Bohnenberger's Machine is a 3-ring gimbal-mounted spinning sphere built by Johann Bohnenberger around 1810 to demonstrate gyroscopic precession in a classroom. It replaced earlier fixed-axis spinning tops, which could not show how a rotating mass behaves when its support is tilted in two independent directions. The machine lets students see angular momentum conservation directly — spin the sphere, tilt the outer ring, and the inner axis precesses visibly. Foucault later used the same gimbal architecture in 1852 to detect Earth's rotation.
Bohnenberger's Machine Interactive Calculator
Vary rotor size, mass, spin speed, and applied torque to see angular momentum and gyroscopic precession change.
Equation Used
The calculator uses the article's precession relation, Omega = tau / L. Angular momentum is computed from a solid spherical rotor using I = (2/5) m r^2 and L = I omega, with spin speed converted from rpm to rad/s.
- Rotor is modeled as a solid sphere.
- Gimbal pivots are frictionless and axes intersect at the rotor center.
- Applied torque is perpendicular to the spin angular momentum.
- Small disturbance precession is assumed.
Operating Principle of the Bohnenberger's Machine
The machine is simple in concept and unforgiving in execution. A heavy sphere — typically brass or ivory in the original 1810 instruments, around 50 mm diameter and 200-400 g — sits on an axle inside the innermost gimbal ring. That ring pivots inside a second ring, which pivots inside a third outer ring fixed to the stand. Three orthogonal pivot axes, all passing through the sphere's centre of mass. Spin the sphere by hand or by pulling a cord wound around its axle, and you have a free gyroscope. Tilt the stand, and the sphere keeps pointing where it was — the gimbals just rotate around it.
Why the three-ring gimbal architecture? Because two rings only give you two degrees of freedom, and a freely-precessing rotor needs three. With three rings, the spin axis can point anywhere on a sphere of orientations while the support frame moves independently. This is the same gimbal mounted gyroscope geometry you find inside a marine compass or an inertial measurement unit, just hand-built in brass.
Tolerances matter more than people expect. Each pivot must be near-frictionless — historically jewelled bearings, today miniature ball races with end-play under 5 µm. If the pivot axes do not intersect at a single point, the rotor's centre of mass shifts as the gimbals move and gravity adds a torque that masks the precession you are trying to demonstrate. If the sphere is unbalanced by even 0.1 g at 25 mm radius, you get a wobble that drowns out the angular momentum demonstration. And if any two gimbal axes line up — gimbal lock — the rotor loses a degree of freedom and the demonstration fails until you nudge the rings apart.
Key Components
- Spinning Sphere or Rotor: The mass that carries the angular momentum. Original Bohnenberger units used a 50 mm ivory or brass sphere around 200-400 g, balanced to within 0.05 g·mm. Modern teaching replicas use a machined brass or steel sphere with a drilled axial bore for the spin axle.
- Inner Gimbal Ring: Holds the rotor axle on two diametrically opposite pivots. This ring rotates freely about an axis perpendicular to the spin axis. Pivot end-play must stay below 5 µm or the rotor wanders and the precession reads dirty.
- Middle Gimbal Ring: Carries the inner ring on a second pair of pivots, orthogonal to the inner ring's pivot axis. This is the ring that visibly precesses when you apply a torque to the spin axis.
- Outer Gimbal Ring: Fixes the assembly to the stand and provides the third orthogonal pivot axis. On Bohnenberger's original design, this ring is a heavy brass hoop screwed to a wooden tripod base.
- Spin-up Cord or Lanyard: A cord wound around the rotor axle. A sharp pull spins the sphere up to 600-1500 RPM — enough angular momentum to give 30-90 seconds of clean precession before bearing friction bleeds it off.
- Pivot Bearings: Originally polished steel points in agate or jewelled cups, like a watch pivot. The friction torque must be a small fraction of the precession torque you intend to apply, otherwise drift contaminates the demonstration.
Real-World Applications of the Bohnenberger's Machine
The machine is a teaching and historical instrument first, an engineering ancestor second. You will not find one driving a production line, but you will find direct descendants — gimbal mounted gyroscopes, IMUs, and stabilised platforms — running in everything from ships to spacecraft. The original purpose was pedagogical, and that is still where you see them: physics teaching collections, science museums, and the occasional research demonstration of angular momentum conservation.
- Physics Education: Teaching collections at the University of Tübingen, where Bohnenberger taught, still hold original 1810-era machines used to introduce gyroscopic precession to first-year mechanics students.
