200 mm length to keep the closure condition valid.
- Coupler Link (Middle Bar): Connects the input and output joints through the second revolute pair. This is where the inversion math actually happens — the coupler does not rotate continuously but oscillates and twists. Its length and twist angle relative to the input bar set the entire motion envelope.
- Output Link (Follower Bar): Carries the mounting flange for whatever the linkage drives — a mixing vessel on a Turbula, an oloid roller in a kinematic mixer. Its motion is the 3D figure-eight inversion. Output flange must be balanced within 5 g·mm or vibration at the input bearings rises sharply above 60 RPM.
- Revolute Joints (Three Skew Pairs): Three precision rolling-element bearings whose axes meet at specific skew angles dictated by the cube-belt geometry. Angular tolerance on each axis is ±0.1° — anything looser and the linkage binds. Most commercial builds use angular contact bearings preloaded against duplex pairs.
- Frame and Bearing Housings: Holds the first and third joint axes in fixed spatial relationship to ground. Frame stiffness directly determines whether the closure condition stays satisfied under load — a flexing weldment behaves like a linkage with shifting link lengths and will bind unpredictably.
Who Uses the Schatz Linkage
The Schatz Linkage shows up wherever you need to mix, blend, or expose something to motion in all three axes at once. Pure rotation does not blend powders that segregate by density. A planar shaker only moves material in two dimensions. The Schatz fixes both problems by inverting the working volume through itself on every cycle, which is why it dominates pharmaceutical and laboratory powder blending. It is also the kinematic basis for several oloid-driven water aerators and a handful of art installations that exploit the visually striking 3D figure-eight motion.
- Pharmaceutical Mixing: Willy A. Bachofen AG Turbula T2F shaker-mixer — the workhorse for blending powders and granulates in pharma R&D labs since 1956, running at 22-101 RPM.
- Industrial Powder Blending: GlaxoSmithKline and Novartis use Turbula-class Schatz mixers for homogenising small-batch active pharmaceutical ingredients where conventional V-blenders cause segregation.
- Water Aeration: Oloid GmbH oloid-based aerators for ponds and small lakes — the oloid is dragged through water along a Schatz-style path to inject oxygen with under 50 W of power per unit.
- Laboratory Sample Preparation: Glen Mills and Inversina-class lab mixers for blending dental cements, ceramic slurries, and battery cathode powders in 50 mL to 2 L containers.
- Kinetic Sculpture and Education: The Paul Schatz Foundation in Dornach, Switzerland operates demonstration units showing the cube-belt inversion as a teaching tool for spatial kinematics.
- Food Powder Processing: Specialty cocoa, spice, and infant-formula blenders where shear-free mixing is required to avoid damaging coated particles.
The Formula Behind the Schatz Linkage
The motion of the Schatz Linkage doesn't have a clean one-line position formula like a four-bar — the output orientation is given by a chain of three rotation matrices around the three skew axes. What you actually need as a designer is the relationship between input RPM and the resulting acceleration the working volume sees, because that acceleration is what does the mixing work. At the low end of the typical operating range (around 20 RPM) the peak acceleration is gentle and you get slow, gravity-dominated tumbling. At the nominal 60 RPM you hit the sweet spot where 3D inversion dominates and powder segregation collapses. Push past 100 RPM and centrifugal effects start pinning material to the vessel wall, killing the very mixing action you wanted.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| apeak | Peak acceleration seen by material at the working volume centre during one inversion cycle | m/s² | ft/s² |
| ω | Input angular velocity (2π × RPM / 60) | rad/s | rad/s |
| Reff | Effective radius from the input axis to the working volume centre | m | ft |
| kschatz | Schatz geometry factor — typically 1.4 to 1.8 for a standard cube-belt build, accounting for the 3D inversion amplification over pure rotation | dimensionless | dimensionless |
Worked Example: Schatz Linkage in a battery-cathode powder blender
You are sizing a Schatz-based laboratory blender for mixing 500 mL batches of NMC811 cathode powder with conductive carbon and binder. The effective radius from the input axis to the vessel centre is 0.12 m, the Schatz geometry factor for your cube-belt build is 1.6, and you need to know what peak acceleration the powder will see across the 22-101 RPM operating range of a Turbula T2F-class machine.
Given
- Reff = 0.12 m
- kschatz = 1.6 dimensionless
- RPM range = 22 to 101 RPM
- RPMnominal = 60 RPM
Solution
Step 1 — convert nominal 60 RPM to angular velocity:
Step 2 — compute peak acceleration at nominal speed:
That's about 0.77 g — strong enough to lift and invert NMC811 powder cleanly without packing it against the vessel wall. This is the sweet spot the Turbula T2F is tuned around.
Step 3 — at the low end of the typical operating range, 22 RPM:
That is roughly 0.10 g — barely above gravity-driven tumbling. At this speed the powder shifts gently and you would need 30+ minutes of run time to fully homogenise a 500 mL batch. Useful for fragile coated particles, useless for most production work.
Step 4 — at the high end, 101 RPM:
That's 2.2 g. The math says mixing should be aggressive, and it is — for about 30 seconds, until centrifugal pinning starts. Above roughly 90 RPM on a 0.12 m Reff build, the dense NMC particles get pressed against the outer wall during the rotation phase of the inversion and the lighter carbon black floats free. You get reverse segregation, which is the opposite of what you wanted.
