Electrical Engineering › Section H · Magnetic circuits and inductors › Topic H.3
Ampere's circuital law and magnetic circuits
Relays, transformers and motors all guide flux round a closed path of iron, often with a small air gap. Ampère's circuital law turns such a path into a circuit you can solve, much as Kirchhoff's laws do for resistors.
Magnetic circuits
A magnetic circuit is a closed path for flux: a ring, a rectangular frame, or a shell core with three legs, carrying one or more windings. The previous two topics gave the matching quantities, set out here side by side:
Each section of core with its own material, area or length plays the part of one resistor. With that picture, the familiar dc laws carry straight across.
Ampère's circuital law
Kirchhoff's voltage law says the rises and drops of voltage round a closed loop sum to zero. Ampère's circuital law is the magnetic version: round any closed loop of a magnetic circuit, the mmf rises equal the mmf drops,
Each winding supplies a rise of \(NI\), and each section of core uses up a drop of \(Hl\), its magnetizing force times its length. When two windings act on one core, give each a sign with the right-hand rule: a winding pushing flux the other way counts as negative, so \(N_1I_1 - N_2I_2 = \sum Hl\).
Flux at a junction
Kirchhoff's current law also has a twin. Where a core branches, the flux entering the junction equals the flux leaving it: \(\Phi_1 = \Phi_2 + \Phi_3\). In a shell core, the winding sits on the centre leg and its flux splits between two outer legs that act like parallel branches. If the two outer paths are identical, each carries half the centre flux. In a simple series circuit, such as a ring or a frame with one winding, there are no junctions, so the same flux passes through every section and only the flux density changes where the area changes.
Why the table method beats reluctance
For resistors you can add resistances and divide. For iron this rarely works, because \(\mu\) and so \(\mathcal{R}\) change with the flux density, as the topic "Reluctance, magnetizing force, and hysteresis" showed. Engineers instead work section by section with a small table, columns for \(\Phi\), \(A\), \(B\), \(H\), \(l\) and \(Hl\). For each section: find \(B = \Phi/A\), read \(H\) from that material's B-H curve, multiply by its length, then add the drops round the loop. The same nonlinearity rules out superposition: you cannot find the flux from each winding separately and add.
Air gaps
Relays, motors and many inductor cores contain a deliberate air gap. Flux spreads out slightly as it crosses a gap, called fringing; for hand calculations it is ignored, so the gap is given the same flux and area as the core beside it. Air is linear, with \(\mu = \mu_0\), so no curve is needed:
A gap only a millimetre long can therefore need far more mmf than the whole iron path. Designers accept that cost because a gap makes the circuit more linear, stops the core saturating early and lets an armature move, as in a relay.
Worked example: a core with an air gap
A sheet-steel frame has an iron path of mean length 0.18 m, a 1.0 mm air gap and a cross-section of \(2.5 \times 10^{-4}\,\text{m}^2\) throughout. A 400-turn coil must set up \(2.0 \times 10^{-4}\,\text{Wb}\). The steel's curve gives \(H = 200\,\text{At/m}\) at 0.80 T. Find the current. 1. Flux density, the same everywhere: \(B = \Phi/A = 0.80\,\text{T}\). 2. Iron drop: \(Hl = 200 \times 0.18 = 36\,\text{At}\). 3. Gap: \(H_g = 0.80/\mu_0 = 6.37 \times 10^5\,\text{At/m}\), so the gap drop is \(H_g \times 0.0010 = 637\,\text{At}\). 4. Ampère's law: \(NI = 36 + 637 = 673\,\text{At}\), so \(I = 673/400 = 1.68\,\text{A}\). 5. Check: the gap takes about 95% of the mmf, as expected for air.