- Science Museums: The Deutsches Museum in Munich displays a Bohnenberger-style 3-ring gimbal as part of its history-of-navigation gallery, alongside Foucault's 1852 gyroscope.
- Historical Instrument Restoration: Workshops like Tesseract Scientific Instruments in Hastings-on-Hudson restore 19th-century Bohnenberger machines for university collections and private buyers, replacing worn jewel pivots and rebalancing the spheres.
- Inertial Navigation Heritage: Sperry Marine and Anschütz both trace their gyrocompass lineage back to the Bohnenberger gimbal architecture — the 3-ring suspension was the direct ancestor of the Anschütz 1908 gyrocompass used on SMS Deutschland.
- Aerospace Demonstration Rigs: Aerospace engineering programmes at MIT and TU Delft use modernised Bohnenberger replicas with electric spin-up motors to show students attitude-control behaviour before introducing CMG (control moment gyroscope) hardware.
- Watchmaking and Precision Mechanics Training: The WOSTEP watchmaking school in Neuchâtel uses miniature gimbal-mounted rotors to teach pivot geometry and friction analysis — the same skills needed to rebuild a Bohnenberger machine.
The Formula Behind the Bohnenberger's Machine
What you actually want to predict is the precession rate — how fast the inner gimbal swings around when you hang a small weight off the spin axis. At the low end of the typical operating range, with the rotor barely spun up, precession is fast and chaotic because angular momentum is small. At the nominal range — say 1000 RPM on a 50 mm brass sphere — the precession rate is a steady 1-3°/s, slow enough to follow with the eye and obviously coupled to the applied torque. At the high end, above 2000 RPM, precession slows to a crawl and the demonstration becomes visually boring even though the physics is identical. The sweet spot for teaching is around 800-1200 RPM.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Ωp | Precession rate of the inner gimbal | rad/s | deg/s |
| τ | Applied torque about the spin axis | N·m | lbf·in |
| Is | Moment of inertia of the rotor about the spin axis | kg·m² | lb·in² |
| ωs | Spin rate of the rotor | rad/s | RPM |
Worked Example: Bohnenberger's Machine in a university physics teaching replica
A physics demonstrator at a Scottish university is rebuilding a Bohnenberger replica for a first-year mechanics course. The brass sphere is 50 mm diameter, 350 g, and the demonstrator hangs a 2 g brass weight on a 30 mm arm off the inner gimbal to apply a known torque. The class needs to see precession that is slow enough to follow but fast enough to measure with a stopwatch — target around 2°/s at nominal spin.
Given
- msphere = 0.350 kg
- rsphere = 0.025 m
- mweight = 0.002 kg
- Larm = 0.030 m
- ωs,nom = 1000 RPM
Solution
Step 1 — compute the rotor moment of inertia for a solid sphere, Is = (2/5) × m × r²:
Step 2 — compute the applied gravitational torque from the 2 g weight on the 30 mm arm:
Step 3 — convert nominal spin to rad/s and compute precession at 1000 RPM:
That is right in the visual sweet spot — one full revolution of the inner gimbal in about 97 seconds, easily timed with a stopwatch. At the low end of the typical operating range, 500 RPM, precession doubles to ≈7.4°/s. The motion looks rushed and the rotor visibly droops because angular momentum is too low to fully resist the torque. At the high end, 2000 RPM, precession halves to ≈1.85°/s — one revolution every ~3 minutes — which is too slow for a 50-minute lecture. Above ~2500 RPM the demonstration is technically more pure but pedagogically dead.
Result
Nominal precession rate is 3. 7°/s, or one inner-gimbal revolution every 97 seconds — exactly the cadence you want for a stopwatch-and-protractor measurement in a first-year lab. Compare that to 7.4°/s at 500 RPM (rotor droops, motion looks chaotic) and 1.85°/s at 2000 RPM (visually boring, students lose interest before one revolution completes). If the measured rate comes in noticeably faster than predicted, suspect three things in this order: (1) the 2 g weight is heavier than marked — cheap brass weights routinely run 5-10% over nominal, so weigh it on a 0.01 g scale, (2) the rotor spin has decayed faster than expected because pivot friction is high, often from dried lubricant in the agate cups, or (3) the sphere is not a uniform solid and the (2/5)mr² assumption is wrong — a hollow or drilled sphere has a different Is and you must compute it from the actual geometry.
