Unit overview
Classical Mechanics II reformulates motion using energy, constraints and symmetry. Lagrangian and Hamiltonian methods reveal conserved quantities, simplify constrained systems and provide the conceptual bridge to statistical and quantum mechanics. The course extends particle dynamics to central-force orbits, rotating bodies and coupled degrees of freedom.
Level: Undergraduate university physics. Prerequisites: introductory mechanics, algebra, trigonometry, calculus and basic laboratory or mathematical reasoning appropriate to the topic. Study approach: work through each concept, reproduce the derivations, attempt the numerical examples independently, then complete the examination questions.
Learning outcomes
- Define the central physical quantities and use SI units consistently.
- Derive and apply governing equations while stating assumptions and validity limits.
- Interpret diagrams, graphs, mathematical models and measurable consequences.
- Solve multi-step numerical and conceptual problems with clear working.
- Evaluate uncertainty, approximations, boundary conditions and experimental limitations where relevant.
- Communicate conclusions using correct notation, units and physical reasoning.
Core textbook content
Classical Mechanics II reformulates motion using energy, constraints and symmetry. Lagrangian and Hamiltonian methods reveal conserved quantities, simplify constrained systems and provide the conceptual bridge to statistical and quantum mechanics. The course extends particle dynamics to central-force orbits, rotating bodies and coupled degrees of freedom. The sections below develop the main concepts from definitions to applications. For each topic, identify the system, list known quantities, choose a model, state approximations, and test the result against units and physical limits.
1. Generalised coordinates and constraints
A system with N particles has 3N Cartesian coordinates, but constraints reduce the independent degrees of freedom. Generalised coordinates q_i describe the configuration with the smallest convenient set of independent variables. Holonomic constraints can be written f(q,t)=0; non-holonomic constraints cannot generally be reduced to such coordinate equations. Ideal constraint forces do no virtual work for allowed virtual displacements. Selecting coordinates that encode constraints avoids solving unnecessary reaction forces.
2. The principle of stationary action
The action S=∫L(q,q̇,t)dt is stationary under small variations of the path with fixed endpoints. The Lagrangian is often L=T−V for conservative systems. Stationarity does not always mean a minimum; it may be a maximum or saddle. Applying the calculus of variations yields the Euler–Lagrange equations. This formulation is coordinate-independent in its physical content and naturally incorporates constraints through suitable coordinates or multipliers.
3. Euler–Lagrange equations and examples
For each generalised coordinate q_i, d/dt(∂L/∂q̇_i)−∂L/∂q_i=0. If a coordinate is absent explicitly from L, its conjugate momentum is conserved. If L has no explicit time dependence, the energy function is conserved under appropriate regularity conditions. The equations can be more efficient than Newton's laws when constraints are complicated or coordinates are curvilinear.
4. Noether's theorem and conservation laws
Continuous symmetries of the action correspond to conserved quantities under Noether's theorem. Invariance under spatial translation gives linear momentum conservation; rotational invariance gives angular momentum conservation; time-translation invariance gives energy conservation. The theorem identifies the structural origin of conservation laws. External fields, time-dependent constraints or symmetry-breaking potentials may remove the corresponding conservation law.
5. Central-force motion and effective potential
A central force depends only on radial distance and points along the radius vector. Angular momentum is conserved, so motion lies in a plane. The radial equation can be written using an effective potential V_eff(r)=V(r)+L²/(2mr²), where L is angular momentum magnitude. Turning points occur where total energy equals V_eff. The centrifugal term is a consequence of angular momentum conservation, not an additional physical force in an inertial frame.
6. Kepler problem and orbital dynamics
For gravitational attraction between two masses, the relative motion reduces to a one-body problem with reduced mass μ=m₁m₂/(m₁+m₂). Bound inverse-square orbits are ellipses with the force centre at one focus. The orbit equation is r(φ)=p/[1+e cos(φ−φ₀)], where p=L²/(μk) and eccentricity e distinguishes circular, elliptical, parabolic and hyperbolic trajectories. Kepler's laws follow from the inverse-square force and angular momentum conservation.
7. Scattering and impact parameters
Scattering describes trajectories deflected by a central potential. The impact parameter b is the perpendicular offset of the incoming asymptote from the force centre. The scattering angle depends on b, energy and potential. In inverse-square Coulomb or gravitational scattering, the Rutherford formula connects differential cross-section to the interaction strength and scattering angle. Cross-sections describe probabilities per incident flux, not literal geometric target areas. Classical predictions fail when de Broglie wavelengths or quantum effects become important.
8. Rotating reference frames
A rotating frame introduces apparent acceleration terms when Newton's laws are expressed in that frame. For angular velocity Ω, acceleration includes Coriolis −2Ω×v_rel, centrifugal −Ω×(Ω×r), and Euler −dΩ/dt×r contributions in addition to relative acceleration. These terms are not new fundamental interactions; they arise from the accelerating coordinates. They explain deflection in rotating systems, including atmospheric circulation and laboratory turntables.
9. Rigid-body kinematics and angular velocity
A rigid body preserves distances between its particles. Its motion can be decomposed into centre-of-mass translation and rotation about the centre of mass. Angular velocity ω defines instantaneous rotational motion, with particle velocity v=V_cm+ω×r′. Angular acceleration is dω/dt; in general, ω and angular momentum need not be parallel for an asymmetric body. Euler angles or rotation matrices describe orientation, but coordinate singularities can occur in angle parameterisations.
10. Moment of inertia and inertia tensor
For rotation about a fixed axis, I=Σm_ir_i² or I=∫r_perp²dm. Rotational kinetic energy is ½Iω². The inertia tensor generalises this to three dimensions and relates angular momentum to angular velocity: L=I·ω. Principal axes diagonalise the tensor. The parallel-axis theorem I=I_cm+Md² shifts the axis; the perpendicular-axis theorem applies to planar laminae. Moment of inertia depends on mass distribution and axis, not only total mass.
11. Euler equations and torque-free rotation
In body-fixed principal axes, Euler's equations are I₁ω̇₁+(I₃−I₂)ω₂ω₃=τ₁, with cyclic permutations. Torque-free rotation conserves angular momentum and kinetic energy, but ω can change direction relative to the body. Rotation about the largest or smallest principal moment is stable to small perturbations, while rotation about the intermediate principal axis is unstable. This is the tennis-racket or Dzhanibekov effect.
12. Coupled oscillations and normal modes
For small oscillations around stable equilibrium, expand potential energy to quadratic order. The equations become M q̈+Kq=0. Seeking q=a e^(iωt) gives (K−ω²M)a=0; non-trivial modes require det(K−ω²M)=0. Eigenvectors specify relative amplitudes and phases, while eigenvalues give squared normal-mode frequencies. Orthogonality is defined with respect to the mass matrix for suitable symmetric systems.
13. Hamiltonian mechanics and phase space
Define canonical momentum p_i=∂L/∂q̇_i and Hamiltonian H=Σp_iq̇_i−L. Hamilton's equations are q̇_i=∂H/∂p_i and ṗ_i=−∂H/∂q_i. Phase space contains positions and momenta; a trajectory represents the complete dynamical state. For many conservative systems H equals total energy, but this is not automatic for every coordinate choice or explicitly time-dependent Lagrangian.
14. Poisson brackets and canonical transformations
The Poisson bracket is {f,g}=Σ(∂f/∂q_i ∂g/∂p_i−∂f/∂p_i ∂g/∂q_i). Time evolution is df/dt={f,H}+∂f/∂t. Canonical transformations preserve the Hamiltonian structure and Poisson-bracket relations. These methods prepare the way for Hamilton–Jacobi theory and quantum commutators. Care is required with generating functions and transformed momenta.
15. Hamilton–Jacobi theory and action variables
Hamilton–Jacobi theory seeks a generating function S(q,t) satisfying ∂S/∂t+H(q,∂S/∂q,t)=0. It can reduce dynamics to quadratures when separation is possible. For periodic systems, action variables J=(1/2π)∮p dq are conserved adiabatic invariants under slow parameter changes. These ideas connect classical integrability, semiclassical quantisation and perturbation theory.
16. Small oscillations and stability
Equilibrium is stable if small displacements produce restoring dynamics and the potential has a local minimum in allowed directions. The Hessian matrix of second derivatives of V, combined with the mass matrix, determines small-oscillation frequencies. Zero eigenvalues can indicate continuous symmetry or marginal directions; negative eigenvalues indicate instability. Linear stability is local and does not establish global stability.
17. Integrated problem: rolling without slipping
For a rigid body rolling without slipping on a stationary surface, centre-of-mass speed and angular speed satisfy v_cm=Rω. Total kinetic energy is translational plus rotational: K=½Mv_cm²+½I_cmω². Static friction enforces the constraint and may do no work at the instantaneous contact point in ideal rolling. The constraint reduces degrees of freedom; it must be incorporated consistently into the chosen mechanics formulation.
Summary equations and principles
Model → assumptions → governing equation → boundary / initial data → solution → dimensional and physical checks.
