The Physical Pendulum

What a physical pendulum is

A simple pendulum is a fiction: a point mass on a massless, inextensible string. Nothing in the real world is a point mass, and every actual swinging object — a grandfather-clock pendulum with its heavy bob and rod, a wrecking ball on a boom, a metronome arm, your own leg swinging as you walk — is a physical pendulum (also called a compound pendulum): an extended rigid body of finite size that rotates about a fixed pivot axis under the torque of its own weight.

The physics is the same rotational Newton's second law you already know, τ = Iα, applied to the body about its pivot. The only real subtlety is that the restoring torque comes from the body's weight acting at its center of mass, while the resistance to angular acceleration is governed by the body's moment of inertia about the pivot — not about the center of mass. Getting those two facts right, and connecting them with the parallel-axis theorem, is the whole game.

Deriving the period

Let the pivot be at O and the center of mass (CM) at distance d from O. When the line O→CM is displaced by angle θ from vertical, gravity mg acts downward at the CM. The gravitational torque about O is

τ = − m g d sin θ

The minus sign marks it as a restoring torque — it always pushes θ back toward zero. Newton's second law for rotation about the fixed axis O reads τ = I θ̈, where I is the moment of inertia about O and θ̈ = d²θ/dt². So:

I θ̈ = − m g d sin θ

For small oscillations, sin θ ≈ θ (in radians; accurate to better than 1% for amplitudes under about 14°). The equation becomes that of a simple harmonic oscillator:

θ̈ + (m g d / I) θ = 0    ⟹    ω = √(m g d / I)

The angular frequency is ω = √(mgd/I), so the period is:

T = 2π / ω = 2π √( I / (m g d) )

Symbols and units: T is the period in seconds (s); I is the moment of inertia about the pivot in kg·m²; m is the total mass in kilograms (kg); g is the gravitational field strength in m/s² (9.81 m/s² at Earth's surface); d is the pivot-to-CM distance in meters (m). Note that m appears in both I (which scales as m) and in mgd, so the mass cancels — the period depends only on geometry and g.

The equivalent simple pendulum

Compare T = 2π√(I/(mgd)) with the simple-pendulum period T = 2π√(L/g). They match when

L_eq = I / (m d)

A simple pendulum of this length — the equivalent length — swings in perfect lockstep with the physical pendulum. Using the radius of gyration k_cm about the center of mass (defined by I_cm = m·k_cm²) and the parallel-axis theorem I = I_cm + md² = m(k_cm² + d²), the equivalent length becomes

L_eq = (k_cm² + d²) / d = d + k_cm² / d

Because k_cm²/d > 0, the equivalent length is always longer than d: the equivalent simple pendulum reaches past the center of mass, down to a point called the center of oscillation. The full small-angle period in this form is:

T = 2π √( (k_cm² + d²) / (g d) )

Center of oscillation = center of percussion

The center of oscillation is the point O′ at distance L_eq from the pivot. A striking theorem, first understood by Christiaan Huygens in his 1673 Horologium Oscillatorium, is that pivot and center of oscillation are interchangeable: if you re-hang the body from O′, it swings with exactly the same period, and the old pivot O becomes the new center of oscillation. This conjugacy is the basis of Kater's reversible pendulum (1817), which measures g to five significant figures without needing to know I, m, or d at all — only the distance between two pivots that give equal periods.

The center of oscillation coincides with the center of percussion: the point where an impulsive transverse force produces no reaction impulse at the pivot. That is the "sweet spot" of a baseball bat, tennis racket, cricket bat, or hammer — strike the ball there and no sting is transmitted to your hands at the grip. It is the same point for the same reason: both are governed by L_eq = I/(md).

Why it matters

• Timekeeping. Every pendulum clock from Huygens (1656) to the 1920s Shortt–Synchronome (accurate to ~1 s/year) is a physical pendulum. The rod has mass, the bob has size — the design tunes I and d to hit T = 2 s (a "seconds pendulum," which ticks each half-swing).
• Measuring g. Kater's pendulum and its descendants set the gravity standard for over a century; a compound pendulum turns a length and a stopwatch into a precise value of g ≈ 9.81 m/s².
• Sports and tools. Bat, racket, golf club, and axe design all hinge on placing the center of percussion where you actually hit — minimizing pivot reaction and vibration.
• Biomechanics. A walking leg swings roughly as a physical pendulum; its natural period sets the energetically cheapest stride cadence, which is why tall people walk with slower, longer strides.
• Engineering. Tuned-mass dampers, gyroscopic stabilizers, and seismic test rigs all rely on I/(md) to set a target frequency.

