14.2 Eddy Currents and Inductance
A changing current induces emfs in nearby metal (eddy currents), in a neighbouring coil (mutual inductance, $\varepsilon_2=-M\,dI_1/dt$) and in its own circuit (self-inductance, $\varepsilon=-L\,dI/dt$). An inductor stores energy $\tfrac12LI^2$ in its magnetic field.
Drop a strong magnet down a copper pipe and it drifts down slowly, as if falling through honey, although copper is not magnetic at all. Put a steel pan on an induction cooktop and the pan gets hot while the glass top stays cool. Switch off a large electric motor and a bright spark jumps across the switch. Hold a phone on a wireless charger and its battery fills with no wire between them. Each of these is electromagnetic induction from Section 14.1 at work, but in a new setting: the changing flux is produced by a current, and the emf appears in a lump of metal, in a second coil, or in the very circuit whose current is changing.
In this section we look first at eddy currents — induced currents that swirl inside solid conductors — with their uses and their nuisances. Then we meet self-inductance, the property that makes a coil resist changes in its own current, and mutual inductance, which couples two coils and is the basis of the transformer. Finally we find the energy stored in an inductor, and see that it lives in the magnetic field itself.
Eddy currents
Faraday's law does not need a wire. Whenever the magnetic flux through any closed path inside a piece of metal changes — because the metal moves through a non-uniform field, or the field through it changes — an emf is induced round that path and a current flows. These currents circulate in closed loops in the bulk of the metal, like eddies (whirlpools) in a stream.
Eddy currents are induced currents that flow in closed loops inside a bulk conductor when the magnetic flux through it changes. By Lenz's law they always oppose the change that produces them: they oppose the relative motion between the conductor and the field (electromagnetic damping), and they heat the conductor at the rate $I^2R$ summed over all the loops.
Because a block of metal has a very low resistance, even a small induced emf drives a large eddy current, so the forces and the heating can be large. Three classic demonstrations show this.
- Waltenhofen's pendulum. A copper plate swinging between the poles of a strong magnet stops within a swing or two: eddy currents in the plate are pushed by the field against the motion. Cut slots into the plate and it swings almost freely — the slots break the large current loops, the paths become long and narrow, their resistance rises and the eddy currents (and the damping) fall sharply.
- Magnet in a copper pipe. The falling magnet changes the flux through every ring of the pipe below it (increasing) and above it (decreasing). The ring below repels it and the ring above attracts it: both push up. The braking force grows with the speed, so the magnet soon falls at a steady terminal speed, exactly like the rod on rails in Section 14.1. In a plastic pipe there are no currents and it falls freely; a pipe with a lengthwise slit brakes much less.
- Thomson's jumping ring. An aluminium ring placed over an iron core leaps upward when an alternating current is switched on in a coil below: the induced current in the ring always opposes the coil's changing field, so the ring is repelled.
Uses of eddy currents.
- Electromagnetic braking in some trains, roller-coasters and exercise bikes: magnets near a metal disc or rail create eddy currents that slow it without contact or wear. The braking force falls as the speed falls, so these brakes are smooth, but they cannot hold a vehicle at rest.
- Electromagnetic damping: the coil of a moving-coil galvanometer is wound on a metal frame, so eddy currents in the frame bring the pointer to rest quickly at its reading (a "dead-beat" galvanometer). Some balances use a metal vane between magnets for the same reason.
- Induction furnaces and induction cooktops: a high-frequency alternating current in a coil produces a rapidly changing field that induces large eddy currents in the metal to be heated — the metal heats itself. A glass or ceramic pan has no free electrons, so nothing happens in it.
- Electric power meters (older types) and speedometers: a magnet rotated by the wheels drags an aluminium cup round with eddy currents against a spring; the deflection measures the speed.
- Metal detectors, coin sorters in vending machines and the separation of aluminium from household waste also rely on eddy currents.
Eddy currents as a nuisance. In the iron cores of transformers, motors and generators the flux changes all the time, and eddy currents waste energy as heat. They are reduced by making the core from thin laminations (sheets) insulated from one another by varnish or oxide, with the sheets lying along the field. The insulating layers cut across the paths in which eddy currents would flow, so each current loop is confined to a thin sheet with a high resistance. Laminating reduces eddy currents greatly; it does not remove them completely.