Use the equations introduced in each section as a connected framework rather than as isolated formulae. Define every symbol, retain units through calculations and identify whether quantities are vectors, scalars, fields, rates or integrated totals. Check special cases and conservation statements before accepting a numerical answer.
University examination and problem set
Attempt all questions. Show derivations, diagrams, units and intermediate steps. Answers should distinguish exact results from approximations and should state assumptions explicitly.
- 1. Define the central quantities and state the assumptions used in classical mechanics ii.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 2. Derive the principal governing relation for one of the systems discussed in this unit.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 3. Explain how the choice of model or coordinate system affects the solution.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 4. A measured or calculated quantity changes by 8%. Determine the effect on a dependent quantity using the relevant sensitivity relation.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 5. Compare two limiting cases and explain the physical meaning of the difference.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 6. Draw and label a diagram showing the geometry, directions and relevant variables for a representative problem.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 7. Identify three sources of systematic limitation and explain how each could be reduced.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 8. Construct a graph that would test a predicted linear relationship. State the expected gradient and intercept.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 9. Solve a multi-step numerical problem using consistent SI units and show intermediate working.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 10. Explain where the approximation used in the unit would fail and propose a more complete model.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 11. Relate a mathematical result to an observable experimental consequence.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 12. Evaluate a proposed experimental method, including controls, uncertainty and safety where relevant.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 13. Show how a conservation law or boundary condition constrains the solution.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 14. Interpret a non-ideal result and distinguish measurement effects from failure of the theoretical model.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
- 15. Write a concise scientific conclusion that includes the result, uncertainty or limitations, and physical interpretation.
Answer guidance: Define symbols and assumptions; show the governing equation; substitute values with units; check dimensions and limiting behaviour; state the physical interpretation. For explanatory items, support the argument with a labelled sketch or a relevant equation.
Independent study and further reading
Use a recognised university-level physics textbook and laboratory manual alongside these notes. Compare notation and sign conventions across sources, but do not combine formulae until their conventions and assumptions have been reconciled. For deeper study, locate peer-reviewed papers or university lecture material on the specific physical model, then reproduce at least one derivation or numerical result independently.
For revision, prepare a one-page concept map linking definitions, governing equations, experimental observables, limiting cases and common failure modes. Rework the examples with altered numerical values and explain why the qualitative outcome does or does not change.
CBE Nexus · Undergraduate Physics · Unit 9 of 36 · Educational resource. Check the assumptions and conventions stated in each derivation before applying equations to a different physical system.
9.1 Generalised coordinates, constraints and configuration space
A mechanical system is described by a configuration: the complete set of positions and orientations needed to specify its state at one instant. For N particles in three-dimensional space, an unconstrained configuration requires 3N Cartesian coordinates. Holonomic constraints reduce the number of independent coordinates. A constraint equation f(q1,…,q3N,t)=0 removes one degree of freedom when it is independent of the others. The number of degrees of freedom is therefore the number of independent configuration variables after all independent constraints have been imposed. Non-holonomic constraints may involve velocities in a way that cannot be integrated into a relation among coordinates alone; rolling without slipping is a standard example. The distinction matters because holonomic constraints can normally be built directly into a minimal coordinate set, whereas non-holonomic constraints require additional equations or variational procedures.
Generalised coordinates q_i are any independent variables that uniquely specify the configuration locally. They may be lengths, angles, or dimensionless parameters. A simple pendulum of fixed length l has one coordinate θ, with x=l sinθ and y=−l cosθ. A bead constrained to a circular wire also has one angular coordinate. A rigid body in three dimensions requires three coordinates for orientation and three for translation, although Euler angles have coordinate singularities and quaternions provide an alternative representation in computational mechanics. Generalised velocities are qdot_i. The position of particle a is r_a(q,t), so its velocity follows by the chain rule: v_a=Σ_i (∂r_a/∂q_i)qdot_i + ∂r_a/∂t. The final term is essential for moving constraints and explicitly time-dependent coordinate transformations.
The kinetic energy T=½Σ_a m_a v_a·v_a becomes a quadratic form in generalised velocities, often T=½Σ_ij M_ij(q,t)qdot_i qdot_j plus possible linear and coordinate-only terms. The matrix M is the mass matrix; it is symmetric and positive definite for independent physical coordinates with positive masses. Its entries encode geometry and coupling. Potential energy V(q,t) describes conservative forces, and the Lagrangian is L=T−V. The units of L are joules, but L is not itself the total energy in every explicitly time-dependent problem. The action S=∫L dt has units joule-seconds and is stationary under permitted variations of the path with fixed endpoints.
Worked example: a particle of mass m moves on a frictionless circular hoop of radius a in a vertical plane. Choose θ measured from the lowest point. Its speed is a θdot, so T=½ma²θdot². Its height above the lowest point is h=a(1−cosθ), hence V=mga(1−cosθ). Thus L=½ma²θdot²−mga(1−cosθ). The coordinate θ alone specifies the configuration, so there is one degree of freedom. This example shows how geometry can remove constraint forces from the final equation: the normal force does no virtual work along the allowed tangential displacement and need not be solved explicitly.
When choosing coordinates, prefer a minimal independent set, exploit symmetry, and avoid singular parameterisations near the configuration of interest. Check that the coordinate transformation is invertible locally. If a coordinate is cyclic, its conjugate momentum is conserved, which can reduce the order of the equations. Always state the coordinate convention, positive direction, reference level for potential energy and whether constraints are fixed or moving. These choices prevent sign errors and make a derivation reproducible.
9.2 Principle of stationary action and Euler–Lagrange equations
The action for a path q_i(t) between fixed endpoint configurations at t1 and t2 is S[q]=∫_{t1}^{t2}L(q,qdot,t)dt. The physical path makes the first variation δS vanish for all sufficiently small admissible variations that leave the endpoints fixed. Write q_i→q_i+εη_i, where η_i(t1)=η_i(t2)=0. To first order, δS=Σ_i∫[(∂L/∂q_i)η_i+(∂L/∂qdot_i)ηdot_i]dt. Integrating the second term by parts gives a boundary contribution [ (∂L/∂qdot_i)η_i ] from t1 to t2, which vanishes because the endpoint variations are zero. The remaining integral is Σ_i∫[∂L/∂q_i−d/dt(∂L/∂qdot_i)]η_i dt. Since the functions η_i are arbitrary in the interior, each coefficient must vanish. Therefore d/dt(∂L/∂qdot_i)−∂L/∂q_i=0. These are the Euler–Lagrange equations.
The derivation is more than a mathematical trick. It gives equations of motion without resolving every force into Cartesian components, and it handles constraints naturally when coordinates are chosen appropriately. For a Cartesian coordinate x with L=½m xdot²−V(x), the equation is m xddot+ dV/dx=0, equivalent to Newton's second law because F=−dV/dx. For a coordinate θ, the generalised force is not necessarily an ordinary force; it is the quantity conjugate to virtual displacement δθ and has units of energy when θ is dimensionless.
For a Lagrangian L(q,qdot,t), define the conjugate momentum p_i=∂L/∂qdot_i. It need not equal m qdot_i in arbitrary coordinates. In polar coordinates for a particle, p_r=m rdot and p_φ=m r² φdot. The angular momentum about the origin is p_φ. If ∂L/∂q_i=0, then q_i is cyclic and dp_i/dt=0. This is a direct route from continuous symmetry to conservation laws.
Worked example: a simple pendulum of mass m and length l has L=½ml²θdot²−mgl(1−cosθ). Then ∂L/∂θdot=ml²θdot, whose time derivative is ml²θddot. Also ∂L/∂θ=−mgl sinθ. The equation is ml²θddot+mgl sinθ=0, or θddot+(g/l)sinθ=0. For small oscillations, sinθ≈θ and the equation becomes θddot+(g/l)θ=0, with angular frequency ω0=√(g/l). The approximation requires |θ|≪1 radian; the exact equation is nonlinear.
If L depends explicitly on time, energy need not be conserved. If L is invariant under a continuous coordinate transformation, a corresponding Noether quantity is conserved. Boundary terms and total time derivatives deserve care: replacing L by L+dF(q,t)/dt changes the action only by endpoint terms and leaves the Euler–Lagrange equations unchanged for fixed endpoints. This freedom is useful in electromagnetic mechanics and canonical transformations. In derivations, distinguish a stationary action from a minimum: the physical path can be a saddle point rather than a strict minimum.
9.3 Generalised forces, virtual work and dissipative forces
A virtual displacement δr_a is an infinitesimal displacement compatible with the constraints at a fixed instant. The virtual work of applied forces is δW=Σ_a F_a·δr_a. Since δr_a=Σ_i(∂r_a/∂q_i)δq_i, one obtains δW=Σ_i Q_i δq_i, where Q_i=Σ_a F_a·∂r_a/∂q_i is the generalised force. This definition is coordinate independent in meaning: Q_i is the force-like quantity conjugate to q_i. If q_i is a length, Q_i has units newtons; if q_i is an angle, Q_i has units newton-metres.