Worked example: a shell core
A shell core has a centre leg of length 0.10 m and area \(8.0 \times 10^{-4}\,\text{m}^2\). Each outer path, outer leg plus yokes, is 0.25 m long with area \(4.0 \times 10^{-4}\,\text{m}^2\). A 300-turn coil on the centre leg carries 0.50 A. Assume \(\mu_r = 2000\), constant at this level. 1. Centre reluctance: \(0.10/(2000\mu_0 \times 8.0 \times 10^{-4})\), which is \(4.97 \times 10^4\,\text{At/Wb}\). 2. Each outer path: \(2.49 \times 10^5\,\text{At/Wb}\); the two in parallel give half, \(1.24 \times 10^5\,\text{At/Wb}\). 3. Total: \(1.74 \times 10^5\,\text{At/Wb}\), so with \(NI = 150\,\text{At}\),
4. Each outer leg carries half, \(4.31 \times 10^{-4}\,\text{Wb}\). 5. Check: both regions sit at 1.08 T, so \(H = 429\,\text{At/m}\) everywhere, and \(429 \times (0.10 + 0.25) = 150\,\text{At}\), matching \(NI\).
The reverse problem: finding the flux
Sometimes the winding is known and the flux is wanted. With a single uniform section this is direct: \(H = NI/l\), read \(B\) from the curve, then \(\Phi = BA\). For a cast-steel ring with a 0.40 m mean path, 160 turns and 1.5 A, \(H = 240/0.40 = 600\,\text{At/m}\); if its curve gives 1.0 T there and the area is \(5.0 \times 10^{-4}\,\text{m}^2\), then \(\Phi = 5.0 \times 10^{-4}\,\text{Wb}\). With several sections, such as iron plus a gap, there is no direct route. Guess a flux, often by first assuming the gap takes all the mmf, work out the \(NI\) it needs, compare with the actual \(NI\), adjust the guess and repeat until the two agree within a few per cent. This is often called the cut-and-try method.
Magnetic circuits at work
A relay coil pulls an iron armature across a gap; as the gap closes, reluctance falls and the flux, and so the pull, grows. A loudspeaker's voice coil sits in the narrow radial gap of a permanent-magnet circuit, where the force \(F = BIl\) from the topic "Magnetic fields, flux density, and permeability" drives the cone. A reed switch closes because flux from a nearby magnet takes the low-reluctance path through two iron reeds and pulls them together. A Hall-effect sensor, a semiconductor chip carrying a fixed current, gives an output voltage in proportion to the flux density through it, so it can measure fields, sense position or measure current without touching the conductor.
Common mistakes to avoid
First, forgetting the air gap or using its length in millimetres: a 1 mm gap is 0.001 m. Second, using one value of \(H\) for sections of different materials; each needs its own curve. Third, adding reluctances of saturating iron as if they were fixed resistors. Fourth, giving opposing windings the same sign in Ampère's law. Fifth, in a shell core, forgetting that each outer leg carries only part of the centre flux. Sixth, subtracting the gap from the iron length twice, or not at all, when the problem quotes the total frame length.
Key takeaways
A magnetic circuit maps onto an electric one: mmf for emf, flux for current, reluctance for resistance, \(Hl\) for \(IR\). Ampère's circuital law says the mmf rises round a loop equal the drops, \(\sum NI = \sum Hl\), and flux into a junction equals flux out. Because iron is nonlinear, solve section by section: \(B = \Phi/A\), \(H\) from the curve, then \(Hl\). An air gap needs \(H_g = B_g/\mu_0\) and usually dominates the mmf. Finding \(\Phi\) from a given \(NI\) with several sections takes trial and adjustment.
Flashcards
Flashcard
mmf ↔ emf, flux ↔ current, reluctance ↔ resistance
Test your understanding
Question 1 of 5
What does Ampère's circuital law state for a closed loop of a magnetic circuit?
Question 2 of 5
An air gap 0.80 mm long carries a flux density of 0.50 T. Ignoring fringing, what mmf does the gap use?
Question 3 of 5
A 200-turn coil drives a series circuit in which the iron needs 40 At and the air gap needs 360 At. What current is required?
Question 4 of 5
The centre leg of a three-leg core carries 6.0 × 10⁻⁴ Wb. One outer leg carries 2.2 × 10⁻⁴ Wb. What does the other outer leg carry?
Question 5 of 5
Why do engineers usually solve iron magnetic circuits section by section with B-H curves, not by adding reluctances?
