Magnetorheological Damper

A liquid that turns into a tunable solid

A magnetorheological damper looks, from the outside, like an ordinary monotube shock absorber: a sealed cylinder with a piston rod sticking out one end. The difference is what is inside. Instead of plain mineral oil, the cylinder is filled with a magnetorheological (MR) fluid — a slurry of micron-scale soft-iron particles suspended in oil. And instead of fixed orifices in the piston, the piston contains a small electromagnet. Drive a few amps through the coil and the fluid changes state in milliseconds: from a free-flowing liquid into a yield-stress solid that can withstand tens of kilopascals of shear before it begins to flow. Cut the current and it relaxes back to a liquid. The damping force the shock applies to its piston rod is therefore not a fixed property of the hardware. It is something the on-board controller can dial up or down, ten times a second, on a scale roughly ten-to-one between its softest and stiffest settings.

That single trick — turning a fluid's effective viscosity into a controllable parameter — collapses a whole class of mechanical-design problems. Suspensions no longer have to compromise between sporty handling and comfortable ride; prosthetic knees no longer have to pick a single damping curve and hope it suits every gait phase; tall buildings no longer have to over-design columns for the worst-case earthquake. The damper itself adapts.

What MR fluid actually is

A commercial MR fluid is, by volume, roughly:

• 20-40% iron carbonyl particles, 1-10 μm diameter, soft-magnetic and nearly spherical (carbonyl iron powder is produced by thermal decomposition of iron pentacarbonyl Fe(CO)₅ and is the standard MR particle).
• 60-80% carrier oil: most often a low-viscosity hydrocarbon or synthetic oil, occasionally silicone or water for special temperature or biocompatibility requirements.
• A few percent surfactants and stabilisers: oleic acid, lithium stearate, fumed silica, organoclays — preventing the heavy iron particles (density 7.86 g/cm³) from settling out of the much lighter oil over months on a shelf.

In the absence of a magnetic field, the suspension behaves like a slightly thick motor oil — the off-state viscosity is roughly 0.2 to 0.3 Pa·s, so its flow resistance contributes a small velocity-proportional damping force on its own. The particles drift independently and undergo Brownian motion; the slurry has essentially no yield stress.

Apply a uniform field H of a few hundred kiloamperes per metre and the picture changes completely. Each iron particle is magnetised and becomes a magnetic dipole aligned with H. Adjacent dipoles attract head-to-tail and repel side-to-side; the system finds its lowest energy by assembling particles into chains, then bundles of chains, that span the gap between the damper's flow channels. To make the fluid flow now, those chains have to be broken — and that requires a finite shear stress, the field-dependent yield stress τ_y(H). For carbonyl iron in oil this saturates near τ_y ≈ 60-100 kPa once the particles themselves are magnetically saturated.

The Bingham-plastic model

The textbook rheological model for MR fluid is the Bingham plastic:

τ(γ̇, H) = τ_y(H) · sgn(γ̇) + η · γ̇         (|τ| > τ_y)
γ̇        = 0                                (|τ| ≤ τ_y)

Below the yield stress, the fluid does not flow at all — it behaves as an elastic solid (with a finite, very stiff pre-yield modulus we usually ignore at the damper-design level). Above τ_y, it flows like a Newtonian fluid with viscosity η on top of a constant offset τ_y. The yield stress is set by the field; the post-yield viscosity is set by the carrier oil. This separation is the conceptual key: the damper force is the sum of a controllable yield component (whose value the magnet sets) plus a passive viscous component.

For a flow-mode damper (oil pushed axially through the field gap by the piston), the resulting force on the piston rod can be approximated as

F(v, I) = F_visc(v) + F_y(I) · sgn(v)
F_visc  = c · v                         passive viscous, ~ piston velocity
F_y     ∝ τ_y(H(I)) · A_eff / g         field-controlled Coulomb-like

where g is the annular gap, A_eff the effective active area, and H(I) the field produced by current I in the coil. The two components do different jobs. The yield component dominates at low piston velocities — it sets the static stiction the damper resists. The viscous component dominates at high velocity — it controls peak forces during sharp impacts. Modern designs tune both by choosing fluid composition, gap geometry, and coil power.

