Electromagnetic Forming: Shaping Metal With a Magnetic Pulse

The Circuit: A Capacitor Bank Ringing Into a Coil

An electromagnetic forming machine is, electrically, a very large series RLC circuit. A bank of pulse-rated capacitors — typically 100–1,000 µF — is charged to 3–25 kV, storing E = ½CV², which lands in the 5–100 kJ range for industrial machines. A triggered spark gap, an ignitron, or a thyristor stack then closes the loop into the forming coil.

Because the circuit is underdamped, the current does not simply flow — it rings. The discharge is a damped sinusoid whose frequency is set almost entirely by bank capacitance and coil inductance, f ≈ 1/(2π√(LC)), and whose amplitude is I₀ ≈ V·√(C/L). With a coil inductance of only 0.1–5 µH, that puts the ring at 5–50 kHz and the peak current in the hundreds of kiloamps. The entire useful event is the first half-cycle: current peaks in a quarter period of roughly 10–30 µs, and the part has left the coil before the second oscillation matters. Stray inductance is therefore the enemy — hence many capacitors in parallel, coaxial busbars, the shortest possible leads — and a crowbar switch is often fitted to clamp the voltage reversal the later, useless rings would impose on the bank.

Where the Force Comes From: Induced Currents and Magnetic Pressure

The coil's current builds a magnetic field in the small gap — usually only 1–3 mm — between coil and workpiece. That field changes at tens of kilohertz, so by Faraday's law it drives an EMF around any closed conducting loop nearby, and by Lenz's law the resulting eddy current flows opposite to the coil current. Two antiparallel currents repel — equivalently, the induced current sheet J sits in the field B and feels a Lorentz body force J × B pointed away from the coil.

The cleanest way to size the effect is as a pressure. A field B running parallel to a surface that excludes it exerts magnetic pressure p = B²/2µ₀; more exactly the net push is (B_outside² − B_inside²)/2µ₀, so the pressure exists only to the extent the metal keeps the field out. The numbers are brutal: 10 T gives ~40 MPa, 20 T about 160 MPa, 50 T roughly 1 GPa. Since annealed aluminium yields around 50–150 MPa, 20–50 T in the gap is comfortably enough to make it flow.

Velocity follows from the impulse divided by areal mass, v ≈ (1/ρt)∫p dt. A 1 mm aluminium sheet weighs 2.7 kg/m², so roughly 0.3–0.8 kPa·s — an average 50 MPa held for about 10 µs — accelerates it to 100–300 m/s. The coupling then collapses on its own: as the part flies away the gap grows, inductance rises, and pressure falls off sharply. Electromagnetic forming is intrinsically a short-range, one-shot push.

Why the Pulse Must Be Short: Skin Depth and Field Diffusion

Everything depends on the workpiece behaving like a magnetic mirror for the duration of the pulse. If the field soaks through, B_inside approaches B_outside, the net pressure collapses, and the energy is wasted as Joule heat. The governing length is the skin depth, δ = √(ρ / π f µ). At a 20 kHz ring that is about 0.6 mm in aluminium and 0.46 mm in copper — comfortably less than the 1–3 mm walls being formed. The design rule is to keep the wall at least one, preferably two to three, skin depths thick; equivalently, the pulse must be short against the magnetic diffusion time τ ≈ µ₀σt², only tens of microseconds for a millimetre of aluminium. Hence a discharge engineered to ring fast rather than long.

This sorts the materials. Aluminium alloys (~30–60 % IACS) and copper (100 % IACS) are ideal. Carbon steel is roughly 10 % IACS, six times more resistive, so its skin depth at the same frequency is around 1.4 mm — larger than the sheet — and its ferromagnetism does not help, because steel saturates near 2 T while these coils reach 10–50 T, driving the effective permeability back toward µ₀. Steel is stronger too, so it needs more pressure to begin with. The fix is a driver: a thin aluminium or copper ring or sheet between coil and workpiece that absorbs the magnetic pressure and transmits it by contact — a trick that also extends EMF to titanium, magnesium, and non-conductive layers.

Three Configurations: Compression, Expansion, and Flat Forming

Compression (crimping) is the workhorse. A helical solenoid surrounds a tube, and the field in the annulus collapses the tube inward onto a mandrel, a splined shaft, a fitting, or another tube — how driveshafts get joined to yokes, filter canisters get sealed, and cable lugs get swaged onto conductors. Because the pressure is a pure hoop load applied all the way round at once, the crimp is concentric and repeatable in a way a segmented mechanical crimper cannot match.

Expansion inverts it: the coil goes inside the tube and drives the wall outward — flaring an end, or slamming the tube into an outer die, as in tube-to-tubesheet expansion in heat exchangers. Flat or sheet forming uses a spiral pancake coil laid parallel to a blank, pushing the sheet into a single-sided die cavity; no matching punch is needed, because the magnetic pressure is the punch. It serves embossing, shearing against a sharp die edge, and local calibration of conventionally pre-formed sheets.

Compression setups very often add a field shaper (concentrator): a massive slotted block of copper or beryllium copper between coil and part. Current induced on its outer face is forced to return along a much smaller inner bore, so the same total current circulates around a smaller area and the field — hence B²/2µ₀ — rises sharply exactly where it is wanted. Field shapers also localise the pressure to a narrow band, are cheap to swap for different diameters, and take the abuse instead of the expensive multi-turn coil.