In Interactive 1 keep the magnet strength at $1.0$ (in units of the standard magnet) and read the solid copper pipe: terminal speed $0.200\ \text{m/s}$, fall time about $5.0\ \text{s}$. Predict the fall time through the copper pipe when the magnet's strength is doubled.
The iron core of a transformer is built from thin insulated sheets instead of one solid block. The main reason is to
Self-inductance
A coil carrying a current $I$ sets up a magnetic field, and some of that field threads its own turns. If the current changes, its own flux linkage changes, and by Faraday's law an emf is induced in the coil itself. By Lenz's law this emf opposes the change of current: it tries to hold the current back when it is rising and to keep it going when it is falling. It is therefore called a back emf.
With no magnetic material nearby (or an iron core well below saturation), the field — and so the flux linkage — is proportional to the current. The constant of proportionality depends only on the coil's size, shape, number of turns and core.
The self-inductance $L$ of a coil is the flux linkage per unit current in it:
$$N\Phi_B=LI,\qquad\text{so}\qquad\varepsilon=-\frac{d(N\Phi_B)}{dt}=-L\frac{dI}{dt}.$$Equivalently, $L$ is the back emf induced per unit rate of change of current. Its SI unit is the henry: $1\ \text{H}=1\ \text{Wb/A}=1\ \text{V s/A}=1\ \Omega\,\text{s}$; dimensions $[\text{ML}^2\text{T}^{-2}\text{A}^{-2}]$. A coil designed to have a large self-inductance is called an inductor.
Self-inductance plays the part of inertia for electric current, just as mass resists changes of velocity: an inductor does not oppose a steady current at all (a steady current in an ideal inductor needs no voltage), but it opposes every change. That is why the current in a circuit containing a large inductor rises gradually when the switch is closed, and why breaking such a circuit suddenly — making $dI/dt$ enormous — produces a large back emf that can make a spark jump across the switch.
Self-inductance of a long solenoid. Take a solenoid of length $l$, cross-sectional area $A$ and $N$ turns, so $n=N/l$ turns per metre, with $l$ much larger than its radius. Inside it the field is uniform, $B=\mu_0nI$ (Ampère's law, Chapter 13), and the flux through each turn is $BA$. The flux linkage is
$$N\Phi_B=(nl)(\mu_0nI)A=\mu_0n^2AlI,\qquad\text{so}\qquad L=\mu_0n^2Al=\frac{\mu_0N^2A}{l}.$$A long solenoid of length $l$, cross-sectional area $A$ and $N$ turns ($n=N/l$ per metre), with end effects neglected, has
$$L=\mu_0n^2Al=\frac{\mu_0N^2A}{l}.$$Filled with a core of relative permeability $\mu_r$, $L=\mu_r\mu_0n^2Al$. $L$ depends only on the geometry and the core — not on the current — and grows as the square of the number of turns. Inductors far apart (no shared flux) combine like resistors: $L=L_1+L_2$ in series, $\dfrac1L=\dfrac1{L_1}+\dfrac1{L_2}$ in parallel.
Potential difference across an inductor. In circuit problems treat an inductor like a cell whose emf is $-L\,dI/dt$. Going through an inductor in the direction of the current, the potential falls by $L\,\dfrac{dI}{dt}$ (plus $IR$ if the coil has resistance $R$). If the current is decreasing, $dI/dt$ is negative and the potential rises: the inductor then acts as a source, pushing the current onward.
- Mark the direction of the current and whether it is increasing or decreasing; write $dI/dt$ with its sign.
- Walk from one end of the branch to the other in the direction of the current, adding up the changes of potential: $-IR$ across a resistor, $+\varepsilon$ across a cell from its negative to its positive terminal ($-\varepsilon$ the other way), $-L\,dI/dt$ across an inductor.
- The sum is $V_\text{end}-V_\text{start}$.
A long air-core solenoid has self-inductance $L$. It is rewound with twice as many turns on the same length and the same tube. Its self-inductance becomes
Mutual inductance
Place two coils near each other. A current $I_1$ in coil 1 sends some of its field through coil 2, giving coil 2 a flux linkage $N_2\Phi_{21}$ proportional to $I_1$. If $I_1$ changes, an emf is induced in coil 2 — although nothing connects the two circuits. This is Faraday's original experiment, and it is how transformers, wireless chargers and the ignition coil of a petrol engine work.
