A cycloidal pendulum is a pendulum constrained by curved metal cheeks so its bob traces a cycloid arc rather than a circular arc. The cycloidal cheeks are the defining component — they progressively shorten the effective suspension length as the swing widens, which forces the bob along a tautochrone curve. The purpose is to eliminate circular error, the small period drift that affects every ordinary pendulum as amplitude changes. The outcome is true isochronism: equal swing time at any amplitude, a property Christiaan Huygens proved mathematically in 1659 and built into early precision clocks.
Cycloidal Pendulum Movement Interactive Calculator
Vary beat time and swing amplitudes to size the pendulum length, cycloidal cheek radius, and ordinary circular-error rate loss.
Equation Used
This calculator uses the article comparison for ordinary circular error, where a 4 degree seconds pendulum swing is about 15 seconds per day slower than a 1 degree swing. It also sizes the ideal Huygens cheek geometry from the pendulum length, using a cheek generating radius of L/4.
- Beat time is one half of the full pendulum period.
- Cycloidal cheek generating circle radius is L/4.
- Daily circular-error rule follows the worked example: 4 deg versus 1 deg gives about 15 s/day.
- Ideal cycloidal cheeks remove amplitude-dependent circular error.
How the Cycloidal Pendulum Movement Actually Works
An ordinary pendulum is only approximately isochronous. The simple T = 2π√(L/g) formula is a small-angle approximation → at amplitudes above roughly 5°, the period grows measurably because a circular arc is not a tautochrone. That growth is called circular error, and on a 1-second seconds pendulum swinging at 4° peak amplitude it adds about 15 seconds per day compared to a 1° swing. If you want a clock that keeps the same rate whether the mainspring is fully wound or nearly run-down, you have to fix that.
Huygens' answer was geometric. If you suspend a flexible cord between two metal cheeks shaped to a specific curve, and you choose that curve correctly, the bob's path is forced into a cycloid. The cycloid is the tautochrone curve — the path along which a particle takes the same time to slide to the bottom regardless of starting height. The cheek profile is itself a cycloid, half the length of the bob's path, because the involute of a cycloid is another cycloid of the same dimensions. Get the cheek geometry right and the pendulum becomes mathematically isochronous to within whatever the suspension and air drag will allow.
When the geometry is wrong, you see it in the rate. Cheeks cut too tight cause the cord to bind near top of swing, killing amplitude and dragging the rate. Cheeks cut too generous let the bob trace something closer to a circular arc, and circular error returns. Real builds also struggle with cord wear at the cheek contact points — silk loses fibres, the effective length drifts, and after a few weeks the rate creeps. That wear problem is exactly why most precision horologists abandoned cycloidal cheeks in favour of a stiff suspension spring kept to small amplitudes, where circular error is small enough to ignore.
Key Components
- Cycloidal Cheeks: Two metal plates flanking the suspension point, machined to a cycloid curve whose generating circle has radius L/4 where L is the pendulum length. They progressively shorten the active cord length as swing amplitude grows, forcing the bob onto a cycloidal path. Profile tolerance must be held to within roughly 0.05 mm across the full cheek for the isochronism to hold below 1 second per day.
- Flexible Suspension Cord: Traditionally silk or fine gut, modern reproductions use steel ribbon. The cord must be perfectly flexible so it conforms to the cheek profile during swing — any stiffness reintroduces a virtual pivot point and breaks the cycloidal constraint. Cord length equals exactly half the arc length of one cheek.
- Pendulum Bob: Concentrated mass at the end of the cord, typically lead-filled brass for thermal stability. Bob mass must dominate cord mass by at least 50:1 so the system behaves as a simple rather than compound pendulum, which is a precondition for the cycloidal geometry to predict the period correctly.
- Suspension Block: Rigid mounting that holds the two cheeks in precise mirror symmetry about the vertical. Any tilt of the block off-vertical introduces an asymmetry between left and right swings, which shows up as a beat-error audible at the escapement and a daily rate dependent on amplitude.
- Escapement Coupling: Crutch or pallet linkage that delivers impulse to the pendulum without imposing a side force. On a Huygens-style verge or anchor escapement the impulse must be tangent to the bob's cycloidal path at bottom dead centre, otherwise the impulse itself perturbs amplitude and reintroduces an effective circular error.
Industries That Rely on the Cycloidal Pendulum Movement
Cycloidal pendulums are rare in production timekeeping today because precision regulator builders found it easier to keep amplitude small than to maintain cycloidal cheeks. But the geometry still appears in physics teaching apparatus, in restored 17th-century clocks, and in any seismic or geophysical instrument where the pendulum must behave identically across a wide amplitude range. You will also see the tautochrone principle borrowed for skateboard half-pipes, roller-coaster valley profiles, and certain optical scanner mirror suspensions where equal-time motion regardless of stroke matters more than peak speed.
- Horological Restoration: Reproduction Huygens-pattern table clocks built by makers such as Buchanan of Chelmsford for private collectors, where the cycloidal cheeks are reconstructed from Huygens' 1673 Horologium Oscillatorium drawings.