For conservative forces, Q_i=−∂V/∂q_i. Non-conservative forces such as viscous drag, friction, and externally applied torques can be included through Q_i^(nc) on the right side of the Lagrange equation: d/dt(∂T/∂qdot_i)−∂T/∂q_i+∂V/∂q_i=Q_i^(nc), for the usual natural-coordinate form. Ideal constraint forces do no virtual work for allowed virtual displacements and therefore do not appear explicitly. This is why a pendulum can be solved without first calculating the tension in its string.
Linear viscous damping F_d=−b xdot can be represented by the Rayleigh dissipation function R=½b xdot². For several coordinates, R=½Σ_ij c_ij qdot_i qdot_j when the damping matrix is symmetric and positive semidefinite. The dissipative generalised force is −∂R/∂qdot_i, so the equations become d/dt(∂L/∂qdot_i)−∂L/∂q_i+∂R/∂qdot_i=Q_i^(other). This representation is convenient for linear velocity-dependent damping, but Coulomb friction and dry friction are not generally described by a quadratic Rayleigh function.
Worked example: a mass m attached to a spring k moves on a horizontal surface with viscous coefficient b. Let x be displacement from equilibrium. L=½m xdot²−½k x² and R=½b xdot². The equation is m xddot+kx+b xdot=0. Dividing by m gives xddot+2β xdot+ω0²x=0, where β=b/(2m) and ω0=√(k/m). The damping ratio is ζ=β/ω0=b/(2√(mk)). If ζ<1 the motion is underdamped; ζ=1 is critical damping; ζ>1 is overdamped. The characteristic roots are −β±√(β²−ω0²), which determine the transient response.
Virtual work also clarifies mechanical advantage and constraint reactions. For an ideal pulley, a single independent coordinate can describe the motion of connected masses; the string constraint imposes equal displacement magnitudes. Tension is internal to the constrained system and can be eliminated from the reduced equation. To recover tension, substitute the acceleration into one particle's Newton equation. In a rolling body, the static friction force can do zero instantaneous work at the contact point in pure rolling, even though it may be essential for enforcing the no-slip constraint. Do not assume every friction force dissipates energy: kinetic friction usually does, while ideal static friction may transfer energy without dissipating it.
For non-ideal constraints, the reaction force may do virtual work and must be included. Lagrange multipliers introduce constraint equations f_α(q,t)=0 into the variational problem. The augmented Lagrangian is L*=L+Σ_α λ_α f_α. The multipliers are solved together with the coordinates and often correspond to constraint forces. Their physical interpretation depends on the form and units of each constraint equation.
9.4 Noether theorem and conservation laws
Noether's theorem states that every differentiable continuous symmetry of the action, subject to its stated conditions, produces a conserved quantity. Consider an infinitesimal transformation q_i→q_i+εΔq_i and t→t+ετ. If the Lagrangian changes only by a total derivative ε dF/dt, the quantity J=Σ_i p_iΔq_i−Hτ−F is conserved, where p_i=∂L/∂qdot_i and H=Σ_i p_i qdot_i−L. For a transformation at fixed time, τ=0, and a symmetry with no boundary term, J=Σ_i p_iΔq_i.
Time-translation symmetry yields energy conservation when L has no explicit time dependence. Spatial translation symmetry yields linear momentum conservation when the system is isolated and the Lagrangian is unchanged by a uniform displacement. Rotational symmetry yields angular momentum conservation. These statements explain why conservation laws arise from symmetry rather than being unrelated empirical rules. A central potential V(r) is rotationally invariant, so angular momentum is conserved and the orbit lies in a fixed plane.
The energy function E=Σ_i qdot_i(∂L/∂qdot_i)−L is constant if L has no explicit time dependence. If T is a homogeneous quadratic function of velocities and V is independent of velocities, then Euler's theorem for homogeneous functions gives Σ_i qdot_i∂T/∂qdot_i=2T, and E=T+V. This equality can fail for explicitly time-dependent coordinates, velocity-dependent potentials or non-standard Lagrangians; one should derive the energy function rather than assume it.
Worked example: a particle of mass m moves in a plane under a central potential V(r). In polar coordinates L=½m(rdot²+r²φdot²)−V(r). The coordinate φ is absent explicitly, so p_φ=∂L/∂φdot=mr²φdot is constant. This is the angular momentum magnitude perpendicular to the plane. The radial equation is m rddot−mrφdot²+V'(r)=0. Substituting φdot=ℓ/(mr²) gives m rddot=ℓ²/(mr³)−V'(r). The conserved angular momentum reduces the two-coordinate problem to one radial equation.
Conservation laws are also diagnostic tools. If a numerical simulation of a conservative system exhibits systematic energy drift, the time step may be too large, the integrator may not be symplectic, or the implementation may contain an error. In an experiment, measured momentum may not be conserved if an external impulse, friction, or an unobserved degree of freedom is present. Define the system boundary before invoking a conservation law. Conservation applies to the isolated system, not necessarily to a chosen subsystem.
Symmetry can be approximate. A weak anisotropy breaks exact rotational symmetry and causes slow precession; a weak external torque changes angular momentum gradually. The corresponding balance law is dJ/dt equal to the generalised torque or symmetry-breaking contribution. This provides a controlled way to understand perturbations rather than treating conservation as all-or-nothing.
9.5 Hamiltonian mechanics and Legendre transformation
Hamiltonian mechanics replaces the coordinates and velocities by coordinates and conjugate momenta. For a regular Lagrangian, the relation p_i=∂L/∂qdot_i can be inverted to obtain qdot_i(q,p,t). The Hamiltonian is the Legendre transform H(q,p,t)=Σ_i p_i qdot_i−L(q,qdot,t), with velocities expressed in terms of q and p after the transformation. The regularity condition requires the Hessian matrix ∂²L/∂qdot_i∂qdot_j to be nonsingular. Singular Lagrangians occur in constrained systems and gauge theories and require more advanced treatment.
Hamilton's equations are qdot_i=∂H/∂p_i and pdot_i=−∂H/∂q_i. They form a first-order system in phase space, whose point consists of all q_i and p_i. The phase-space trajectory contains both configuration and momentum information. For a natural one-dimensional system with L=½m qdot²−V(q), p=m qdot and H=p²/(2m)+V(q), the familiar total mechanical energy is recovered. Hamilton's equations give qdot=p/m and pdot=−V'(q), which combine to Newton's equation.
Worked example: for a harmonic oscillator L=½m xdot²−½kx². The momentum is p=m xdot. The Hamiltonian is H=p xdot−L=p²/m−[p²/(2m)−½kx²]=p²/(2m)+½kx². Hamilton's equations are xdot=p/m and pdot=−kx. Differentiating the first and using the second gives xddot=−(k/m)x. Constant-H curves are ellipses in the (x,p) phase plane. The direction of traversal is fixed by the equations and the oscillator period is 2π√(m/k).
The Hamiltonian is not automatically equal to energy for every system. For a time-dependent potential V(q,t), H may depend explicitly on time and dH/dt=∂H/∂t along a solution. For electromagnetic coupling, canonical momentum differs from mechanical momentum. A charged particle has L=½m v²+q A·v−qφ, so p=mv+qA. The Hamiltonian becomes H=(p−qA)²/(2m)+qφ. This distinction is central in gauge theory and quantum mechanics.
Hamiltonian methods are useful for stability analysis, perturbation theory, canonical transformations, statistical mechanics and quantisation. The equations preserve the symplectic structure of phase space. In numerical work, symplectic integrators often reproduce long-term orbital behaviour better than generic integrators, even when their pointwise error is not always the smallest. They are particularly valuable for planetary dynamics and molecular simulations where conservation over many periods matters.
9.6 Poisson brackets and canonical transformations
For phase-space functions A(q,p,t) and B(q,p,t), the Poisson bracket is {A,B}=Σ_i[(∂A/∂q_i)(∂B/∂p_i)−(∂A/∂p_i)(∂B/∂q_i)]. The time evolution of A is dA/dt={A,H}+∂A/∂t. A quantity with no explicit time dependence is conserved if its Poisson bracket with H vanishes. The fundamental brackets are {q_i,q_j}=0, {p_i,p_j}=0 and {q_i,p_j}=δ_ij. The bracket is bilinear, antisymmetric, obeys the product rule and satisfies the Jacobi identity.
Poisson brackets provide a compact algebra for angular momentum. For L=r×p, the components obey {L_i,L_j}=ε_ijk L_k, where ε_ijk is the Levi-Civita symbol. This algebra expresses the structure of rotations and becomes the commutator algebra of angular momentum in quantum mechanics after quantisation. A canonical transformation changes phase-space variables while preserving the form of Hamilton's equations. Generating functions provide systematic constructions; for a type-2 generating function F2(q,P,t), p_i=∂F2/∂q_i and Q_i=∂F2/∂P_i, while the transformed Hamiltonian is K=H+∂F2/∂t.