The mutual inductance $M$ of a pair of coils is the flux linkage of one coil per unit current in the other:
$$N_2\Phi_{21}=MI_1,\qquad\text{so}\qquad\varepsilon_2=-M\frac{dI_1}{dt}.$$Its unit is the henry. $M$ depends on the shapes, sizes and numbers of turns of the coils, on their separation and relative orientation, and on any core between them — not on the currents.
A remarkable fact makes $M$ a property of the pair: the flux linkage of coil 2 per unit current in coil 1 always equals the flux linkage of coil 1 per unit current in coil 2, $M_{21}=M_{12}=M$, however different the coils are. We can check it for the most important case.
Two coaxial solenoids. A long solenoid $S_1$ (radius $r_1$, $n_1$ turns per metre) lies inside a longer solenoid $S_2$ (radius $r_2\gt r_1$, $n_2$ turns per metre), both of length $l$. (i) A current $I_2$ in the outer solenoid gives a field $\mu_0n_2I_2$ everywhere inside it, so the flux linkage of the inner coil is $(n_1l)(\mu_0n_2I_2)(\pi r_1^2)$: $M_{12}=\mu_0n_1n_2\pi r_1^2l$. (ii) A current $I_1$ in the inner solenoid gives a field $\mu_0n_1I_1$ inside it and (almost) zero outside it, so although the outer turns enclose a larger area, the flux through each is only $\mu_0n_1I_1\pi r_1^2$. Its flux linkage is $(n_2l)(\mu_0n_1I_1)(\pi r_1^2)$: $M_{21}=\mu_0n_1n_2\pi r_1^2l$ — the same.
(i) Reciprocity: $M_{12}=M_{21}=M$. The emf induced in either coil by a given rate of change of current in the other is the same.
(ii) Coaxial long solenoids of common length $l$, the inner of radius $r_1$:
$$M=\mu_0n_1n_2\pi r_1^2\,l\qquad(\text{the smaller, inner area}),$$multiplied by $\mu_r$ if a core of relative permeability $\mu_r$ fills the inner solenoid.
(iii) Coupling: for any two coils $M=k\sqrt{L_1L_2}$ with $0\le k\le1$; $k=1$ (perfect coupling) only when every line of flux of each coil threads all the turns of the other. Hence $M\le\sqrt{L_1L_2}$.
Another useful pair: a small loop of radius $r$ at the centre of a large coplanar loop of radius $R\gg r$. A current $I$ in the large loop gives a field $\dfrac{\mu_0I}{2R}$ at its centre (Biot–Savart law, Chapter 13), nearly uniform over the small loop, so $M=\dfrac{\mu_0\pi r^2}{2R}$. Computing it the other way — the flux of the small loop's dipole field through the large loop — is hard, but reciprocity guarantees the same answer.
For two coaxial long solenoids of equal length, the inner of radius $r_1$ and the outer of radius $r_2$, the mutual inductance depends on
Energy stored in an inductor
To set up a current in an inductor, the source must push charge against the back emf. When the current is $I$ and rising at $dI/dt$, the back emf is $L\,dI/dt$, and the source does work against it at the rate
$$P=\left(L\frac{dI}{dt}\right)I,\qquad\text{so}\qquad dW=LI\,dI.$$Building the current up from $0$ to $I$ needs $W=\displaystyle\int_0^ILI\,dI=\tfrac12LI^2$. This work is not lost as heat: it is stored, and when the current falls the inductor gives it back (that is the energy in the spark when an inductive circuit is broken).
An inductor of self-inductance $L$ carrying a current $I$ stores energy
$$U=\tfrac12LI^2=\frac{(N\Phi_B)^2}{2L}=\tfrac12(N\Phi_B)I.$$The energy depends on the current, not on how fast it is changing; it is the magnetic counterpart of $\tfrac12CV^2$ for a capacitor and of $\tfrac12mv^2$ for a moving body ($L$ plays the part of mass, $I$ of velocity).