Bohnenberger's Machine vs Alternatives
Bohnenberger's Machine sits in the same family as the demonstration top and the modern gyrocompass, but solves a different problem. Here is how it stacks up against the realistic alternatives a physics teacher or instrument restorer would actually consider.
| Property | Bohnenberger's Machine | Simple Spinning Top | Modern Electric Gyroscope (e.g. Pasco ME-8960) |
|---|---|---|---|
| Degrees of rotational freedom | 3 (full gimbal) | 1 (spin axis only, support fixed) | 3 (full gimbal, motor-driven) |
| Typical spin-up RPM | 600-1500 RPM by hand cord | 100-300 RPM by hand | 3000-6000 RPM electric |
| Demonstration duration before bleed-off | 30-90 seconds | 10-30 seconds | Indefinite while powered |
| Pivot friction sensitivity | High — needs jewel or precision ball pivots, end-play < 5 µm | Low — single contact point on table | Medium — modern bearings tolerate 10-20 µm |
| Cost (2024 teaching market) | £800-3000 restored original; £200-600 modern replica | £10-50 | £400-900 |
| Best application fit | University physics demo, history-of-science exhibit | Children's toy, intro classroom demo | Quantitative undergraduate lab |
| Ability to demonstrate Earth rotation | Yes (Foucault used this architecture in 1852) | No | Yes if rotor isolated and run long enough |
Frequently Asked Questions About Bohnenberger's Machine
That is almost always a sign that the spin rate has decayed below the point where applied torque dominates the rotor's residual angular momentum. Below roughly 200-300 RPM on a 50 mm brass sphere, pivot friction torque becomes comparable to your applied gravitational torque, and the system starts behaving like a pendulum on a slow rotor instead of a true gyroscope.
The fix is to spin up harder and check pivot friction. A clean instrument should hold useful precession behaviour for at least 60 seconds after spin-up. If yours dies in 15 seconds, the agate cups need cleaning or the steel pivot tips have mushroomed and need re-pointing on a watchmaker's lathe.
If the demonstration is the goal, buy the modern replica. Original Bohnenbergers have ivory spheres that have shrunk and gone out of round over 200 years, the brass gimbals are work-hardened and crack at the pivot bosses, and the agate cups are usually pitted. You will spend £1500-3000 on restoration before it spins cleanly.
If the historical artefact is the goal — a museum case, a department's history wall — restore the original and accept that it will only be spun for special occasions. The two use cases need different machines.
The most common cause on a hand-built replica is that the rotor is not actually a uniform solid sphere. If the sphere has a drilled axial bore for the spin axle (most do), the moment of inertia is higher than (2/5)mr² because mass is biased toward the equator. Recompute Is using the actual geometry — for a typical 4 mm bore through a 50 mm sphere the correction is small, but for a hollow or shell rotor it can be 30-50%.
Second cause: your applied torque arm is shorter than you think. The arm should be measured from the spin axis to the centre of mass of the hanging weight, not to its hook. A 2 mm error on a 30 mm arm is nearly 7%.
Gimbal lock occurs when two of the three gimbal axes align — typically when you tilt the outer ring 90° and the middle and inner ring axes become parallel. The rotor instantly loses a degree of freedom, the precession stops cleanly, and any further applied torque just rotates the whole stack as a rigid body. Students see it as the machine "freezing."
Avoid it by setting up the demonstration with the rotor spin axis at roughly 30-45° to vertical at rest, never aligned with any gimbal pivot axis. If you hit lock during the demo, just nudge the outer ring a few degrees off-axis and the rotor recovers immediately.
Yes, and most modern teaching replicas do. A pair of miniature angular-contact bearings like SKF W 619/4-2Z (4 mm bore) at each gimbal pivot gives you running friction comparable to a clean jewel pivot, with vastly better tolerance to handling. The compromise is preload — you must preload the bearings just enough to remove end-play (under 5 µm) without adding measurable drag torque.
The one place jewels still beat bearings is at the rotor axle itself, where the friction torque budget is tightest. For the outer two gimbal rings, ball bearings are the practical choice.
Bohnenberger built the gimbal architecture but never isolated the rotor well enough or spun it long enough to see the Earth-rotation effect — pivot friction killed the rotor in under two minutes, far below the ~30 minutes you need to see a measurable shift at mid-latitudes. Foucault's contribution was an electric spin-up and a much lower-friction suspension, not a new geometry.
For a teaching replica, do not attempt the Earth-rotation demonstration. The friction budget is too tight for a hand-spun rotor. Stick to the precession demonstration, which works cleanly in the time available.
References & Further Reading
- Wikipedia contributors. Johann Gottlieb Friedrich von Bohnenberger. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
