A balance wheel and hairspring form an oscillating time base in many mechanical watches and clocks. The wheel provides rotational inertia; the spring provides restoring torque. Explore their ideal harmonic motion by changing full oscillations per second and one-sided swing amplitude. A selected frequency alone does not establish the accuracy of a complete watch.
Balance Wheel Interactive Calculator
Vary the oscillator frequency and swing angle to see period, beats per hour, stiffness-to-inertia ratio, and the animated balance motion.
Equation Used
The balance wheel and hairspring are modeled as a torsional harmonic oscillator. The worked example states T = 2*pi*sqrt(I/k) and a typical watch frequency of 4 Hz, which gives 28,800 beats per hour because each full oscillation has two beats.
- Ideal linear torsional oscillator.
- Frequency is full oscillations per second.
- One oscillation produces two escapement beats.
- Amplitude does not change period in the ideal isochronous model.
Swing angle is one-sided amplitude about equilibrium. Two beats make one complete oscillation. The linear torsional oscillator model omits drive, damping and amplitude-dependent spring effects.
Inertia and restoring torque
The balance rotates about its staff. The hairspring’s inner attachment turns with the staff while its outer attachment remains fixed. As the balance moves away from equilibrium, the spring stores elastic energy and produces a torque that tends to bring it back.
In the linear model, restoring torque is proportional to angle and opposite in direction. Inertia carries the balance through equilibrium into the opposite excursion. This gives sinusoidal motion with a period independent of amplitude under the model’s assumptions.
A real timepiece needs an escapement or another drive to replace energy lost through friction and air resistance. The exact impulse and locking arrangement depends on the escapement. The animation concentrates on the balance and spring; it does not count escape-wheel teeth or solve drive energy.
Components
- Balance: supplies moment of inertia about the staff.
- Hairspring: supplies restoring torque.
- Staff and bearings: locate the rotating assembly.
- Inner collet and outer stud: connect the spring to the moving and fixed parts.
A portable mechanical oscillator
Balance-and-spring oscillators are used in mechanical watches and some clocks. They do not require the long hanging member of a pendulum, but real timekeeping can still vary with position, temperature, magnetic effects, amplitude and the surrounding movement.
IWC’s explanation of balance adjustment distinguishes complete oscillations from beats: a 4 Hz movement has eight beats per second. That convention gives 28,800 beats per hour; it is not an accuracy certification.
Some designs regulate effective spring length; others adjust balance inertia. A regulator does not inherently remove mass from the wheel. Actual adjustment and service depend on the particular movement.
Ideal torsional oscillator
For moment of inertia I and linear torsional stiffness k, T = 2π√(I/k), and f = 1/T. With I in kg·m² and k in N·m/rad, the period T is in seconds. Here f means complete oscillations per second, not individual beats.
The calculator takes frequency as an input and returns T = 1/f, beats per hour = 7200f, and k/I = (2πf)² in s⁻². It does not determine k and I separately.
For one-sided amplitude A in degrees, ideal angle is A sin(2πft), and peak angular speed is 2πfA degrees/s. Peak-to-peak excursion is 2A. Increasing A increases peak speed without changing period in this ideal linear model.
Real spring behavior, damping, escapement impulses, bearing losses and positional effects are omitted. The drawing’s spring shape is illustrative, not a stress calculation or an amplitude rating.
Worked example: four complete oscillations per second
At f = 4 Hz, period is 0.25 s, beat rate is 28,800 per hour and k/I = 631.6547 s⁻². With one-sided amplitude 270°, peak angular speed is 6785.8401°/s. At amplitude 180°, peak speed falls to 4523.8934°/s while the other three outputs remain unchanged.
For a separate hypothetical inertia I = 1.10 × 10⁻⁹ kg·m², the required ideal stiffness is k = I(2πf)² = 6.9482 × 10⁻⁷ N·m/rad. This is arithmetic for a chosen inertia, not a specification for replacement spring stock.
If stiffness alone changes by −5%, frequency becomes 4√0.95 = 3.89872 Hz. At +5%, it becomes 4√1.05 = 4.09878 Hz. Relative to an otherwise unchanged train calibrated for 4 Hz, those constant frequency errors correspond to approximately −2187.7 and +2133.7 seconds per day.
Increasing inertia lowers frequency at fixed stiffness; it would worsen the first example’s slow rate. These relationships explain sensitivity, not a repair procedure or the available range of a real regulator.
What this oscillator model explains
| Change | Ideal consequence | Not established |
|---|---|---|
| Higher k at fixed I | Higher frequency. | Spring compatibility or positional performance. |
| Higher I at fixed k | Lower frequency. | Allowable balance modifications. |
| Higher amplitude at fixed f | Higher peak speed. | Real isochronism, power reserve or safe clearance. |
| Higher chosen frequency | More beats per hour. | Better accuracy, service life or shock resistance. |
Compare actual timepieces using their tested performance and manufacturer information. Frequency alone does not rank mechanical, quartz or tuning-fork designs by accuracy or maintenance needs.
Balance and hairspring questions
Two under the watch beat convention used here. At 4 Hz that is eight beats per second or 28,800 per hour.
Amplitude and period are independent in the ideal linear torsional oscillator. The peak-speed output still changes with amplitude.
No. It is one-sided amplitude from equilibrium. The full peak-to-peak excursion is twice that angle.
No. It returns stiffness divided by inertia, not spring dimensions, material selection, terminal shape or compatibility.
No. A regulator may change effective spring length. Inertia adjustment is a different mechanism and depends on the balance design.
No. Accuracy also depends on the oscillator, escapement, adjustment, temperature, position and disturbances. Those effects are not calculated.
No. Magnetism is one possible cause among several. Proper movement-specific examination is needed; the animation is not a diagnostic test.
No. Measurement setup, lift angle, position, winding state and the movement’s specifications matter. The example amplitude is not a service threshold.
Further reading
- OpenStax, Pendulums, including the ideal torsional oscillator.
- IWC, The balance wheel: a matter of adjustment, on balance assembly and frequency.
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