Result
Nominal peak acceleration is 7. 57 m/s² (0.77 g) at 60 RPM, which is exactly why pharma and battery labs cluster around that speed for Turbula-class machines. Across the range, you swing from 1.02 m/s² at 22 RPM (slow, gentle tumble) through 7.57 m/s² at the 60 RPM sweet spot up to 21.5 m/s² at 101 RPM (where centrifugal pinning kicks in and segregation reverses). If your measured mixing time is 2-3× longer than predicted, suspect one of these: (1) the cube-belt twist angle is off by more than 1°, which converts smooth inversion into a stuttering motion that wastes input energy as torque ripple, (2) joint-bearing preload is too high and the linkage is dissipating energy at the revolute pairs instead of the working volume, or (3) Reff was measured from the wrong reference — it must be measured from the input shaft axis to the geometric centre of the loaded vessel, not the vessel mounting flange.
When to Use a Schatz Linkage and When Not To
Schatz Linkages compete with V-blenders and planetary mixers for small-batch homogenisation work. The choice usually comes down to batch size, particle fragility, and how much segregation risk you can tolerate. Here's how the three stack up on the dimensions that actually matter:
| Property | Schatz Linkage (Turbula) | V-Blender | Planetary Mixer |
|---|---|---|---|
| Typical operating speed | 22-101 RPM | 10-25 RPM | 30-300 RPM |
| Batch size sweet spot | 50 mL to 100 L | 5 L to 5,000 L | 1 L to 500 L |
| Segregation risk for density-mismatched powders | Very low — 3D inversion | High at high fill | Moderate |
| Mechanical complexity | High — overconstrained spatial linkage | Low — single rotating axis | Moderate — 2 axes |
| Capital cost (lab-scale) | $8,000-25,000 | $3,000-10,000 | $5,000-30,000 |
| Shear imparted to particles | Very low (tumbling only) | Very low | High (impeller blades) |
| Maintenance interval (industrial duty) | Bearing inspection at 2,000 hr | Bearing inspection at 4,000 hr | Seal/blade inspection at 1,000 hr |
| Best application fit | Pharma APIs, battery cathodes, fragile coated particles | Bulk dry powders, large batches | Wet mixing, dough, viscous slurries |
Frequently Asked Questions About Schatz Linkage
This is almost always a link-length error of 0.3-1.0 mm rather than an axis-angle error. The Schatz is overconstrained — the closure condition demands that link lengths and twist angles satisfy the cube-belt construction simultaneously. If the lengths are off, the linkage is geometrically possible only at certain angles, and it binds elsewhere because the bars must elastically deform to close the loop.
Diagnostic check: disconnect the input motor and rotate the linkage by hand. If torque resistance spikes at a repeatable angle, measure the four critical link lengths against your CAD model — you'll usually find one is off by more than the 0.1 mm tolerance the closure condition tolerates on a 100 mm bar.
At a 10:1 density ratio the V-blender will segregate during discharge no matter how well it mixes during operation. The dense fraction migrates to the bottom of each V-arm during the dwell phase. Schatz inversion eliminates this because the vessel passes through every spatial orientation on every cycle — there is no persistent 'bottom' for dense particles to settle toward.
Rule of thumb: above about 3:1 density ratio or for batches under 10 kg, the Schatz/Turbula wins on homogeneity. Below 3:1 and above 50 kg, the V-blender wins on cost and throughput.
Periodic torque spikes that repeat once per input revolution point to a twist-angle error on one of the three joint axes. A 1° error on the second-axis twist angle converts the smooth inversion into a motion with a brief geometric stiction zone, where the linkage has to elastically deform the bars to pass through the closure point. The motor sees that as a torque demand that scales with frame stiffness.
Quick check: disconnect the load (vessel) and measure current with the linkage running unloaded. If the spike persists, it is geometric, not load-related. Recheck the second-axis twist angle with a precision protractor or coordinate-measuring inspection — that's where the error usually lives.
Yes for the geometry, no for the dynamics. Doubling all link lengths preserves the closure condition because it is scale-invariant — the cube-belt math doesn't care whether the cube is 50 mm or 500 mm on a side. But peak acceleration scales linearly with Reff, so a 2× bigger Schatz at the same RPM doubles the acceleration on the working volume, and bearing loads scale with the cube of the linear dimension at constant ω.
Practical consequence: scaled-up Schatz mixers run at lower RPM than lab-scale units. The Bachofen T10B (10 L) runs at 16-49 RPM versus 22-101 RPM for the T2F (2 L), which keeps the powder acceleration in the same 0.5-1.5 g window where mixing works best.
Real-world Schatz prototypes deviate from theory mostly because of joint compliance, not geometry error. Each revolute pair has 5-20 µm of radial play in the bearing race, and that play accumulates through three joints in series. Over a 100 mm output radius, 15 µm of bearing play at each joint shows up as roughly 0.5-1 mm of path wobble at the output flange.
If the distortion is asymmetric — meaning the path is fine in one half-rotation and wobbly in the other — the cause is usually a bent input bar or a misaligned bearing housing pulling one of the joint axes off its design line. Indicate the input bar between centres and check housing concentricity before blaming the linkage math.
The pure cube-belt Schatz does not have a stationary singularity in the same sense a planar four-bar does — the closure condition guarantees continuous motion through 360° of input. However, the angular velocity ratio between input and output is highly non-uniform, with peak output angular velocity reaching roughly 3× the input at two points per cycle.
That peak ratio is what limits practical RPM. If you size the input motor for nominal torque only, those 3× velocity peaks will demand 3× instantaneous power that a small DC motor can't deliver, and you'll see RPM dips synchronised with the cycle. Size for peak power, not average — typically 2.5-3× the steady-state torque requirement.
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