Worked example: for a one-dimensional harmonic oscillator H=p²/(2m)+½mω²q², evaluate {q,H}=∂H/∂p=p/m, which equals qdot. Similarly {p,H}=−∂H/∂q=−mω²q, which equals pdot. The bracket therefore reproduces Hamilton's equations directly. For A=H, {H,H}=0, and if H has no explicit time dependence, dH/dt=0.
Canonical transformations can simplify a problem by selecting action-angle variables. For a periodic one-dimensional system, the action variable J=(1/2π)∮p dq is constant for an integrable time-independent system. The angle variable advances uniformly in time, with angular frequency Ω=∂H/∂J. For the harmonic oscillator, J=E/ω and H=ωJ, so Ω=ω. This representation is useful in celestial mechanics, accelerator physics and adiabatic invariance.
A canonical transformation is not merely any invertible coordinate change. It must preserve the symplectic form, equivalently the fundamental Poisson brackets. In numerical implementations, test the transformation's Jacobian against the symplectic condition. A transformation that preserves energy but not the symplectic structure may still distort phase-space geometry over long times. Always distinguish canonical momentum from mechanical momentum and specify the generating-function convention used.
9.7 Central-force motion and effective potential
A central force has the form F(r)=F(r) rhat and derives from a spherically symmetric potential V(r). The torque is r×F=0, so angular momentum ℓ is conserved. The motion is planar, and in polar coordinates the Lagrangian is L=½m(rdot²+r²φdot²)−V(r). Since p_φ=mr²φdot=ℓ, the radial motion is equivalent to one-dimensional motion in the effective potential V_eff(r)=V(r)+ℓ²/(2mr²). The radial energy equation is E=½m rdot²+V_eff(r). The second term is the centrifugal barrier; it is a consequence of angular momentum conservation, not an additional physical force in the original inertial frame.
Circular orbits occur at stationary points of V_eff: dV_eff/dr=V'(r)−ℓ²/(mr³)=0. Stability requires a local minimum, d²V_eff/dr²>0. Small radial perturbations about a stable circular orbit oscillate with frequency κ²=(1/m)V_eff''(r0). The orbital angular frequency is Ω=ℓ/(m r0²). Comparing κ and Ω determines whether nearby orbits close or precess. For a Newtonian gravitational potential V=−GMm/r, circular orbits satisfy ℓ²=GMm²r0 and Ω²=GM/r0³.
Worked example: a satellite of mass m in a circular orbit of radius r around a planet of mass M has gravitational force GMm/r². Equating centripetal acceleration gives v²=GM/r, so the period is T=2πr/v=2π√(r³/GM). The total energy is E=½mv²−GMm/r=−GMm/(2r). The negative sign indicates a bound orbit. Escape requires total energy at least zero in the ideal two-body model, giving escape speed v_esc=√(2GM/r), which is √2 times the circular speed at the same radius.
For an inverse-square attractive force, the orbit is a conic section r(φ)=p/[1+e cos(φ−φ0)], where p=ℓ²/(GMm²) and e is the eccentricity. Eccentricity determines the orbit: e=0 circle, 0<e<1 ellipse, e=1 parabola and e>1 hyperbola. Kepler's equal-area law follows from constant areal velocity dA/dt=ℓ/(2m). The period of an ellipse depends on its semimajor axis a through T²=4π²a³/(GM), independent of eccentricity for a fixed a.
For a general central potential, the orbit need not close. The angular change between successive radial turning points is Δφ=2∫_{rmin}^{rmax}[ℓ/(r²√(2m(E−V_eff)))]dr. If Δφ is not 2π times a rational ratio, the orbit precesses and fills an annular region. This integral is often evaluated numerically. It provides a practical route to compare theoretical orbit shapes with observations and to infer departures from an inverse-square law.
9.8 Kepler problem, Runge–Lenz vector and orbital perturbations
The two-body gravitational problem can be reduced to a single effective particle of reduced mass μ=m1m2/(m1+m2) moving in the potential −Gm1m2/r, with the centre of mass moving uniformly. The conserved angular momentum is L=μr×rdot. In addition to energy and angular momentum, the inverse-square problem has a conserved Laplace–Runge–Lenz vector A= p×L−μk rhat, where k=Gm1m2 and p=μrdot. Its direction points toward periapsis and its magnitude is related to eccentricity. This additional constant explains why the Kepler ellipse has a fixed orientation in an ideal Newtonian two-body problem.
The vector is conserved only for the exact inverse-square central force. Small corrections to the potential cause periapsis precession. For a weak perturbing potential δV(r), one can calculate the orbit correction using perturbation theory, action-angle variables or numerical integration. The relativistic correction to planetary motion is a famous example: the orbit is nearly Newtonian but the perihelion advances slightly each revolution. In practical astronomy, other planets, oblateness and tidal effects can produce much larger perturbations than the relativistic correction for some bodies.
A useful way to derive the conic equation is to set u=1/r. Binet's equation for a central force F(r) is d²u/dφ²+u=−μF(1/u)/(L²u²). For F=−k/r²=−ku², the equation becomes u''+u=μk/L². Its solution is u=(μk/L²)[1+e cos(φ−φ0)], giving the conic form. The derivation assumes nonzero angular momentum and motion in a plane.
Worked example: for a planet orbiting a star with gravitational parameter GM, an ellipse with semimajor axis a has specific orbital energy ε=−GM/(2a). If a spacecraft transfers from a circular orbit r1 to a circular orbit r2 using a Hohmann transfer, the transfer ellipse has semimajor axis at=(r1+r2)/2. Vis-viva gives v²=GM(2/r−1/a). At r1, the transfer speed is √[GM(2/r1−1/at)], and at r2 it is √[GM(2/r2−1/at)]. The two impulsive velocity changes are the differences from the circular speeds √(GM/r1) and √(GM/r2). This illustrates how central-force theory informs mission design.
When applying the two-body model, state whether masses are treated as point masses, whether the central body is spherical, and whether atmospheric drag or third-body effects are neglected. For close binaries, satellites, and spacecraft, these assumptions can be tested against the required accuracy. A model should be made more complex only when the residual error or physical scale justifies it.
9.9 Rigid-body kinematics and angular velocity
A rigid body is an extended system in which distances between all pairs of material points remain constant. Its configuration requires the centre-of-mass position and orientation. The velocity of a point P relative to a reference point O fixed in the body is v_P=v_O+ω×r_P/O, where ω is the angular velocity vector. The acceleration is a_P=a_O+α×r+ω×(ω×r), with α=ωdot. The second term is tangential acceleration and the third is centripetal acceleration. These vector equations hold instantaneously and are independent of the choice of Euler-angle convention.
For a rigid body rotating about a fixed axis, T_rot=½Iω², where I=∫r_perp² dm is the moment of inertia about that axis. For general three-dimensional rotation, angular momentum is not necessarily parallel to angular velocity. In body-fixed principal axes, L=(I1ω1,I2ω2,I3ω3), while T=½(I1ω1²+I2ω2²+I3ω3²). The principal moments are eigenvalues of the inertia tensor. This distinction is important for asymmetric bodies and tumbling spacecraft.
Worked example: a uniform solid disk of mass M and radius R rotates about its symmetry axis. Its surface mass density is σ=M/(πR²). A ring of radius r and thickness dr has mass dm=2πrσdr. Therefore I=∫r²dm=2πσ∫_0^R r³dr=½MR². If it rotates at angular speed ω, its rotational kinetic energy is ¼MR²ω². If a constant torque τ acts, angular acceleration is α=τ/I=2τ/(MR²), assuming the axis is fixed and I is constant.
The centre-of-mass decomposition separates the kinetic energy of a system into translational and rotational parts: T=½M v_cm²+½Σ_a m_a |v'_a|². For a rigid body, the second term is ½ω·I·ω. This decomposition is valuable because external forces determine centre-of-mass acceleration while external torque about the centre of mass determines angular momentum change. Internal forces cancel in the total linear momentum balance, although internal stresses are essential to maintain rigidity.
Angular velocity is an axial vector, and finite rotations do not commute: rotating first about x and then y generally differs from the reverse order. Small rotations can be added approximately as vectors, but large orientation changes require rotation matrices, quaternions or other proper attitude representations. In computational rigid-body dynamics, numerical drift can violate orthogonality of a rotation matrix; re-orthogonalisation or quaternion normalisation may be required.
9.10 Inertia tensor, principal axes and parallel-axis theorem
The inertia tensor about an origin O is I_ij=∫(r²δ_ij−x_i x_j)dm. In matrix form it is symmetric. The rotational kinetic energy is T=½ωᵀIω and angular momentum is L=Iω. Off-diagonal products of inertia represent coupling between angular-velocity components. Since I is real and symmetric, it has orthogonal eigenvectors and real eigenvalues; these eigenvectors define principal axes. In the principal-axis frame, the tensor is diagonal and the rotational energy is a sum of three independent quadratic terms.
For a thin uniform rectangular plate of sides a and b in the xy-plane, the moments about its centre are I_x=Mb²/12, I_y=Ma²/12 and I_z=M(a²+b²)/12. The perpendicular-axis theorem I_z=I_x+I_y applies to a planar lamina. It does not apply to a general three-dimensional body. The parallel-axis theorem states I_O=I_C+Md² for a specified axis parallel to an axis through the centre of mass, where d is the perpendicular separation. For the full tensor, the shift is I_O=I_C+M[(d·d)1−ddᵀ].
