Kinematics in Cartesian, Polar and Spherical Coordinates
High-school kinematics lives almost entirely in Cartesian coordinates, where the unit vectors are the same at every point and differentiating a position vector means differentiating three independent components. University mechanics cannot afford that luxury. Orbits, rotating machinery, central-force problems and anything with spherical or cylindrical symmetry are far more naturally described in coordinates whose basis vectors change from point to point — and the moment a basis vector depends on position, the product rule adds terms to every velocity and acceleration that have real physical names: the centripetal term, the Coriolis term, the Euler term.
This chapter derives those terms from scratch. The method is always the same: write the position vector in the chosen coordinates, then differentiate with respect to time, being careful to differentiate the basis vectors too. Once you have done this honestly for plane polar coordinates, the cylindrical and spherical results follow with more algebra but no new ideas.
The one habit that matters. In any non-Cartesian system, never write "" by analogy with the Cartesian case. Always start from (or its spherical analogue) and differentiate twice, tracking , and so on. The extra terms are the physics.
1. Position, Velocity and Acceleration as Vectors
For a particle tracing a path , velocity and acceleration are defined coordinate-free as
These definitions are independent of the coordinate system; only their components change from one system to another. The speed is and the velocity is always tangent to the path. What differs between coordinate systems is how much of the work is done by changing components and how much by changing basis vectors.
2. Cartesian Coordinates: The Easy Case
With and constant basis vectors (),
Every component evolves independently — the statement that motion along perpendicular Cartesian axes is decoupled. This is why projectile motion splits into two one-dimensional problems, and it is exactly the feature the other coordinate systems lack.
3. Plane Polar Coordinates: A Rotating Basis
In the plane, a point is located by with basis vectors (radially outward) and (in the direction of increasing , i.e. counter-clockwise from ). In terms of the fixed Cartesian basis,
Both depend on , and depends on time, so by the chain rule
These two little results — and — are the entire content of polar kinematics. Everything below is the product rule applied with them.
4. Velocity and Acceleration in Polar Form
Start from . Differentiating once, with the product rule:
The first term is the radial velocity (moving in or out); the second is the transverse velocity (swinging around). Differentiating again:
Substituting and and collecting components:
Each term has a name:
- — genuine radial acceleration from changing radial speed.
- — the centripetal term: present whenever the particle is going around, directed inward, equal to .
- — the angular or Euler term: from angular acceleration.
- — the Coriolis term: present only when the particle is simultaneously moving radially and rotating. It is the reason a figure skater spins up when pulling their arms in.
Worked example — uniform circular motion. If is constant and is constant, then and , leaving
exactly the familiar centripetal result — but now derived, not asserted, and as a special case of a formula that also handles spiralling and non-uniform rotation.
Worked example — a bead on a rotating rod. A rod pivoted at the origin rotates at constant while a frictionless bead slides outward with . The transverse component of is : even with zero angular acceleration, a transverse (Coriolis) acceleration is required, supplied by the normal force from the rod. If no transverse force acted, angular momentum would be conserved and would have to fall as the bead moved out; here the rod does push transversely, so grows and stays fixed.
5. Cylindrical Coordinates
Cylindrical coordinates are plane polars in the -plane plus an independent Cartesian . Since is constant, the results simply append a -term to the polar expressions:
These are the natural coordinates for helical motion, for a charged particle in a uniform magnetic field, and for anything with axial symmetry.
6. Spherical Coordinates: The Basis Vectors
Spherical coordinates use = distance from origin, = polar angle from the -axis (colatitude), = azimuthal angle in the -plane. The orthonormal basis is
All three depend on the angles, so all three have non-zero time derivatives whenever or changes. Working them out with the chain rule (the same procedure as in Section 3, just longer) gives
7. Velocity and Acceleration in Spherical Form
Starting from and differentiating once:
The three components are the radial, polar and azimuthal velocities; the factor in the last reflects that circles of constant shrink toward the poles. Differentiating a second time and collecting terms yields the standard result
You are rarely asked to reproduce all nine terms from memory, but you are expected to be able to derive them and to recognise the centripetal (, ) and Coriolis (, , ) contributions.
Worked example — the conical pendulum. A bob swings so that (constant, the string length) and (constant cone half-angle), with steady azimuthal rate . Then , , , and the acceleration reduces to
Its magnitude is and it points horizontally, toward the axis — the centripetal acceleration of uniform circular motion on the circle of radius . Setting this equal to the horizontal component of tension over mass recovers the standard result .
8. Common Pitfalls
- Treating , , as constant. They are the whole difficulty. Their time derivatives generate every centripetal, Coriolis and Euler term.
- Writing by dropping the "extra" terms. The correct radial component is and the transverse is .
- Confusing the two polar angles in spherical coordinates. In physics convention is measured from the -axis (colatitude, to ) and is azimuthal ( to ); many maths texts swap the symbols.
- Forgetting the in the azimuthal velocity. , not ; the discrepancy vanishes only in the equatorial plane.
- Mixing bases in one equation. Do not add a component to an component; convert everything to one basis first.
- Assuming the polar and the spherical are the same object. They are defined by different geometry; keep track of which coordinate system you are in.
Summary
- Velocity and acceleration are defined coordinate-free as and ; only the components and basis vectors change between systems.
- Cartesian: constant basis, so and the axes decouple.
- Plane polar: , give and .
- The named terms: centripetal, Coriolis, Euler/angular.
- Cylindrical: the polar results with an independent appended.
- Spherical: ; the acceleration has centripetal and Coriolis terms in every component, derivable from the three basis-vector derivatives.
- Method for any system: write in that system and differentiate twice, always differentiating the basis vectors too.