Where is the energy? For a long solenoid, $L=\mu_0n^2Al$ and $B=\mu_0nI$, so
$$U=\tfrac12\mu_0n^2AlI^2=\frac{(\mu_0nI)^2}{2\mu_0}\,(Al)=\frac{B^2}{2\mu_0}\times\text{volume}.$$The energy is spread through the space where the field is, with the same amount in every cubic metre. Although derived for a solenoid, this result holds for any magnetic field in vacuum (or air).
A magnetic field $B$ in vacuum (or air) stores energy per unit volume
$$u_B=\frac{B^2}{2\mu_0},$$in $\text{J/m}^3$ — the magnetic counterpart of $u_E=\tfrac12\varepsilon_0E^2$.
In Interactive 2 set $N=500$, $l=25\ \text{cm}$, $A=4.0\ \text{cm}^2$, air core: $L=0.503\ \text{mH}$, peak stored energy $1.01\ \text{mJ}$ and a back emf of $0.201\ \text{V}$ while the current falls. Predict the peak stored energy when you double the turns to $N=1000$.
LR circuits are not named in the JEE Main 2026 syllabus list (they are listed for JEE Advanced 2026). This box is for interest only; nothing here is used in the exercises, the Exam corner or the quiz.
Connect an inductor $L$ in series with a resistance $R$ to a cell of emf $\varepsilon$. Kirchhoff's loop rule gives $\varepsilon-L\dfrac{dI}{dt}-IR=0$, and the current grows smoothly towards $\varepsilon/R$:
$$I=\frac{\varepsilon}{R}\left(1-e^{-t/\tau}\right),\qquad\tau=\frac LR.$$After one time constant $\tau$ the current has reached $63\%$ of its final value. If the cell is then removed and the coil is shorted through $R$, the current decays as $I=I_0e^{-t/\tau}$, and the stored energy $\tfrac12LI_0^2$ ends up as heat in $R$. The equations have exactly the same form as those of the rod on rails in Section 14.1, with $L$ in place of $m$.
1. $L$ is a property of the coil, not of the current. $L=N\Phi_B/I$ is constant for a given coil and core; doubling the current doubles the flux, not $L$. A steady current produces no back emf, however large $L$ is.
2. $N^2$, and $n$ versus $N$. $L=\mu_0N^2A/l=\mu_0n^2Al$ — mixing the two (e.g. $\mu_0n^2A/l$) is the usual slip. Twice the turns on the same length gives $4L$; twice the turns and twice the length (same $n$) gives $2L$.
3. Coaxial solenoids use the inner area. $M=\mu_0n_1n_2\pi r_1^2l$, the same whichever coil carries the current. And $M\le\sqrt{L_1L_2}$ always.
4. Signs along a branch. Through an inductor in the direction of the current the potential changes by $-L\,dI/dt$ with the signed $dI/dt$: a decreasing current makes the inductor raise the potential.
5. Energy depends on $I$, not on $dI/dt$. $U=\tfrac12LI^2$ is largest when the current is largest, even if the back emf is then zero. And laminating a core reduces eddy-current loss; it does not stop the flux or remove eddy currents completely.
The current through an inductor is doubled. The energy stored in it
Exercises
Take $\mu_0=4\pi\times10^{-7}\ \text{T m/A}$ and $g=9.8\ \text{m/s}^2$. Solenoids are long and air-cored unless told otherwise; end effects are neglected.
- 1 A solid copper plate swinging between the poles of a magnet stops quickly, but a plate with deep slots cut in it swings for a long time. Explain.
- 2 Why is a transformer core made of thin insulated sheets, and why must the sheets lie along the magnetic field rather than across it?
- 3 Why does an induction cooktop heat a steel pan but not a glass dish? Where is the heat produced?
- 4 A magnet of mass $40\ \text{g}$ falls through a vertical copper pipe $1.5\ \text{m}$ long at a steady $0.25\ \text{m/s}$. Find the time taken and the heat produced in the pipe.
- 5 An eddy-current brake slows a $2.0\times10^5\ \text{kg}$ train on level track from $20\ \text{m/s}$ to $5.0\ \text{m/s}$. How much heat is produced (ignore other losses)? Why can such a brake not hold the train at rest?
- 6 Why is the coil of a moving-coil galvanometer wound on a metal frame?
- 7 A copper disc spins freely on an axle. When a magnet is brought close to its rim, it slows down. Where does its kinetic energy go?