Worked example: a uniform rod of length L and mass M has moment of inertia about its centre perpendicular to the rod I_C=ML²/12. About an axis through one end, parallel to the first, I_end=ML²/12+M(L/2)²=ML²/3. The result can also be obtained directly by integrating x² dm with dm=(M/L)dx from 0 to L. Agreement between independent derivations is a useful check.
A physical pendulum uses the inertia about its pivot. If its centre of mass is a distance d from the pivot, the small-angle period is T=2π√(I_p/(Mgd)). The equivalent simple-pendulum length is l_eq=I_p/(Md). This relation is used in experiments to determine g or moments of inertia. Uncertainty in pivot location and the centre-of-mass position can dominate the result, so the geometry should be measured carefully.
The intermediate-axis theorem explains the instability of rotation about the intermediate principal moment for a torque-free asymmetric rigid body. Rotation about the maximum or minimum principal axis is stable to small perturbations, while rotation about the intermediate axis is unstable. This can be derived by linearising Euler's equations around steady rotation. The effect is observable in a tossed rectangular object and matters for spacecraft attitude dynamics. It is not a statement that every real body will tumble visibly: damping, internal flexibility and external torques alter the motion.
9.11 Euler equations and torque-free rigid-body motion
In a body-fixed principal-axis frame, Euler's equations are I1ωdot1+(I3−I2)ω2ω3=τ1, I2ωdot2+(I1−I3)ω3ω1=τ2, and I3ωdot3+(I2−I1)ω1ω2=τ3. These follow from the inertial-frame angular momentum balance τ=dL/dt and the rotating-frame derivative identity (dL/dt)_inertial=(dL/dt)_body+ω×L. The cross product term is essential: it accounts for the fact that the body-fixed basis rotates.
For torque-free motion, τ=0. Both rotational kinetic energy T=½ΣIiωi² and angular momentum magnitude L²=ΣIi²ωi² are conserved. The intersection of these two surfaces constrains the angular-velocity trajectory. Linearising around rotation about the first principal axis gives coupled perturbation equations. The sign of the resulting squared frequency determines stability. Rotation about the intermediate principal axis produces exponential growth of a perturbation.
Worked example: a symmetric top has I1=I2=I and I3 distinct. If no external torque acts, the component of angular momentum along the symmetry axis is conserved, and the body exhibits regular precession and nutation depending on initial conditions. Under gravity, a fixed point and centre-of-mass offset produce torque τ=r_cm×Mg. For rapid spin and small nutation, the approximate precession rate is Ω_p≈Mgr/(I3ω3). This approximation assumes the spin angular momentum dominates the transverse components and the top is not near a separatrix.
A gyroscope resists changes in its angular momentum direction, but it does not resist torque in the sense of cancelling it. The torque changes the angular momentum vector according to τ=dL/dt. The resulting motion can be perpendicular to the applied torque and may appear counterintuitive if one thinks only in terms of scalar rotational dynamics. Always use vector equations for gyroscopic motion.
In a laboratory, a rotating platform with known applied torque can be used to estimate moment of inertia. Measure angular acceleration from a fitted angular-velocity-versus-time graph, then calculate I=τ/α. Repeat at several torques to test linearity. Bearing friction introduces an offset torque, so fit τ=Iα+τ_f rather than forcing the line through the origin. Report uncertainties in torque, pulley radius, hanging mass and angular acceleration. Avoid sudden release that could cause the apparatus to topple or entangle clothing.
9.12 Rolling motion and contact constraints
For a rigid wheel of radius R rolling without slipping on a stationary horizontal surface, the velocity of the instantaneous contact point is zero relative to the surface. Therefore v_cm=Rω and a_cm=Rα along the direction of travel, with signs determined by the coordinate convention. The total kinetic energy is T=½Mv_cm²+½I_cmω². Substituting the no-slip relation gives T=½(M+I_cm/R²)v_cm². The effective inertial term includes both translation and rotation.
For a body rolling down an incline of angle θ, the gravitational component Mg sinθ drives motion and static friction provides the torque required for rotation. Translation gives Ma=Mg sinθ−f, while rotation about the centre gives Iα=fR. Using a=αR yields a= g sinθ/[1+I/(MR²)]. The friction force points up the slope for a body accelerating downhill in the usual case. The result depends on shape through I/(MR²): a solid cylinder accelerates faster than a thin hoop of the same mass and radius.
Worked example: for a uniform solid sphere I=2MR²/5. Its acceleration down a frictionless-looking but no-slip incline is a=(5/7)g sinθ. A thin hoop has I=MR² and a=(1/2)g sinθ. Thus the sphere reaches the bottom first if both start together and roll without slipping. The static friction is not necessarily dissipative because the point of contact is instantaneously at rest relative to the surface.
The no-slip condition can fail if the required static friction exceeds its maximum value μ_s N. One must calculate the friction required by the rolling equations and compare it with μ_s Mg cosθ. If the condition fails, the body slips and kinetic friction applies; the relation v=Rω is no longer valid. A correct solution must test the assumed contact regime rather than impose rolling without checking feasibility.
For a moving surface such as a conveyor belt, the contact-point velocity equals the belt velocity, not zero. The constraint is v_cm−Rω=v_belt with sign convention chosen consistently. For a wheel on a rotating drum or inside a curved track, the geometry changes the relation. This highlights that rolling is a kinematic contact condition, not a universal equation independent of the surface.
A useful experiment compares the acceleration of objects with different mass distributions on an adjustable incline. Measure the angle, travel distance and time, repeat trials, and use a photogate if available. Compare observed acceleration with the theoretical prediction. Sources of discrepancy include slipping, rolling resistance, deformation, air drag, non-uniform objects and inaccurate release. Video tracking can provide an independent estimate of centre-of-mass acceleration and angular speed.
9.13 Coupled oscillations and normal modes
A system of N small oscillations near a stable equilibrium can be approximated by a quadratic Lagrangian L=½ qdotᵀM qdot−½qᵀKq, where M is the mass matrix and K is the stiffness matrix. The equations are M qddot+Kq=0. Seeking normal-mode solutions q=a exp(iωt) gives (K−ω²M)a=0. A non-zero vector a exists when det(K−ω²M)=0. The resulting eigenvalues ω² are squared normal-mode frequencies and the eigenvectors give relative amplitudes and phases of the coordinates.
For a symmetric positive mass and stiffness matrix, the normal modes can be chosen orthogonal under the mass-weighted inner product a_rᵀM a_s. Transforming to modal coordinates decouples the linear equations into independent oscillators. This is the mathematical reason that coupled motion can be understood as a superposition of normal modes. The number of modes equals the number of independent degrees of freedom, although zero-frequency modes may occur for free translation or rotation.
Worked example: two equal masses m connected by three identical springs k between fixed walls, with displacements x1 and x2. The equations are m x1ddot+2kx1−kx2=0 and m x2ddot+2kx2−kx1=0. The in-phase mode x1=x2 has ω1=√(k/m), because the middle spring is unstretched. The out-of-phase mode x1=−x2 has ω2=√(3k/m), because the middle spring also contributes a restoring force. An arbitrary initial displacement is a combination of these modes and can show beats.
For unequal masses or spring constants, solve the generalised eigenvalue problem numerically or symbolically. Check that all ω² are non-negative for a stable equilibrium. A negative eigenvalue indicates an exponentially growing perturbation and therefore instability within the linearised model. A zero eigenvalue may represent a symmetry mode, such as uniform translation of an isolated chain.
Normal modes are measured in structures, molecules, acoustic cavities and electrical networks. In molecular spectroscopy, vibrational modes determine characteristic absorption frequencies. In civil engineering, modal frequencies and mode shapes help identify resonance risks in bridges and buildings. In a laboratory, accelerometers or optical displacement sensors record the response to a small impulse; Fourier analysis reveals peaks corresponding to natural frequencies. Damping broadens resonances and shifts peak locations slightly, so a simple undamped model should be treated as a first approximation.
Mode shapes should be normalised consistently before comparing amplitudes. Eigenvectors are defined only up to an arbitrary scale and sign. When two eigenfrequencies are close, measurement noise or small parameter changes can mix the apparent mode shapes. Report the normalisation convention, boundary conditions and whether the mode is calculated or measured.
9.14 Non-inertial frames, Coriolis and centrifugal terms
A reference frame rotating with angular velocity Ω relative to an inertial frame has the derivative relation (dA/dt)_inertial=(dA/dt)_rot+Ω×A for any vector A. Applying this twice to position yields the acceleration relation a_inertial=a_origin+a_rot+2Ω×v_rot+Ωdot×r+Ω×(Ω×r), where a_origin is the acceleration of the rotating-frame origin. Newton's second law in the rotating frame therefore includes the Coriolis force −2mΩ×v_rot, Euler force −mΩdot×r and centrifugal force −mΩ×(Ω×r), in addition to real forces and any translational inertial force −m a_origin.