Inside the damper: piston, coil, gap

A typical MR damper is a monotube cylinder. The piston rod terminates in a piston head about 30-60 mm in diameter, which contains:

• A copper coil of a few hundred turns wound around a central iron core, generating an axial-then-radial magnetic field when energised.
• An annular flow gap between the coil bobbin and the cylinder wall — typically 0.5 to 2 mm wide — through which the MR fluid must squeeze as the piston moves.
• A gas-charged accumulator (usually nitrogen behind a floating piston at one end) to accommodate the volume change as the rod enters and exits the cylinder.
• Lead wires routed through the hollow piston rod up to an external controller.

The magnetic circuit is closed mostly through the piston-iron, jumps the annular gap radially through the MR fluid, returns through the cylinder wall (also made of magnetic steel), and short-circuits across the steel end of the piston. The fluid in the gap is therefore the magnetic "load" of the circuit; that is where the field does work to align chains and where the controlled yield stress appears. With a current of 1-3 amps the gap field reaches a few hundred kA/m and the yield-stress component saturates near 100 kPa, multiplied across the gap geometry to give a piston-rod force of order kilonewtons.

Why MR is fast

The single most underrated property of MR damping is its response time. The fluid itself responds to a step change in field in much less than a millisecond — the chains assemble or disassemble on a timescale governed by particle drag in the oil, which is microseconds for micron particles in light oil. The damper's overall response is limited not by the fluid but by the L/R time constant of the coil and the bandwidth of the current-driver electronics. Together these are typically under 5 ms in commercial automotive units.

Compare:

Damper type	Adjustability	Response time	Bandwidth
Passive hydraulic	None (fixed valving)	—	—
Adjustable passive (clickers)	Manual, between rides	Minutes	0 Hz
Active hydraulic (servo valve)	Continuous	~50 ms	~5 Hz
Magnetorheological	Continuous	<5 ms	50-200 Hz
Electrorheological	Continuous	<5 ms	50-200 Hz

5 ms is fast enough to act on a single road bump as the wheel sees it — wheel-hop frequencies are 10-20 Hz, so a controller updating every 1-2 ms has loads of margin. For a passenger car body bouncing at 1 Hz on its primary suspension, MR is basically instantaneous. For an active hydraulic system at 50 ms, the wheel has already moved through most of its travel before the valve has fully opened.

Closed-loop control

Because the damper's force is electronically programmable, an MR suspension is a closed-loop control problem. The canonical strategy is the skyhook controller (Karnopp, 1974): conceptually, the body is attached to an inertial "sky" by an ideal damper, and the controller picks a command that emulates that virtual damper using the real damper-to-wheel forces it has available. In practice, modern automotive implementations layer several strategies on top of each other:

• Skyhook / ground-hook blends, weighted by piston-velocity sign and accelerometer signals on the body and each wheel.
• Feedforward from steering angle, brake, and throttle inputs that anticipate weight transfer.
• Pre-emptive maps indexed by driving mode ("Tour" / "Sport" / "Track") which set base damping curves and weighting between front/rear.
• Road-surface estimation from wheel-speed sensors and inertial measurements that detects a rough surface and softens the rear for ride comfort.

Sample rates of 1 kHz are routine; the damper is updated faster than any mechanical mode the suspension can excite.

Worked example: what force does the damper deliver?

Take a typical automotive MR damper with the following geometry:

Active length (axial) L     = 30 mm
Annular gap width g          = 1.0 mm
Mean gap radius R            = 15 mm
Effective active area A_eff  = 2π R L          = 2.83 × 10⁻³ m²
Carrier viscosity η          = 0.25 Pa·s
On-state yield stress τ_y    = 60 kPa

For a piston velocity v of 0.5 m/s (a brisk bump), the volumetric flow rate is Q = A_p · v where A_p ≈ 1 × 10⁻³ m² is the piston cross-section, giving Q ≈ 5 × 10⁻⁴ m³/s. The mean fluid velocity in the gap is u = Q / (2π R g) ≈ 5.3 m/s, and the shear rate γ̇ = u / g ≈ 5300 s⁻¹.