- 1 Find the self-inductance of a solenoid of $800$ turns, length $40\ \text{cm}$ and cross-sectional area $5.0\ \text{cm}^2$, in mH.
- 2 A current of $2.0\ \text{A}$ in a $200$-turn coil produces a flux of $5.0\times10^{-4}\ \text{Wb}$ through each turn. Find its self-inductance.
- 3 The current in a $0.30\ \text{H}$ coil falls steadily from $4.0\ \text{A}$ to $1.0\ \text{A}$ in $0.050\ \text{s}$. Find the induced emf.
- 4 The current in a $20\ \text{mH}$ inductor is $I=(3t^2+2t)\ \text{A}$, with $t$ in seconds. Find the back emf at $t=2.0\ \text{s}$.
- 5 The solenoid of part 1 is extended to twice the length with twice the number of turns (same turns per metre and area). Find its new self-inductance.
- 6 Inductors of $4.0\ \text{mH}$ and $6.0\ \text{mH}$, far apart, are connected (a) in series and (b) in parallel. Find the combined inductance in each case.
- 7 A branch runs from A to B through a $5.0\ \Omega$ resistor and an ideal $10\ \text{mH}$ inductor. A current of $2.0\ \text{A}$ flows from A to B and is increasing at $100\ \text{A/s}$. Find $V_A-V_B$. What is it if the current is decreasing at the same rate?
- 8 Explain why a spark often appears at a switch when a circuit containing a large inductor is opened, but not when it is closed.
- 1 When the current in coil A changes at $40\ \text{A/s}$, an emf of $0.20\ \text{V}$ is induced in coil B. Find their mutual inductance, in mH.
- 2 For the pair of part 1, the current in B rises steadily from $0$ to $3.0\ \text{A}$ in $0.010\ \text{s}$. Find the emf induced in A.
- 3 A long solenoid with $1500$ turns per metre and cross-sectional area $10\ \text{cm}^2$ has a $50$-turn coil wound tightly round its middle. Find their mutual inductance, in μH.
- 4 The current in the solenoid of part 3 falls uniformly from $3.0\ \text{A}$ to zero in $2.0\ \text{ms}$. Find the emf induced in the small coil.
- 5 Two coils have $L_1=0.20\ \text{H}$, $L_2=0.80\ \text{H}$ and $M=0.30\ \text{H}$. Find the coupling coefficient.
- 6 What is the largest possible mutual inductance of two coils of self-inductances $9.0\ \text{mH}$ and $4.0\ \text{mH}$?
- 7 A current of $2.0\ \text{A}$ in coil 1 produces a flux of $2.0\times10^{-5}\ \text{Wb}$ through each of the $400$ turns of coil 2. Find $M$, and the flux linkage of coil 1 when instead $2.0\ \text{A}$ flows in coil 2.
- 8 A loop of radius $2.0\ \text{cm}$ lies at the centre of a coplanar loop of radius $50\ \text{cm}$. Find their mutual inductance.
- 1 Find the energy stored in a $0.20\ \text{H}$ inductor carrying $3.0\ \text{A}$.
- 2 A coil stores $25\ \text{mJ}$ when it carries $0.50\ \text{A}$. Find its self-inductance.
- 3 Find the energy density of a magnetic field of $0.50\ \text{T}$ in air.
- 4 The current in a $50\ \text{mH}$ coil rises from $2.0\ \text{A}$ to $6.0\ \text{A}$. Find the increase in stored energy.
- 5 A solenoid with $2000$ turns per metre, cross-sectional area $2.0\ \text{cm}^2$ and length $30\ \text{cm}$ carries $3.0\ \text{A}$. Find the field inside, the energy density and the total energy stored, in mJ.
- 6 At an instant when the current in a $0.40\ \text{H}$ inductor is $2.0\ \text{A}$ and increasing at $5.0\ \text{A/s}$, at what rate is energy being stored in it?
- 7 The solenoid of part 5 is filled with a core of relative permeability $400$ (assume no saturation) and carries the same current. Find the energy stored.
- 8 A fixed length of wire is wound into a long solenoid of a given length. It is unwound and rewound on the same length but on a tube of half the radius. Show that the self-inductance does not change.