These are fictitious or inertial forces in the sense that they arise from the accelerating coordinate frame, not from an interaction with another body. They are nevertheless essential for predicting motion as measured in that frame. The centrifugal term points away from the rotation axis for constant Ω. The Coriolis term is perpendicular to the rotating-frame velocity and Ω, so it changes direction but does no instantaneous work in a frame with constant Ω. The Euler term appears when the angular velocity changes.
Worked example: a particle is released from rest relative to a platform rotating at constant Ω. In the rotating frame, its initial velocity is zero, but the centrifugal acceleration Ω²r causes it to move outward. As it acquires radial velocity, the Coriolis acceleration deflects its path. The inertial-frame description is simply straight-line motion in the absence of real forces; the curved path in the platform frame is a coordinate effect. Both descriptions predict the same physical event when transformed consistently.
On Earth, the Coriolis acceleration is approximately −2Ω_E×v in a local frame. It contributes to large-scale atmospheric and oceanic circulation and to the deflection of long-range projectiles. Its magnitude depends on latitude and velocity; it is not the primary cause of water draining direction in ordinary sinks, where basin geometry and initial flow dominate. The Foucault pendulum demonstrates Earth's rotation through the gradual rotation of the oscillation plane relative to the ground.
For a rotating fluid, centrifugal effects can produce a pressure gradient. In steady rotation with fluid at rest relative to the frame, the pressure gradient balances gravity and centrifugal acceleration. This leads to a parabolic free surface in a rotating container. In engineering, rotating machinery, centrifuges, gyroscopes and vehicle dynamics all require careful treatment of non-inertial frames.
When solving problems, first identify the frame and whether it is inertial. Draw the real forces separately from inertial forces. State Ω and its direction, use the right-hand rule for cross products, and verify limiting cases: if Ω→0, all rotational inertial terms vanish; if v_rot=0, the Coriolis term vanishes; if Ωdot=0, the Euler term vanishes. Confusing centrifugal force with centripetal acceleration is a frequent source of sign errors.
9.15 Small oscillations of a physical pendulum and stability
A physical pendulum is a rigid body oscillating under gravity about a fixed horizontal pivot. If the centre of mass is distance d from the pivot and the moment of inertia about the pivot is I_p, the exact equation is I_p θddot+Mgd sinθ=0. For small θ, sinθ≈θ, giving simple harmonic motion with ω0=√(Mgd/I_p) and period T=2π√(I_p/(Mgd)). The result includes the entire mass distribution through I_p, not merely the distance to the centre of mass.
The equilibrium at the lowest centre-of-mass position is stable because potential energy has a local minimum. The inverted position is unstable because it is a local maximum. More generally, for a conservative system with equilibrium q0, expand V(q)≈V(q0)+½δqᵀKδq. If the Hessian K is positive definite and the mass matrix is positive definite, small perturbations oscillate. If K has a negative eigenvalue, at least one perturbation grows exponentially. If an eigenvalue is zero, higher-order terms or symmetries must be examined.
Worked example: a uniform rod of length L pivoted at one end has I_p=ML²/3 and d=L/2. Its small-angle period is T=2π√[(ML²/3)/(MgL/2)]=2π√[2L/(3g)]. A simple pendulum with equivalent length l_eq=2L/3 has the same small-angle period. The period is independent of mass, but depends on length and gravitational acceleration.
For larger amplitudes, the period increases because the restoring torque is no longer proportional to angle. The exact period can be expressed using a complete elliptic integral: T=4√(I_p/(Mgd)) K(sin(θ_max/2)). For small maximum amplitude θ_max, expansion gives T≈T0[1+θ_max²/16+11θ_max⁴/3072+…], with θ_max in radians. This amplitude dependence matters when high precision is required.
A stability calculation should specify the equilibrium, allowed perturbations and constraints. A system may be stable in one direction and unstable in another. Gyroscopic effects can stabilise certain configurations dynamically even when the static potential curvature is not positive in every direction. Such gyroscopic stability requires the full equations, including velocity-dependent terms, rather than a potential-energy test alone.
9.16 Hamilton–Jacobi theory and action-angle variables
Hamilton–Jacobi theory seeks a generating function S(q,α,t) satisfying ∂S/∂t+H(q,∂S/∂q,t)=0. This nonlinear first-order partial differential equation can transform the dynamics into a problem of finding constants α and a complete integral S. The canonical momenta are p_i=∂S/∂q_i. For a time-independent Hamiltonian, one often separates S=W(q,α)−Et, producing the time-independent Hamilton–Jacobi equation H(q,∂W/∂q)=E.
For a one-dimensional conservative system, W(q)=∫p(q,E)dq, with p=±√[2m(E−V(q))]. The action integral J=(1/2π)∮p dq measures the enclosed phase-space area divided by 2π. For periodic motion, the conjugate angle increases uniformly, and the frequency is Ω=∂H/∂J. This converts nonlinear periodic motion into uniform motion in angle space when the system is integrable.
Worked example: for the harmonic oscillator, p=√[2m(E−½mω²q²)] on the positive branch. The closed phase-space orbit is an ellipse with area 2πE/ω. Thus J=E/ω and H=ωJ. Therefore Ω=∂H/∂J=ω, recovering the oscillator frequency. The result also shows why action is a natural measure of oscillation amplitude and energy.
In slowly varying systems, action variables are often adiabatic invariants: they remain approximately constant when parameters change on a timescale much longer than the oscillation period and the system stays away from separatrices. This principle appears in planetary dynamics, plasma physics and accelerator physics. It should not be applied blindly near resonances or separatrix crossings, where the adiabatic approximation can fail.
Hamilton–Jacobi theory also provides a bridge to geometrical optics. The eikonal equation for wave propagation has a mathematical structure analogous to a Hamilton–Jacobi equation, with rays playing a role similar to trajectories. This connection is useful conceptually, but the physical variables and approximations must be defined for each application.
9.17 Scattering in central potentials and differential cross-section
In a central potential, an incoming particle with speed v∞ and impact parameter b has angular momentum L=μbv∞ and energy E=½μv∞², taking the potential to vanish at infinity. The effective potential determines the closest approach and whether the trajectory is deflected or captured. The scattering angle χ is obtained by integrating the angular change from the turning point to infinity and comparing with the undeflected straight path. For an inverse-square repulsive or attractive Coulomb potential, this yields the Rutherford scattering relation.
The differential cross-section dσ/dΩ measures the effective area of incoming trajectories scattered into a unit solid angle around a direction. For an azimuthally symmetric interaction, dσ/dΩ=(b/sinχ)|db/dχ|. This formula follows by equating the annular incident area 2πb db to the scattered solid angle annulus 2π sinχ dχ. It assumes a one-to-one relation between impact parameter and scattering angle; if multiple trajectories contribute, their cross-sections must be summed.
For Coulomb scattering with potential V(r)=κ/r, the Rutherford differential cross-section is proportional to [κ/(4E)]² cosec⁴(χ/2), with the precise coefficient depending on whether E and κ are defined for the relative two-body motion. The strong increase at small angles reflects the long range of the Coulomb force. Experimental scattering distributions can therefore reveal the spatial and charge structure of a target, provided multiple scattering and detector acceptance are controlled.
Worked example: if a central interaction produces a known monotonic relation b(χ), calculate db/dχ analytically and substitute into dσ/dΩ. Check dimensions: b/|dχ/db| has dimensions of area per steradian, with steradian dimensionless in SI. Integrating the differential cross-section over a detector's angular acceptance gives the expected effective area for that angular range. Multiplying by incident flux and target number gives a predicted event rate.
The classical model is valid when de Broglie wavelengths are small compared with the characteristic interaction scale and quantum effects are negligible. At low energies or small length scales, quantum scattering replaces classical trajectories with probability amplitudes. Classical scattering nevertheless remains important in beam transport, molecular collisions and charged-particle experiments.
9.18 Numerical integration of mechanical systems
Many nonlinear mechanical equations have no closed-form solution. Numerical integration approximates the trajectory at discrete times. For a first-order system ydot=f(t,y), explicit Euler uses y_(n+1)=y_n+h f(t_n,y_n), while the midpoint and Runge–Kutta methods use additional slope evaluations to improve accuracy. For second-order mechanics, one may integrate position and velocity together or use specialised methods such as velocity Verlet. The time step h must resolve the shortest relevant period and any rapid transient.
A numerical result should be tested through step refinement, comparison with an analytic limiting case, dimensional consistency and conservation-law monitoring. Local truncation error is the error introduced in one step; global error accumulates over many steps. Stability is distinct from accuracy: a method may produce a numerically growing solution even when the true system is stable if the step is too large. Stiff systems can require implicit methods or adaptive step control.
For Hamiltonian systems, symplectic integrators preserve the geometric structure of phase space. Velocity Verlet is widely used for separable Hamiltonians H=T(p)+V(q): p_(n+1/2)=p_n−(h/2)∇V(q_n), q_(n+1)=q_n+h M^−1 p_(n+1/2), p_(n+1)=p_(n+1/2)−(h/2)∇V(q_(n+1)). The method is time reversible and second-order accurate. It does not conserve the exact energy at every step, but often keeps energy error bounded and oscillatory over long integrations.