The Bingham post-yield viscous pressure drop is

ΔP_visc = 12 η L u / g²
        = 12 × 0.25 × 0.030 × 5.3 / (10⁻³)²
        ≈ 480 kPa

The yield component (Coulomb-like in the duct flow approximation) is

ΔP_y ≈ (2 L / g) · τ_y
     = (2 × 0.030 / 10⁻³) × 60 × 10³
     = 3.6 × 10⁶ Pa = 3.6 MPa

Multiplying by piston area gives an on-state damping force of roughly (3.6 + 0.48) × 10⁶ Pa × 10⁻³ m² ≈ 4 kN; off-state (τ_y → 0) gives 0.48 kN. The 10:1 dynamic range and the dominance of the yield term over the viscous term at moderate velocities both emerge from the geometry. The actual MagneRide automotive damper produces 1-7 kN over its operating range, in line with this estimate.

Where MR dampers are used

• Automotive semi-active suspension — Cadillac MagneRide. Introduced on the 2003 Cadillac Seville STS and 2003 Chevrolet Corvette (C5 Z06), MagneRide replaces a conventional shock with an MR damper at each corner, polled at 1 kHz by a body-control module. After GM's lead, the technology was licensed to Audi (R8), Ferrari (599 onward), Aston Martin, and many other premium platforms. The dampers themselves are produced by BWI (formerly Delphi).
• Civil engineering — large-scale MR seismic dampers. The largest deployed MR damper to date is a 350-ton unit on Tokyo's Dictionary Hall, installed for seismic protection. MR dampers are also used in cable-stayed bridge cable dampers (Dongting Lake Bridge, China) and base-isolation systems for hospitals and high-tech fabs. The civil-scale dampers are physically enormous — piston rods of several centimetres' diameter — but the underlying fluid, coil, and Bingham model are identical to a car shock.
• Prosthetics — Össur Rheo Knee. The Rheo Knee uses a compact MR damper in the knee joint, adapting its damping every step based on accelerometer and load-cell readings. Swing-phase damping is dialled low for free leg motion; stance-phase damping is dialled up to support body weight; rapid response (under 10 ms) lets it recognise a stumble and lock before the patient falls.
• Industrial vibration isolation. MR dampers are used to balance high-spin washing machines, isolate gun recoils (M101 howitzer), damp helicopter rotor lag, and absorb shock in heavy off-highway vehicle seats.
• Adaptive resistance machines. High-end rehabilitation and fitness equipment uses MR dampers to provide a velocity-independent "constant load" that a hydraulic damper cannot match.

Aside: the 350-ton seismic damper

Why does the largest MR damper in the world live in a Tokyo library? Because Japan is the world's premier laboratory for earthquake engineering and the building (the Nihon University Dictionary Hall, or "Tokyo Dictionary Hall" in some sources) was designed around a hybrid passive-plus-MR seismic isolation system. A 350-ton hydraulic damper would have to commit to one set of valves at design time, optimised for the most likely earthquake; an MR damper can adjust its yield stress in real time as the controller infers earthquake severity from the first half-second of strong motion. Smaller MR dampers are used as cable dampers on cable-stayed bridges, where wind-induced rain-on-cable vibrations need a damping curve that depends on weather.

MR versus other controllable dampers

The two main competitors for an MR damper are electrorheological (ER) fluid and active hydraulic (servo-valve) systems.