E1. JEE Main · MCQ A long air-core solenoid has self-inductance $L$. Another solenoid of the same cross-section has twice the number of turns and half the length. Its self-inductance is (A) $2L$ (B) $4L$ (C) $8L$ (D) $L/2$
E2. JEE Main · Numerical When the current in a coil rises steadily from $2.0\ \text{A}$ to $4.0\ \text{A}$ in $0.050\ \text{s}$, the back emf is $8.0\ \text{V}$. Find the energy, in mJ, stored in the coil when the current is $4.0\ \text{A}$. Give the answer to the nearest integer.
E3. JEE Main · MCQ A branch of a circuit runs from A to B through a $3.0\ \Omega$ resistor, then a $12\ \text{V}$ cell (negligible internal resistance) with its positive terminal towards B, then an ideal $2.0\ \text{mH}$ inductor. A current of $2.0\ \text{A}$ flows from A to B and is increasing at $2000\ \text{A/s}$. Then $V_A-V_B$ is (A) $-2.0\ \text{V}$ (B) $22\ \text{V}$ (C) $2.0\ \text{V}$ (D) $-10\ \text{V}$
E4. JEE Main · Numerical A solenoid of radius $1.0\ \text{cm}$ with $2000$ turns per metre lies coaxially inside a solenoid of radius $3.0\ \text{cm}$ with $500$ turns per metre; both are $1.0\ \text{m}$ long. Find their mutual inductance in μH. (Take $\mu_0=4\pi\times10^{-7}\ \text{T m/A}$ and $\pi^2=10$.) Give the answer to the nearest integer.
E5. JEE Main · MCQ Which of the following does not depend on eddy currents? (A) an induction furnace melting steel (B) electromagnetic braking of a train (C) dead-beat damping of a moving-coil galvanometer (D) the heating element of an electric kettle
E6. JEE Main · Numerical A $2.0\ \text{H}$ inductor carries a current of $3.0\ \text{A}$. The current is reduced to $1.0\ \text{A}$ in $0.20\ \text{s}$. Find the average rate, in watts, at which the inductor gives up its stored energy during this time. Give the answer to the nearest integer.
Research question. How does the braking of a magnet falling down a metal pipe depend on the pipe's material and wall thickness — and does the terminal speed follow $v_t=mg/k$ with $k$ proportional to the metal's conductivity?
Why it matters. Eddy-current brakes stop roller-coasters at theme parks around the world, slow high-speed trains in Germany and Japan without touching the rail, and control the resistance of exercise bikes. Their designers must predict the braking force from the conductor's material and shape. The same physics limits how much energy is lost in transformer cores across every national grid. A magnet and a few pipes reproduce the whole story on a desk.
What to do.
- Collect pipes of the same length (about $1\ \text{m}$) and similar bore: copper, aluminium, brass and plastic; if possible copper pipes of two wall thicknesses, and one copper pipe with a lengthwise saw-cut.
- Choose a neodymium magnet that slides freely in the pipes. Measure its mass.
- Drop it through each pipe five times and time the fall with a stopwatch or a phone video. Find the average time and the terminal speed $v_t\approx\text{length}/\text{time}$.
- Compute $k=mg/v_t$ for each pipe. Look up the electrical conductivity of each metal and plot $k$ against conductivity.
- Compare the solid and the slit copper pipes, and the thin and thick walls.
Tools: Metal and plastic pipes, a strong magnet, stopwatch or phone camera, metre rule, balance, conductivity table, spreadsheet.
Hand in: A results table with uncertainties, the graph of $k$ against conductivity, and a paragraph explaining the effects of the slit and of wall thickness in terms of the paths and resistance of the eddy currents.
Going further. Stand the pipe on a kitchen scale and film the reading while the magnet falls. Explain the extra reading using Newton's third law and compare it with $mg$. Compare with Interactive 1, which uses the same model.
🧪 Self-check quiz
Q1. The henry is equivalent to
Q2. The self-inductance of a long solenoid of fixed length and cross-section is proportional to
Q3. Eddy-current losses in the core of a transformer are reduced by
Q4. The energy stored in an inductor carrying current $I$ is $U$. The flux linkage of the inductor is
Q5. Two coils have self-inductances $16\ \text{mH}$ and $4.0\ \text{mH}$. Their mutual inductance cannot be