Worked numerical protocol: integrate a Kepler orbit from specified initial position and velocity. First calculate the expected specific energy ε and angular momentum h_spec. Choose a time step much smaller than the orbital period, integrate for several periods, then plot the orbit and monitor relative energy and angular-momentum errors. Repeat with h/2 and h/4. A converged trajectory should change less under refinement. If the orbit spirals systematically, inspect the integrator, step size, force evaluation and coordinate singularities.
Report the algorithm, step size or tolerance, initial conditions, units, precision, and convergence evidence. A smooth graph is not proof of accuracy. Include a table of maximum relative energy error and final phase error for successive step sizes. For chaotic systems, trajectories may diverge exponentially even when each computation is correct; compare statistical or qualitative properties rather than expecting long-term pointwise agreement.
9.19 Laboratory investigations and data analysis
A university laboratory can test rotational dynamics by measuring the acceleration of a falling mass attached to a string wound around a pulley. If the hanging mass m descends with acceleration a and the pulley has radius R and moment of inertia I, the equations are mg−T=ma and TR=Iα, with α=a/R. Eliminating T gives a=mg/[m+I/R²]. Plotting 1/a against 1/m or fitting the linearised relationship can estimate I, provided friction torque is included or shown to be negligible. Use a photogate or motion sensor and repeat each mass several times.
To measure a moment of inertia independently, use a torsion pendulum. A body suspended by a wire with torsion constant C obeys Iθddot+Cθ=0 for small twist, so T=2π√(I/C). Calibrate C using a reference body of known I, then measure the unknown body. The wire must remain in its elastic regime; large angular deflection can make the restoring torque nonlinear. Damping should be small or included in the analysis.
For a physical pendulum, measure period for several pivot positions. Since T²=4π²I_p/(Mgd), and I_p=I_cm+Md², plotting T² against d can be used to estimate g and I_cm, subject to the chosen linearisation. Avoid pivot positions too close to the centre of mass where the period becomes large and uncertainty rises. Use multiple oscillations per timing run and divide the total time by the number of cycles.
Uncertainty analysis should distinguish random variation from systematic effects. For independent measured quantities x_i, the standard uncertainty in a derived quantity y is u_y²=Σ_i(∂y/∂x_i)²u_i². If variables are correlated, covariance terms must be included. For a fitted straight line, report slope, intercept, their uncertainties, residual plot and goodness-of-fit assessment. Significant curvature suggests a missing physical effect or an inappropriate linearisation.
Laboratory safety includes securing rotating apparatus, keeping loose clothing and hair away from shafts, using guards where available, checking hanging masses and pulleys before release, and never standing under suspended masses. Stop the apparatus before adjusting sensors or changing the string. Record apparatus identification, calibration status, environmental conditions and any deviations from the planned method. A strong report separates observation from interpretation and explains whether data support the model within uncertainty.
9.20 Advanced worked examples and integrated problem solving
Example 1 — Atwood machine with a massive pulley. Two masses m1 and m2 are connected by a light string over a pulley of radius R and moment I. Assume m2>m1 and no slip. Equations are m2g−T2=m2a, T1−m1g=m1a, and (T2−T1)R=I(a/R). Adding the translational equations and substituting the torque relation gives a=(m2−m1)g/[m1+m2+I/R²]. The tensions are T1=m1(g+a) and T2=m2(g−a). The pulley increases the effective inertia and reduces acceleration relative to a massless pulley. Check limits: I→0 recovers the ideal Atwood result; m1=m2 gives a=0.
Example 2 — Bead on a rotating hoop. A bead of mass m slides without friction on a vertical circular hoop of radius a rotating at constant angular speed Ω about its vertical diameter. Let θ be measured from the downward vertical. The bead's position has speed squared a²θdot²+a²Ω²sin²θ. The Lagrangian is L=½ma²(θdot²+Ω²sin²θ)+mga cosθ, choosing gravitational potential V=−mga cosθ. The Euler–Lagrange equation is θddot=sinθ(Ω²cosθ−g/a). Equilibria satisfy sinθ=0 or cosθ=g/(aΩ²). Linearise around each equilibrium to determine stability. This example shows how rotation changes the effective potential and can create additional equilibrium positions when Ω²a exceeds g.
Example 3 — Central-force circular orbit stability. For V(r)=−k/r+α/r², V_eff=−k/r+[α+L²/(2m)]/r². A circular orbit requires dV_eff/dr=0, giving r0=2[α+L²/(2m)]/k when the bracket is positive. The second derivative is positive for positive k and positive bracket, so the orbit is radially stable in this model. The extra inverse-square potential changes the effective angular barrier and the orbital precession. Interpret α's units carefully: it must have units energy times length squared.
Example 4 — Coupled oscillator eigenfrequencies. Given M=diag(m,2m) and K=k[[2,−1],[−1,2]], solve det(K−ω²M)=0. The determinant is (2k−mω²)(2k−2mω²)−k²=0. With λ=mω²/k, this becomes (2−λ)(2−2λ)−1=0, or 2λ²−6λ+3=0. Thus λ=(3±√3)/2 and ω²=(k/m)(3±√3)/2. The mode vector follows by substituting each eigenvalue into either row. The two modes are not simply in-phase and out-of-phase because the masses differ.
Example 5 — Rotating frame balance. A bead at rest relative to a horizontal turntable at radius r experiences a real inward static friction force providing centripetal acceleration in the inertial frame. In the rotating frame, the bead is at rest and the outward centrifugal force balances the inward friction. If the required friction μ_s N is exceeded, the bead slips. This is not a contradiction: the two frames use different force balances but predict the same impending motion.
For every advanced problem, use a disciplined sequence: define the system; draw the geometry and choose coordinates; list assumptions and constraints; write kinetic and potential energies or Newton–Euler balances; identify conserved quantities; derive the governing equations; solve analytically as far as possible; evaluate limiting cases; substitute numerical values with SI units; and state the physical meaning of the result. A numerical answer without a model and checks is incomplete at university level.
9.21 Common misconceptions, modelling limits and revision synthesis
A Lagrangian is not simply another name for energy. It is a function whose stationary action yields equations of motion; for many natural systems it equals T−V. The Hamiltonian is a Legendre transform and equals total mechanical energy only under suitable conditions. Canonical momentum may differ from mechanical momentum, especially in electromagnetic fields. A cyclic coordinate is one absent explicitly from the Lagrangian, not merely one that happens to have constant velocity. Its conjugate momentum is conserved.
The centrifugal force is not a real interaction in an inertial frame. It is an inertial force introduced in a rotating frame. The Coriolis force is velocity dependent and perpendicular to the relative velocity for constant rotation; it does no work in that frame. Angular momentum and angular velocity are not generally parallel for an asymmetric rigid body. The scalar relation L=Iω is valid only about a fixed axis or along a principal axis in the appropriate context.
Rolling without slipping is a constraint that must be checked against available static friction. The contact point is instantaneously at rest only relative to the stationary surface under pure rolling. Static friction can do zero work at the contact point, whereas kinetic friction generally dissipates mechanical energy. A body with greater moment of inertia does not always roll slower unless mass, radius, geometry and contact conditions are specified.
A stable equilibrium is not necessarily one with zero force only; it must restore small perturbations. Linear stability analysis is local and may not describe large excursions. A stable circular orbit requires a minimum of the effective potential, not merely a solution of the force-balance equation. A numerical integrator that conserves energy approximately may still have phase error, and an attractive plot does not establish convergence.
Modelling assumptions should be visible: ideal rigid body, point masses, massless strings, frictionless pivots, conservative forces, small oscillations, weak perturbations, and isolated system are approximations. Explain when each is reasonable and what observation would reveal its failure. Advanced mechanics is not a collection of formulas; it is the systematic construction, testing and refinement of mathematical models of physical systems.
Revision synthesis: start with configuration and degrees of freedom; select generalised coordinates; form T and V; construct L; derive Euler–Lagrange equations; identify symmetries and conserved quantities; transform to Hamiltonian variables when useful; analyse central forces using effective potentials; describe rigid bodies with inertia tensors and Euler equations; reduce small coupled motion to normal modes; add inertial forces when using accelerating frames; and validate predictions through experiment or converged computation. The same reasoning connects classical mechanics to continuum mechanics, molecular dynamics, celestial mechanics, statistical physics and quantum theory.
Classical mechanics diagrams and physical interpretation
University examination questions with model answers
- 1. Define a holonomic constraint and give one example.Model answer: A holonomic constraint is an integrable relation among coordinates and possibly time, f(q,t)=0. A pendulum string of fixed length satisfies x²+y²=l².
- 2. Derive the Euler–Lagrange equation from the stationary-action principle.Model answer: Vary S=∫Ldt, integrate the velocity-variation term by parts, use zero endpoint variations, and set the coefficient of each arbitrary interior variation to zero: d/dt(∂L/∂qdot_i)−∂L/∂q_i=0.
