Loop Torque Lab · Guide · glidepathscience.org
GuideLoop Torque Lab
Contents
This Guide covers the ideas and equations behind every Walkthrough step, Concept Challenge question and Predict round. Read Part 1 before the Intro track and Part 2 before the Advanced track, about ten minutes each. Axes: $x$ to the right, $y$ away from you, $z$ up.
Part 1 · The essentials
From moving charges to F = IL × B
A magnetic field pushes sideways on a moving charge, $\vec F = q\vec v \times \vec B$. A current is moving charge: in copper, $n = 8.49 \times 10^{28}$ free electrons per cubic meter drift slowly against the current, at $v_d = I/(neA)$. A length $L$ of wire holds $nAL$ of them, and their pushes add to
$\vec F = I\vec L \times \vec B$ $F = ILB\sin\theta$
with $\vec L$ along the current and θ the angle between the wire and $\vec B$. The carriers’ sign cancels: electrons have negative $q$ and move backward, so $q\vec v$ points along $I$ either way. Only the part of the wire across $\vec B$ counts; a wire along $\vec B$ feels nothing.
Close the loop: no force, but a torque
Bend the wire into a closed loop in a uniform field. Every piece has a partner carrying current the other way through the same field, so the forces add to zero: $\vec F_{net} = 0$, for any shape and any tilt. But the forces on opposite sides do not act along one line. Like two hands on a steering wheel, they make a torque.
Which way? The loop turns so that its moment $\vec\mu$ (below) swings toward $\vec B$. “Rises” and “drops” mean $+z$ and $-z$, whichever way you turn the view.
τ = NIAB sin θ, μ and stability
A coil of $N$ turns enclosing area $A$ has a magnetic moment
$\mu = NIA$ $\tau = NIAB\sin\theta = \mu B\sin\theta$
$\vec\mu$ points along the coil’s axis, by the right-hand rule from the current (fingers along $I$, thumb along $\vec\mu$), and θ is the angle between $\vec\mu$ and $\vec B$. The torque is largest with the coil’s plane along $\vec B$ (θ = 90°) and zero with $\vec\mu$ along $\vec B$. Only the area counts, not the shape: a circle and a square of equal area feel equal torques.
Released, the loop swings, overshoots and settles with $\vec\mu$ along $\vec B$: a stable balance, like a compass needle. Anti-aligned is a balance too, but an unstable one: the slightest tilt grows and it flips.
A loop partly in a field
Now let the field fill only a region with an edge, and slide a loop with $\vec\mu$ along $\vec B$ halfway in. The two sides cut by the edge carry opposite currents through the same field, so their forces still cancel; the side inside has no partner. The loop is pulled in with
$F = NILB$
whether a quarter or three-quarters of it is inside: what matters is how fast the flux through it changes as it moves, $BL$ per meter. Fully inside, nothing pulls. Reverse the current and the loop is pushed out.
Magnets and long wires
Above a magnet’s north pole the field lines fan out, so the field on each piece of a ring tilts outward. With $\vec\mu$ along the field, each piece’s force points outward and tips toward the magnet: the outward parts cancel round the ring and the rest add. An aligned loop is pulled toward stronger field; a reversed one is pushed away.
Beside a long straight wire, $B = \mu_0 I_w/(2\pi r)$ with $\mu_0 = 4\pi \times 10^{-7}$ T·m/A. For a rectangular coil in the wire’s plane, the two long sides are in different fields, so their forces do not cancel; the short sides do. Currents running the same way attract, so the near side wins.
Motors: the commutator and back EMF
Released with little friction, a loop only swings about $\vec\mu$ along $\vec B$: the torque reverses each time $\vec\mu$ passes $\vec B$. A split-ring commutator reverses the coil’s current at exactly that moment, so the torque keeps its sign and the rotor spins.
Turning, the coil’s sides cut the field and generate a back EMF $\varepsilon$ that opposes the battery, so
$I = (V - \varepsilon)/R$
The faster it turns, the smaller the current. A jammed motor has no back EMF, so $I = V/R$, and $I^2R$ can burn out its coil.
Meters: the radial field
A moving-coil meter turns against a hairspring, $\tau_{spring} = \kappa\varphi$. Its curved poles and soft-iron core make the field radial, so the coil’s sides always cross $\vec B$ at right angles and $\tau = NIAB$ at every angle. The needle settles where the torques balance:
$\varphi = NIAB/\kappa$
proportional to $I$: equal steps of current give equal steps on the scale. In a uniform field the torque falls as $NIAB\cos\varphi$, and the meter reads low, worse at large currents.