Why Speed Changes the Metal: Inertial Ductility and Vanishing Springback

A part travelling 200 m/s and deforming over a 20 mm span is straining at roughly 10⁴ s⁻¹ — three to four orders of magnitude faster than a mechanical press. Metals behave measurably differently there. Balanethiram and Daehn at Ohio State reported in the early 1990s that aluminium 6061 and OFHC copper reached forming limits two to several times their quasi-static values under electromagnetic launch, and coined the term hyperplasticity for the effect.

Three mechanisms are credited. Inertial stabilisation of necking: a neck can only grow by accelerating material sideways into it, and at high velocity the metal's own inertia resists that, so the instability that normally ends a tensile test is delayed. Constrained impact with the die: hitting a rigid surface at 100+ m/s puts the sheet into through-thickness compression exactly where it is thinning, suppressing fracture — which is why free-formed parts gain far less than die-impacted ones, and why much of the apparent "strain-rate ductility" is really an impact effect. And genuine constitutive changes at high strain rate, including local adiabatic heating.

Springback is the other prize. In stamping, elastic recovery on unloading pulls the part back off the die and forces engineers to overbend and compensate. In EMF the part arrives with kinetic energy, and the impact sends a reversed through-thickness stress wave that relaxes those bending residual stresses instead of storing them, so the feature holds its shape far more faithfully — springback on a calibrated radius is routinely pulled down to a degree or less. That drives electromagnetically assisted stamping, where a press does the bulk draw and a pulse coil calibrates the sharp radii afterwards — pushed by Sergey Golovashchenko during his years in Ford Research on aluminium body panels.

Magnetic Pulse Welding: The Same Physics, Turned Up

Increase the velocity and add a small oblique approach angle, and forming turns into magnetic pulse welding (MPW) — a genuine solid-state weld with no bulk melting. A flyer tube is accelerated across a standoff of about 1–3 mm and strikes the inner part at 200–500 m/s with a collision angle of roughly 5–20°. At the sweeping collision line the transient pressure reaches gigapascals, the metals momentarily behave hydrodynamically, and a thin jet squirts out ahead of the contact point, scrubbing away oxide films and adsorbed contamination. Two atomically clean surfaces are then pressed together hard enough to bond metallurgically, leaving the characteristic wavy interface.

That is word-for-word the physics of explosive welding — MPW is the same process family with a capacitor bank standing in for the explosive charge, which makes it clean, repeatable, and factory-safe. Like EXW it excels at pairs fusion welding cannot handle: aluminium to copper for busbars and battery terminals, aluminium to steel for automotive tube joints, aluminium to magnesium and titanium. Because interfacial heating is intensely localised and lasts only microseconds — any melt is confined to isolated pockets in the wave vortices — the brittle intermetallics (Fe–Al, Cu–Al) that ruin fusion joints never grow thick.

How It Fails, and Who Builds It

The dominant limitation is the coil, and the reason is Newton's third law. The magnetic pressure that pushes the workpiece away pushes the coil back with exactly the same force. A compression coil therefore sees hundreds of MPa of outward hoop pressure on every shot, as a microsecond hammer blow, thousands of times a day, while absorbing Joule heating from currents in its own skin depth. Coils are built from hardened conductors — beryllium copper, Cu–Cr–Zr — reinforced with steel or composite jackets and epoxy-encapsulated, and are still wear parts. Life ranges from a few thousand shots at near-gigapascal pressures to well over a million at modest ones, which is why production cells put a sacrificial field shaper in the line of fire. Turn-to-turn arcing at 10–25 kV is the other common coil death.

On the part side, too little energy gives an incomplete crimp; too much tears the workpiece or lets a tube wrinkle and buckle before it collapses uniformly. Pressure falls off at coil ends, so short coils crimp non-uniformly, and a workpiece that cannot carry a closed current loop — a slit tube, a slotted blank — simply refuses to form. The process is also inefficient: typically only 5–20 % of bank energy becomes plastic work, the rest lost to coil heating and ringing.

Commercially the technology dates to late-1950s work at General Dynamics' General Atomic division in San Diego, the founding patent granted to Harvey and Brower in 1961, and the Magneform machine line that grew out of it, later built and sold by Maxwell Laboratories. Today the field is served by Bmax in France (grown from Pulsar Ltd's Israeli technology), PST Products and Poynting in Germany, plus research at Fraunhofer IWU and Ohio State's Impulse Manufacturing Laboratory. Its everyday footprint is larger than its fame: crimped driveshafts, oil and fuel filter seams, shock-absorber and gas-strut closures, assembled camshafts, aluminium space-frame joints, and aluminium-to-copper transitions in electrical gear.

Electromagnetic forming against the conventional ways of shaping sheet and tube

Property	Electromagnetic forming	Mechanical press / stamping	Hydroforming
How force is applied	Body force — Lorentz J×B inside the metal, no tool touches it	A punch and die squeeze the surface	Pressurised fluid on one face, rigid die on the other
Event duration	~20–100 µs, single shot	~0.1–1 s per stroke	Seconds to tens of seconds
Strain rate	10³–10⁴ s⁻¹	~1–10 s⁻¹	~0.1–1 s⁻¹
Springback	Largely eliminated by high-speed die impact	Significant — needs overbend and compensation	Reduced but still present
Material limits	Needs a good conductor (Al, Cu); steel needs a driver ring	Any formable metal	Any formable metal
Tooling	One die plus a coil that fatigues; no lubricant	Matched punch and die, lubricant, guided ram	One die plus seals and a pressure intensifier