- Physics Education: PASCO and Cenco tautochrone demonstration kits used in undergraduate mechanics labs to show that a ball released from any point on a cycloidal track reaches the bottom in the same time.
- Museum Conservation: Working replicas of the 1657 Salomon Coster pendulum clock at the Museum Boerhaave in Leiden, where cycloidal cheeks are part of the original specification documented in Huygens' patent.
- Seismology: Long-period reference pendulums in calibration vaults where cycloidal suspension keeps the period independent of amplitude as the building settles, used in early 20th-century Wiechert seismographs.
- Precision Mechanics: Optical galvanometer scanner suspensions where a cycloidal flexure profile linearises the angular response across a wider deflection range than a simple torsion fibre.
- Civil Engineering Demonstration: University-built tautochrone ramps demonstrating bridge-cable pendulum dynamics, including the setups at Delft TU and ETH Zurich in mechanical engineering teaching collections.
The Formula Behind the Cycloidal Pendulum Movement
The cycloidal pendulum's period depends only on the radius of the generating circle that defines the cheek profile, not on amplitude. That is the whole point — at the low end of practical amplitude (1-2°) the cycloidal pendulum and a simple pendulum agree to within a few parts per million, so the cycloid offers little benefit. At nominal amplitudes around 4-6° the cycloid starts to earn its keep, holding rate steady where a simple pendulum would lose around 12-20 seconds per day. At the high end, 15-25° amplitude, the simple pendulum has drifted minutes per day while the cycloidal pendulum still keeps theoretical rate — but only if the cheek profile is accurate over the full sweep. The sweet spot for a real build is 4-8° amplitude, where the cheek is short enough to machine accurately and the isochronism advantage is meaningful.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| T | Period of one full oscillation | s | s |
| r | Radius of the generating circle of the cycloid (= L/4, where L is the equivalent pendulum length) | m | in |
| g | Local gravitational acceleration | m/s² | ft/s² |
| L | Equivalent simple-pendulum length, equal to 4r | m | in |
Worked Example: Cycloidal Pendulum Movement in a university physics tautochrone demonstration
A teaching technician at a German Gymnasium physics department is building a lecture-bench cycloidal pendulum to demonstrate isochronism alongside a matched simple pendulum. The brief calls for a 1-second period (T = 1.0 s, half-period beat) so it ticks in time with a wall clock for visual comparison. The technician needs to fix the cycloidal generating-circle radius r, then verify that the period stays constant across the full demonstration amplitude range from 2° (gentle release) through 6° (typical lecture release) up to 20° (deliberate large-swing demonstration).
Given
- T = 1.0 s
- g = 9.81 m/s²
- Amplitude range = 2 to 20 degrees
Solution
Step 1 — solve the cycloidal period formula for the generating-circle radius r at the nominal target period of 1.0 s:
So the cheeks must be machined to a cycloid generated by a circle of radius 62.1 mm, and the cord length equals half the arc of one cheek, which for a cycloid equals 4r = 248.4 mm. At 6° nominal amplitude the bob traces the cycloid cleanly and the period sits at 1.000 s.
Step 2 — at the low end of the demonstration range, 2° amplitude, compare against an equivalent simple pendulum of length L = 4r = 0.2484 m:
At 2° the simple pendulum and the cycloidal pendulum agree to better than 7 seconds per day. The cycloid is doing almost no useful work — the cheeks are essentially decorative at this amplitude.
Step 3 — at the high end, 20° amplitude, the simple pendulum's circular-error correction term grows sharply:
That is a 7.5 ms gain per swing, or 651 seconds per day — roughly 11 minutes lost on a wall clock by the simple pendulum. The cycloidal pendulum, in theory, still reads 1.0000 s. In a real lecture-bench build with silk cord and brass cheeks, you'll measure something like 1.0001-1.0003 s at 20° because of cord stiffness and cheek-profile machining error, giving residual drift of 10-25 seconds per day — still 25× better than the simple pendulum at the same amplitude.
Result
The cycloidal cheeks must be cut from a generating circle of radius r = 62. 1 mm with a suspension cord of 248.4 mm, giving a nominal period of 1.000 s at 6° amplitude. At 2° the cycloid offers no measurable benefit over a simple pendulum, at 6° it starts to earn its place with sub-second daily drift, and at 20° it preserves rate where the simple pendulum loses 11 minutes per day. If you measure a period drift larger than 25 s/day at full amplitude, the most likely causes are: (1) cheek profile machined as a circular arc rather than a true cycloid — a common shortcut that reintroduces full circular error, (2) silk cord with too much bending stiffness so it bridges across the cheek instead of conforming to it, shortening the effective active length unpredictably, or (3) cheek mounting block tilted off-vertical by more than 0.5°, which makes the left and right swings asymmetric and shows up as a beat error before it shows up as a rate error.