Part 2 · Going further
The integrals: ∮ dl = 0 and τ = μ × B
For any closed loop in a uniform field, $\vec B$ comes out of the integral:
$\vec F = \oint I\,d\vec l \times \vec B = I\left(\oint d\vec l\right) \times \vec B = 0$
because the displacements round a closed path add to nothing. Any net force needs $\vec B$ to vary across the loop. The torque, by the identity $\oint \vec r \times (d\vec l \times \vec B) = \left(\tfrac12\oint \vec r \times d\vec l\right) \times \vec B$, is
$\vec\tau = \vec\mu \times \vec B$ $\vec\mu = NIA\,\hat n$
for any flat loop: $\tfrac12\oint \vec r \times d\vec l$ is its vector area. The engine adds $\vec r \times d\vec F$ over 192 pieces and agrees.
Energy and small oscillations
Integrating the torque gives the orientation energy in a uniform field, $U = -\vec\mu\cdot\vec B = -\mu B\cos\theta$: lowest aligned, highest anti-aligned, so turning a loop from θ₁ to θ₂ takes $W = \mu B(\cos\theta_1 - \cos\theta_2)$. In any field the exact energy at constant current is $U = -NI\Phi$, with Φ the flux through one turn; then $\tau = -\partial U/\partial\alpha$ for any turn α. The Sandbox plots both, and $-\vec\mu\cdot\vec B$ at the center drifts from $-NI\Phi$ wherever the field varies across the loop.
Near alignment, $J\ddot\theta = -\mu B\,\theta$, so
$\omega^2 = \mu B/J$
For a ring about a diameter, $J = \tfrac12 mR^2$, and $\omega^2 = 2\pi NIB/m$: the radius cancels. A square about an axis through its center parallel to two sides has $J = ma^2/6$, so $\omega^2 = 6NIB/m$. Larger swings take a little longer, by about $1 + \theta_0^2/16$.
Forces in fields that vary: ∇(μ·B) and NI∇Φ
For a small loop the force is $\vec F = \nabla(\vec\mu\cdot\vec B)$, exact in a linear field. In the axial gradient $\vec B = (B_0 + Gz)\hat k - \tfrac G2(x\hat\imath + y\hat\jmath)$ an aligned ring feels $F_z = \mu G$. Exactly, for any loop at constant current, virtual work gives
$\vec F = NI\nabla\Phi$ $\tau_\alpha = NI\,\partial\Phi/\partial\alpha$
which needs only $\nabla\cdot\vec B = 0$. At a sharp edge, moving by $dx$ adds a strip as long as the edge’s cut across the loop, so $\partial\Phi/\partial x = B \times$ (the cut’s length): $BL$ for a square, $2\sqrt{R^2 - x^2}\,B$ for a circle centered at $x$.
How good is the small-loop estimate? Beside a long wire, for a coil whose near side is $a$ from the wire and far side $a + b$, the exact force over the estimate at the center is $1 + b^2/[4a(a+b)]$: it undershoots, and improves as the coil shrinks. Far away a small loop’s force falls as $d^{-2}$ beside a wire and as $z^{-4}$ on a magnet’s axis, where $B \propto z^{-3}$.
Fringes and Earnshaw
No real field ends sharply. Round a thin rectangle across a sharp edge, $\oint \vec B\cdot d\vec l \ne 0$ although no current threads it, which $\nabla \times \vec B = 0$ forbids. So the field bulges out past the edge, with a long tail: at the edge of two pole plates 14 cm apart the mid-plane field is 78 % of $B_0$, and still 20 % one gap width out. A half-in loop then feels less than $NILB$, but the work the field does pulling it in from infinitely far, $NI\Delta\Phi$, is unchanged.
Can a static field hold a loop in place? With $\vec\mu$ fixed, $U = -\vec\mu\cdot\vec B$, and where no current flows each component of $\vec B$ satisfies Laplace’s equation, so $\nabla^2 U = 0$: $U$ has no minimum, and the loop cannot hover stably (Earnshaw). A moment that follows $\vec B$, like an atom’s, has $U = \pm\mu|\vec B|$, and $|\vec B|$ can have a minimum (Wing’s theorem): that is how magnetic traps hold atoms.
The motor’s torque–speed line
With the commutator, $\tau = NIAB\sin\theta$ and $I = (V - NAB\omega\sin\theta)/R$ for θ from 0 to π. Averaged over a turn (ignoring the brushes’ gap),
$\langle\tau\rangle = \frac{NAB}{R}\left(\frac{2V}{\pi} - \frac{NAB\,\omega}{2}\right)$
a straight line: the stall torque at rest, zero at the no-load speed $\omega_0 = 4V/(\pi NAB)$. The mechanical power $\langle\tau\rangle\omega$ peaks at half the no-load speed, at $P_{max} = \tau_{stall}\omega_0/4$. A gap of ±6° at each switch trims both slightly; the engine includes it.