- 3. What is a cyclic coordinate?Model answer: A coordinate absent explicitly from L. Its conjugate momentum ∂L/∂qdot is conserved because its time derivative equals ∂L/∂q=0.
- 4. Explain Noether's theorem using rotational symmetry.Model answer: If the action is invariant under continuous rotations, angular momentum is conserved. The conserved quantity is the generator of the rotation.
- 5. Distinguish Lagrangian and Hamiltonian formulations.Model answer: L(q,qdot,t)=T−V for a natural system and produces second-order equations. H(q,p,t)=Σp_iqdot_i−L is a Legendre transform and produces first-order canonical equations.
- 6. Show that a harmonic oscillator Hamiltonian is conserved.Model answer: H=p²/(2m)+½mω²q². With no explicit time dependence, dH/dt=∂H/∂t=0 along Hamilton's equations.
- 7. State the fundamental Poisson brackets.Model answer: {q_i,q_j}=0, {p_i,p_j}=0, {q_i,p_j}=δ_ij.
- 8. Derive the effective potential for central-force motion.Model answer: Use conservation of pφ=mr²φdot=ℓ in T=½m(rdot²+r²φdot²). The radial energy becomes ½m rdot²+V(r)+ℓ²/(2mr²)=E.
- 9. What condition gives a stable circular orbit?Model answer: dV_eff/dr=0 and d²V_eff/dr²>0 at the circular radius.
- 10. State Kepler's three laws in the Newtonian model.Model answer: Bound orbits are ellipses with the attracting mass at a focus; equal areas are swept in equal times; T² is proportional to a³.
- 11. Define the Runge–Lenz vector and explain its significance.Model answer: For the inverse-square Kepler problem it is a conserved vector pointing toward periapsis; it explains the fixed orientation of the ideal Kepler ellipse.
- 12. Explain why angular momentum may not be parallel to angular velocity.Model answer: For a general rigid body L=Iω. If I is not a scalar multiple of the identity, its action changes vector direction; alignment occurs along principal axes.
- 13. State and apply the parallel-axis theorem.Model answer: I_O=I_C+Md² for parallel axes separated by d. For a rod about its end, I=ML²/12+M(L/2)²=ML²/3.
- 14. Write Euler's equations for a rigid body in principal axes.Model answer: I1ωdot1+(I3−I2)ω2ω3=τ1, with cyclic permutations for components 2 and 3.
- 15. What is the intermediate-axis instability?Model answer: Torque-free rotation about the intermediate principal moment is unstable to small perturbations, unlike rotation about the largest or smallest principal moments.
- 16. Derive the acceleration of a rolling body down an incline.Model answer: Combine Ma=Mg sinθ−f, Iα=fR and a=αR to obtain a=g sinθ/[1+I/(MR²)].
- 17. When does pure rolling fail?Model answer: When the static friction required by the no-slip equations exceeds μ_sN; slipping then requires a different kinetic-friction model.
- 18. Identify the Coriolis and centrifugal accelerations.Model answer: Coriolis acceleration is 2Ω×v_rot in the inertial acceleration transformation; centrifugal acceleration in the rotating-frame equation is −Ω×(Ω×r).
- 19. How are normal-mode frequencies obtained?Model answer: Assume q=a exp(iωt) in Mqddot+Kq=0 and solve det(K−ω²M)=0.
- 20. Explain the physical meaning of a negative eigenvalue of M⁻¹K.Model answer: It gives ω²<0 and an exponentially growing linear perturbation, indicating instability near the equilibrium.
- 21. Describe the action-angle variables of a one-dimensional periodic system.Model answer: J=(1/2π)∮p dq is the action and its conjugate angle advances uniformly; Ω=∂H/∂J.
- 22. Derive the differential scattering cross-section relation.Model answer: Equate incident annular area 2πb db with scattered solid-angle annulus 2πsinχ dχ to obtain dσ/dΩ=(b/sinχ)|db/dχ|.
- 23. Why are symplectic integrators useful in long-term orbital simulation?Model answer: They preserve the symplectic geometry and often bound long-term energy error rather than producing systematic secular drift.
- 24. Give two sources of systematic error in a rotational-inertia experiment.Model answer: Bearing friction and inaccurate effective pulley radius; other examples include string slip, sensor calibration and non-negligible string mass.
- 25. A disk has M=2.0 kg and R=0.20 m. Find its axial moment of inertia.Model answer: I=½MR²=0.5×2.0×0.20²=0.040 kg m².
- 26. A spring has k=200 N m⁻¹ and a mass m=0.50 kg. Find the undamped angular frequency.Model answer: ω=√(k/m)=√400=20 rad s⁻¹.
- 27. A satellite is in a circular orbit of radius 7.0×10⁶ m around Earth. Write its speed expression.Model answer: v=√(GM_E/r). Substituting GM_E≈3.986×10¹⁴ m³s⁻² gives approximately 7.55×10³ m s⁻¹.
- 28. A uniform rod of length 1.2 m and mass 3 kg pivots at one end. Find I.Model answer: I=ML²/3=3×1.2²/3=1.44 kg m².
- 29. What is the physical meaning of a conjugate momentum?Model answer: It is p_i=∂L/∂qdot_i, the momentum variable paired with coordinate q_i in the variational and Hamiltonian descriptions; it need not be mv.
- 30. Explain why a stationary action need not be a minimum.Model answer: The second variation can be positive, negative or indefinite. The physical path is stationary, and may be a saddle point.
- 31. A hoop and solid cylinder roll from rest down the same incline. Which accelerates faster?Model answer: The solid cylinder: I/(MR²)=1/2, so a=(2/3)g sinθ; the hoop has a=(1/2)g sinθ.
- 32. What does an explicit time dependence in H imply?Model answer: Along a Hamiltonian trajectory dH/dt=∂H/∂t; the Hamiltonian is generally not conserved if that derivative is nonzero.
- 33. Why must a normal-mode eigenvector be normalised by convention?Model answer: Its scale and sign are arbitrary. A stated normalisation is needed for meaningful comparison of amplitudes.
- 34. Describe the physical pendulum small-angle approximation.Model answer: Replace sinθ by θ for |θ|≪1 rad, giving T=2π√(I_p/(Mgd)).
- 35. How can experimental data test the rolling model?Model answer: Measure acceleration for objects with known I/(MR²), compare with g sinθ/[1+I/(MR²)], and assess residuals and uncertainty.
- 36. What is the reduced mass in a two-body problem?Model answer: μ=m1m2/(m1+m2); it converts relative motion into an equivalent one-body problem in the central potential.
- 37. What is the difference between a real force and an inertial force?Model answer: A real force arises from physical interaction. An inertial force appears because equations are written in an accelerating or rotating frame.
- 38. State one limitation of the rigid-body approximation.Model answer: Real materials deform and vibrate; the approximation fails when elastic deformation or internal modes affect the motion.
- 39. Why is dimensional analysis important in a mechanics derivation?Model answer: It detects inconsistent equations, checks units of derived parameters and helps identify plausible dependence on physical scales.
- 40. Give a complete workflow for solving a constrained Lagrangian problem.Model answer: Choose independent coordinates, express positions and velocities, form T and V, construct L, apply Euler–Lagrange equations, impose constraints, solve, and check units and limiting cases.
Revision checklist, key equations and further study
Core equations: Euler–Lagrange: d/dt(∂L/∂qdot_i)−∂L/∂q_i=0; Hamilton: qdot_i=∂H/∂p_i and pdot_i=−∂H/∂q_i; Poisson evolution: dA/dt={A,H}+∂A/∂t; central effective potential: V_eff=V+ℓ²/(2mr²); rigid-body angular momentum: L=Iω; rolling acceleration: a=g sinθ/[1+I/(MR²)]; normal modes: det(K−ω²M)=0; rotating-frame acceleration includes 2Ω×v_rot and Ω×(Ω×r).
Before an examination, practise deriving rather than memorising the Euler–Lagrange equations, reducing a central-force problem, diagonalising an inertia tensor, deriving Euler's equations, solving a two-degree-of-freedom normal-mode problem and distinguishing inertial-frame from rotating-frame descriptions. For each derivation, identify assumptions and verify dimensions.
SEO topic coverage: Classical Mechanics II university notes; advanced classical mechanics; Lagrangian mechanics derivation; Euler–Lagrange equation examples; Hamiltonian mechanics; Hamilton's equations; Poisson brackets; Noether theorem; central force motion; effective potential; Kepler problem; rigid body dynamics; moment of inertia tensor; Euler equations; gyroscope precession; rolling motion; coupled oscillations; normal modes; Coriolis force; centrifugal force; Hamilton–Jacobi equation; action-angle variables; classical scattering; numerical mechanics; undergraduate physics worked examples; university physics examination questions with answers.
These terms are integrated as descriptive subject language rather than repeated unnaturally. The page is designed as a readable undergraduate learning resource with a clear canonical URL and structured LearningResource metadata.