Technology	Mechanism	Yield stress	Drive	Response	Issues
MR fluid	Magnetic field aligns iron particles	60-100 kPa	1-3 A, 12 V (10-50 W)	<5 ms	Particle settling, sealing rod
ER fluid	Electric field polarises dielectric particles	~5 kPa	3-5 kV (low current)	<5 ms	High voltage, low force, contamination sensitive
Active hydraulic	Servo valve gates pressurised oil	—	Hydraulic pump (kW)	~50 ms	Complex hoses, leaks, energy intensive
Passive hydraulic	Fixed orifices	—	None	—	No adaptation

ER fluid was studied first — Winslow described the effect in 1949 — but its yield stress is at least an order of magnitude lower than MR, and the kilovolt drive voltages are difficult to package safely in a car or a building. Active hydraulic systems can in principle apply any force, but they need a continuously running pump, complex hose routing, and tens of milliseconds to slew valves; they were abandoned for high-end automotive suspensions in the 1990s when MR became commercially available. The fundamental MR-fluid invention is usually attributed to Jacob Rabinow (1948) at the U.S. National Bureau of Standards; modern commercialisation began with Lord Corporation in the early 1990s, who licensed key patents and were first to ship MR fluid in volume.

What goes wrong

The two recurring engineering headaches with MR fluid are particle settling and thermal/abrasive wear. Iron is far denser than oil, so a static damper left for weeks will see iron settle to the bottom of the cylinder. Modern designs combat this with thixotropic additives that build a weak gel in the fluid at rest, plus a small return flow during normal piston cycling that re-suspends particles. Long-term settling is the main reason MR dampers are sometimes specified with sealed-for-life service intervals: opening the unit and stirring the fluid is rarely necessary, but the design assumes the vehicle moves regularly. Secondary issues include in-use abrasion (hard iron particles wearing seals and piston bores — countered by hard-coated piston surfaces and hardened seal materials), and elevated viscosity at low temperatures (the off-state penalty grows in cold climates).

One unusual virtue: an MR damper has only one moving fluid path, no valves, and no high-pressure hoses. If the coil shorts or the wiring is cut, the damper degrades gracefully to a passive damper with the off-state viscous force — never to nothing. That fail-safe behaviour is a deliberate design property and a major reason MR has been certified for safety-critical aerospace and prosthetic uses where active hydraulics would be a regulatory non-starter.

Variants and modes

• Flow mode (valve mode). Fluid is pumped through a fixed annular gap by piston motion. The dominant mode for damping applications — including all automotive and most seismic units. Force scales with τ_y(H).
• Shear mode. Fluid is sheared between two parallel surfaces moving relative to each other. Used in MR brakes and clutches, where rotational motion needs a controllable torque.
• Squeeze-film mode. Fluid is compressed normal to the field. Used in precision vibration isolation and small actuator applications where strokes are short and stiffness is high.
• Pinch mode. A non-uniform field causes the fluid to thicken near field maxima, blocking small flow channels. Used in microfluidic and micro-valve applications.

A practical damper can combine modes — e.g. a flow-mode primary damping path plus a shear-mode rod-seal region that uses the off-state low viscosity for low friction.

Common pitfalls and misconceptions

• "MR fluid is a magnetic fluid (ferrofluid)." No. Ferrofluids contain 10 nm magnetic nanoparticles and remain liquid even under strong fields — they only change density. MR fluids use 1-10 μm particles, three orders of magnitude larger, and develop a true yield stress. The two are easily confused but solve different problems.
• "MR dampers consume a lot of power." They don't. A passenger-car MR damper draws 10-50 W steady state per corner — comparable to a single LED headlamp — and zero when the controller decides the off-state is appropriate. Active hydraulic systems can pull several kilowatts continuously.
• "You can drive the damper harder to get more force." Only up to a point. Once the iron particles are magnetically saturated, τ_y stops rising regardless of coil current. Designers pick gap geometry so that magnetic saturation is reached at the rated coil current; pumping more current just heats the coil.
• "It works the same upside down." Not for long. Settling is gravity-driven, and a damper installed inverted that sees long static intervals will accumulate iron at the new "bottom" (which used to be the top). Production units specify an installation orientation.
• "Off-state means zero force." No: even at I = 0, the carrier oil produces a velocity-proportional viscous force. That force is the lower bound of the damper's controllable range and is typically 10% of the on-state peak.