Worked example: a uniform rod

Take a uniform rod of length L = 1.00 m and mass m, pivoted at one end. Its moment of inertia about the center is I_cm = mL²/12, and the CM is at d = L/2. By the parallel-axis theorem:

I = I_cm + m d² = mL²/12 + m(L/2)² = mL²/3

Then

T = 2π √( I / (m g d) ) = 2π √( (mL²/3) / (m g · L/2) ) = 2π √( 2L / (3g) )

For L = 1.00 m and g = 9.81 m/s²: T = 2π√(2·1.00/(3·9.81)) = 2π√(0.06796) ≈ 1.64 s. The equivalent length is L_eq = I/(md) = (mL²/3)/(m·L/2) = 2L/3 ≈ 0.667 m — so the center of oscillation sits two-thirds of the way down the rod, well below the CM at L/2. A simple pendulum would need a string of only 0.667 m, not 1.00 m, to match this rod's swing.

Periods of common rigid bodies

Body & pivot	I about pivot	d (pivot→CM)	Period T
Point mass on string (simple)	mL²	L	2π√(L/g)
Uniform rod, pivot at end	mL²/3	L/2	2π√(2L/(3g))
Uniform rod, pivot at L/4 from end	7mL²/48	L/4	2π√(7L/(12g))
Uniform disk, pivot at rim	(3/2)mR²	R	2π√(3R/(2g))
Uniform ring/hoop, pivot at rim	2mR²	R	2π√(2R/g)
Solid sphere, pivot on surface	(7/5)mR²	R	2π√(7R/(5g))
Rod, pivot AT center of mass	mL²/12	0	∞ (no restoring torque)

The last row is the key sanity check: when d = 0 the pivot is at the CM, gravity exerts no torque, and the "pendulum" has infinite period — it does not swing at all, it just sits in neutral equilibrium.

Which pivot swings fastest?

Writing T = 2π√((k_cm² + d²)/(gd)), treat d as adjustable. Minimizing the bracket over d gives dT/dd = 0 at d = k_cm — the pivot whose distance from the CM equals the radius of gyration. There:

L_eq = 2 k_cm ,    T_min = 2π √( 2 k_cm / g )

Push the pivot toward the CM (d → 0) and T → ∞ because the restoring torque vanishes; push it far out toward the ends and I grows faster than d, so T also climbs. The sweet spot for the fastest swing is exactly one radius of gyration from the center of mass — a fact used to design the "seconds pendulum" length efficiently.

Beyond the small-angle approximation

The clean formula assumes sin θ ≈ θ. For larger amplitudes θ₀ the true period grows and depends on amplitude (the swing is isochronous only in the small-angle limit). The exact period involves a complete elliptic integral of the first kind:

T = T₀ · (2/π) K(sin(θ₀/2))
  ≈ T₀ · [ 1 + (1/16)θ₀² + (11/3072)θ₀⁴ + … ]

where T₀ = 2π√(I/(mgd)) is the small-angle period. At θ₀ = 20° (0.349 rad) the correction is about +0.76%; at 90° it is about +18%. Huygens' insight that only a cycloidal path is truly isochronous for all amplitudes is why precision clocks keep amplitudes tiny.

Common misconceptions

• "Use the moment of inertia about the center of mass." No — the formula needs I about the pivot. Use the parallel-axis theorem, I = I_cm + md², to get there. Forgetting the +md² term is the single most common error.
• "Heavier pendulums swing slower." No — mass cancels in I/(mgd). A brass and an aluminum rod of identical shape have identical periods. What matters is the mass distribution, captured by I/m, not the total mass.
• "The equivalent length equals d." No — L_eq = d + k_cm²/d, always strictly greater than d. The center of oscillation lies below the center of mass, never at it.
• "Any object hangs and swings the same way." The period changes with where you pivot. Pivoting a rod at its end (T = 2π√(2L/3g)) is not the same as pivoting at L/4; and pivoting exactly at the CM gives no oscillation at all.
• "Small-angle period works for big swings." Only for amplitudes under ~15°. Beyond that the period lengthens measurably and you need the elliptic-integral correction.
• "Center of percussion is a separate, unrelated point." For a pivoted rigid body it is identically the center of oscillation — same distance L_eq = I/(md), same physics.

A brief history

Galileo noticed the near-isochronism of a swinging lamp in Pisa Cathedral around 1602 and proposed pendulum timekeeping, but never built a working clock. Christiaan Huygens built the first pendulum clock in 1656 and, in his masterwork Horologium Oscillatorium (1673), gave the full theory of the compound pendulum — including the reciprocity of pivot and center of oscillation and the cycloidal isochrone. In 1817 Henry Kater turned that reciprocity into the reversible pendulum, the gold standard for measuring g for the next hundred years. The same mathematics later illuminated the center of percussion in the physics of impacts, tools, and sports equipment.