Choosing the Cycloidal Pendulum Movement: Pros and Cons
Cycloidal cheeks are not the only path to isochronism, and for most precision clocks they are not even the preferred one. The trade is between geometric correction (cycloidal cheeks) and amplitude restriction (small-swing simple pendulum) versus electronic correction (modern quartz disciplining). Here is how they compare on the dimensions a horologist or instrument builder actually decides on.
| Property | Cycloidal Pendulum | Simple Pendulum (small amplitude) | Quartz-Disciplined Pendulum |
|---|---|---|---|
| Theoretical accuracy across amplitude range | Isochronous at all amplitudes | Requires amplitude held below 2° for sub-1 s/day rate | Independent of pendulum amplitude entirely |
| Practical accuracy (s/day) | 10-30 s/day in real builds | 1-5 s/day with stable amplitude | <0.1 s/day |
| Build complexity | High — cheek profile must be cut to 0.05 mm | Low — straight knife-edge or spring suspension | High — requires electronics and reference oscillator |
| Maintenance interval | 3-6 months — silk cord wears at cheek contact | 5-10 years between suspension-spring service | Years — battery only |
| Cost (typical 2024 build) | £800-2500 for hand-cut cheeks and bob | £200-600 for an equivalent regulator suspension | £300-1500 for a Hipp-toggle or quartz-locked build |
| Best application fit | Demonstration, restoration, museum work | Production precision clocks and regulators | Modern observatory and lab timekeeping |
Frequently Asked Questions About Cycloidal Pendulum Movement
Two reasons, both practical. First, the silk or gut suspension cord wears at the cheek contact points within weeks of running, and that wear changes the effective length unpredictably — you trade circular error for cord-wear error. Second, by the early 18th century clockmakers realised they could simply restrict amplitude to 1-2° using a heavy bob and a stiff suspension spring. At that amplitude the residual circular error is below 1 s/day, which beats what a real cycloidal cheek build delivers in service. George Graham's deadbeat escapement made small-amplitude operation easy, and cycloidal cheeks fell out of use in production work.
The profile error budget scales with the swing amplitude you actually use. For a 6° demonstration build, holding the cheek profile to 0.1 mm gives sub-1 minute per day rate stability. For a 15-20° large-swing demonstration, you need profile accuracy below 0.05 mm to keep the daily rate within 30 seconds. The error is concentrated at the upper portion of the cheek where the cord contacts only at extreme swing — a profile error there only affects the largest amplitudes, which is why low-amplitude builds are forgiving.
Only partially. A steel ribbon has bending stiffness, so it doesn't perfectly conform to the cheek profile — it bridges across the curve and pivots from a virtual point that is not on the cheek surface. The result is that you recover most of the cycloidal benefit at small amplitudes but lose it progressively as amplitude grows, because the ribbon resists wrapping tighter against the cheek. If you must use ribbon for durability, machine the cheeks to a profile that is slightly flatter than a true cycloid to compensate, and accept that the build will be tuned for one specific amplitude rather than truly isochronous.
The cheeks are doing their job — the rate change you're seeing comes from the escapement, not the pendulum. Mainspring tension changes the impulse delivered through the verge or anchor, which changes the bob amplitude. With cycloidal cheeks, the period should be amplitude-independent, so the clock should not drift. If it does, the most likely cause is a side-component in the impulse delivery: the crutch pin is pushing the bob off the cycloidal path, or the pallet impulse is not tangent to the bob's velocity at bottom dead centre. Check the pallet geometry against Huygens' original drawings — modern reproductions often get the pallet angles wrong by 1-2°.
Build the cycloidal pendulum, but pair it with a matched simple pendulum on the same bench. The whole pedagogical point is the visible comparison — a small-amplitude regulator running at 1° looks identical to a circular-arc pendulum at 1°, and students can't see what Huygens solved. With a side-by-side rig at 15-20° amplitude, the simple pendulum visibly drifts ahead of the cycloidal one within a single lecture. PASCO sells exactly this paired demonstration; it works because the difference becomes obvious without needing a stopwatch resolution.
Friction at the cheek contact. Every time the cord wraps and unwraps against the cheek profile, you lose energy to bending and to sliding friction at the contact line. A simple pendulum on a knife-edge or spring suspension only loses energy to air drag and pivot friction, which is far less. Expect a cycloidal pendulum's Q factor to be 30-50% lower than an equivalent simple pendulum. If the amplitude decay is much faster than that — halving in under 60 seconds for a metal bob — check that the cord is genuinely flexible silk rather than a stiff synthetic, and that the cheek surface is polished rather than left with machining marks.
No — cycloidal cheeks fix amplitude-driven period error, not temperature-driven period error. Those are independent problems. A brass-cord suspension lengthens by roughly 19 ppm/°C, which on a 1-second pendulum gives 0.8 s/day per °C. To handle that you need a separate temperature compensation scheme — a gridiron rod, an Invar suspension, or a mercury-jar bob. Huygens-era clocks combined cycloidal cheeks with seasonal hand-regulation, which is why their certified rates were given for a specific temperature band.
References & Further Reading
- Wikipedia contributors. Cycloid. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