Ballistic throws and damping
A pulse of charge $Q$, short against the meter’s period, kicks the coil to $J\dot\varphi_0 = NAB\int I\,dt = NABQ$. Undamped, it swings to
$\varphi_{max} = NABQ/\sqrt{\kappa J}$
whatever the pulse’s shape: a ballistic galvanometer measures charge. Moving, the coil makes an EMF $NAB\dot\varphi$ that drives a current through its circuit, $R_{total} = R_{coil} + R_{ext}$, and that current brakes it with $(NAB)^2\dot\varphi/R_{total}$. It is critically damped when
$R_{total} = (NAB)^2/2\sqrt{\kappa J}$
Open terminals give no braking; shorted terminals, the most, which is how meters are carried.
Reference
Common mistakes
- Using cos for a wire. $F = ILB\sin\theta$ with θ between the wire and $\vec B$: zero along $\vec B$.
- “The forces cancel, so it doesn’t turn.” In a uniform field $\vec F_{net} = 0$, but $\vec\tau = \vec\mu \times \vec B$ need not be.
- Taking θ from the coil’s plane. θ is between $\vec\mu$ (the axis) and $\vec B$; the torque is largest with the plane along $\vec B$.
- Forgetting $N$, or the area in square meters. $\mu = NIA$ with $A$ in m².
- Scaling the edge force by the part inside. Partly in, $F = NILB$ whatever the fraction.
- “A magnet always attracts a loop.” Only an aligned one; a reversed loop is pushed toward weaker field.
- Treating zero torque as stable. Anti-aligned has τ = 0 but flips at the slightest tilt.
- Ignoring back EMF. A running motor’s current is $(V - \varepsilon)/R$, far below $V/R$.
- Using $-\vec\mu\cdot\vec B$ where the field varies across the loop. The exact energy is $-NI\Phi$.
Equation sheet
| Part 1 | |
|---|---|
| a wire | $F = ILB\sin\theta$ |
| drift | $v_d = I/(neA)$ |
| moment | $\mu = NIA$ |
| torque | $\tau = NIAB\sin\theta$ |
| uniform field | $F_{net} = 0$ |
| partly in, sharp edge, μ along B | $F = NILB$ |
| a long wire | $B = \mu_0 I/(2\pi r)$ |
| $\mu_0 = 4\pi \times 10^{-7}\ \text{T·m/A}$ | |
| motor | $I = (V - \varepsilon)/R$ |
| meter | $\varphi = NIAB/\kappa$ |
| Part 2 | |
|---|---|
| torque | $\vec\tau = \vec\mu \times \vec B$ |
| energy | $U = -\vec\mu\cdot\vec B$ |
| any field | $U = -NI\Phi$ |
| work | $W = \mu B(\cos\theta_1 - \cos\theta_2)$ |
| oscillation | $\omega^2 = \mu B/J$ |
| inertia | $J = ma^2/6$, $J = \tfrac12 mR^2$ |
| small loop | $\vec F = \nabla(\vec\mu\cdot\vec B)$ |
| gradient | $F_z = \mu G$ |
| virtual work | $\vec F = NI\nabla\Phi$ |
| by a wire, exact/estimate | $= 1 + b^2/[4a(a+b)]$ |
| motor line | $\langle\tau\rangle = \frac{NAB}{R}\left(\frac{2V}{\pi} - \frac{NAB\,\omega}{2}\right)$ |
| ballistic | $\varphi_{max} = NABQ/\sqrt{\kappa J}$ |
| critical | $R_{total} = (NAB)^2/2\sqrt{\kappa J}$ |
Symbols: $I$ current (A) · $L$ length of wire (m) · $B$ field (T) · θ angle between the wire, or $\vec\mu$, and $\vec B$ · $N$ turns · $A$ area (m²) · μ moment (A·m²) · τ torque (N·m) · Φ flux through one turn (Wb) · $J$ moment of inertia (kg·m²) · $G$ gradient (T/m) · $V$ supply (V) · ε back EMF (V) · $R$ resistance (Ω) · κ hairspring (N·m/rad) · φ deflection · $Q$ charge (C).
Further reading
Samuel J. Ling, Jeff Sanny and William Moebs, University Physics Volume 2 (OpenStax, 2016), free online under CC BY-NC-SA 4.0: Chapter 11, Magnetic Forces and Fields (the force on a current-carrying conductor; force and torque on a current loop). D. J. Griffiths, Introduction to Electrodynamics, §6.1 (forces and torques on magnetic dipoles). And this lab’s Field Manual (H): how the model works, and what it leaves out.
Loop Torque Lab v1.3 · © 2026 Dr. David Meier · CC BY-NC-SA 4.0 · glidepathscience.org