A Magnetic Ore Separator is a mineral processing machine that pulls ferromagnetic and paramagnetic particles out of a flowing ore stream using a controlled magnetic field. The core component is a rotating drum or roll housing a stationary magnet assembly — that fixed magnet creates a field gradient at the drum surface so magnetic particles cling to the shell while gangue falls clear. The purpose is to upgrade run-of-mine ore before downstream grinding or flotation. A single LIMS wet drum can produce a magnetite concentrate at 65% Fe from a feed grade of 30%.
Magnetic Ore Separator Interactive Calculator
Vary field strength, drum gap, slurry solids, and feed velocity to see the magnetic force index and operating margins.
Equation Used
The article notes that magnetic pull scales with B x dB/dx, not field strength alone. This calculator converts shell field from gauss to tesla, approximates the gradient as field divided by the drum gap, and reports a relative force index for comparing separator settings.
- Field gradient is approximated across the trough-to-drum gap.
- Particle susceptibility, size, and volume are held constant.
- The result is a relative force index, not an absolute particle force.
- Velocity margin is compared with the article limit of 0.3 m/s.
The Magnetic Ore Separator in Action
A Magnetic Ore Separator works by exploiting the difference in magnetic susceptibility between valuable minerals and gangue. Feed slurry or dry powder runs across the surface of a rotating drum, belt, or roll. Inside the rotating shell sits a fixed array of permanent magnets — usually ceramic ferrite for low-intensity duty or NdFeB rare-earth blocks for high-intensity work. Magnetite, ilmenite, and tramp iron stick to the shell, ride around the magnetic arc, and drop into the concentrate launder once they pass beyond the field. Non-magnetic gangue obeys gravity and discharges into the tails launder.
The geometry of the magnet assembly matters more than raw field strength. You want a high field gradient — that is, a fast change in field strength over a short distance — because the force on a paramagnetic particle scales with B × dB/dx, not with B alone. A low intensity magnetic separator (LIMS) wet drum running 600-1000 gauss at the shell surface will recover liberated magnetite cleanly, but it will not touch hematite. For hematite or ilmenite you need a wet high-intensity magnetic separator (WHIMS) running 8000-15000 gauss with a matrix of fine steel rods or expanded mesh that sharpens the gradient locally.
If the gap between the drum shell and the trough bottom is wrong — too tight and you grind the shell, too wide and the field at the bottom of the slurry layer is too weak to grab fines — you lose recovery fast. Typical wet drum gap is 3-6 mm. If you measure recovery dropping after a few months, the usual culprits are demagnetised ferrite blocks behind a worn shell (check with a gauss meter at the shell surface — should read within 10% of factory spec), a slurry density drift outside the 30-45% solids window, or feed velocity above 0.3 m/s tearing magnetic particles off before they reach the discharge arc.
Key Components
- Rotating Drum or Shell: A non-magnetic stainless steel cylinder, typically 600-1500 mm diameter, rotates at 15-35 RPM around a fixed internal magnet assembly. Shell thickness sits at 3-5 mm — go thicker and the field at the surface drops, go thinner and abrasive feed wears through inside 6 months.
- Stationary Magnet Assembly: An arc of ferrite or rare-earth NdFeB blocks fixed to a stationary shaft inside the drum. The arc covers 110-130° of the drum circumference, defining where particles pickup and where they release. Field strength at the shell ranges from 600 gauss for LIMS cobber duty up to 15000 gauss for WHIMS finishing.
- Feed Box and Trough: Distributes slurry evenly across the drum width at 30-45% solids. Trough-to-drum gap holds at 3-6 mm. Uneven feed distribution causes localised overload and a streaky concentrate.
- Concentrate and Tailings Launders: Two parallel collection chutes positioned at the magnetic discharge arc and the gravity discharge point. Splitter blade adjustment is critical — 5 mm of misalignment moves middlings into the wrong stream and drops grade by 1-2 percentage points.
- Drive and Bearings: Geared motor delivers 15-35 RPM at the drum shaft. Pillow block bearings carry the radial load from slurry hydraulics plus the magnetic pull on tramp iron. Bearings are sealed against slurry ingress — failed seals are the number one cause of drum replacement on wet duty.
Real-World Applications of the Magnetic Ore Separator
Magnetic Ore Separators show up anywhere a ferromagnetic or paramagnetic mineral needs to be split from non-magnetic gangue, or anywhere tramp iron threatens downstream equipment. The same basic mechanism — drum, roll, or belt with a fixed magnet inside — scales from small lab cobbers to 3 m wide production drums handling 500 t/h. Industry naming varies: iron ore plants call them Magnetic Ore Separators, mineral sands operators call them rare-earth roll separators, and recyclers call them overband magnets, but the physics is identical.
- Iron Ore Beneficiation: LKAB's Kiruna concentrator in Sweden runs banks of LIMS wet drums to upgrade magnetite from a 45% Fe feed to a 68% Fe pellet feed.
- Mineral Sands: Iluka Resources' Jacinth-Ambrosia operation in South Australia uses rare-earth roll Magnetic Ore Separators to split ilmenite and monazite from zircon and rutile in heavy mineral concentrate.
- Coal Preparation: Dense medium circuits at Peabody's Australian operations use LIMS drums to recover magnetite from the dense medium slurry, returning it to the cyclone feed at 99.5%+ recovery.
- Foundry and Scrap Recycling: Eriez overband suspended magnets pull tramp iron out of bauxite and limestone feed on conveyor belts ahead of crushers, protecting cone liners from catastrophic damage.
- Tungsten and Tin Concentrators: Wolframite and cassiterite plants in Portugal run induced roll separators after gravity tables to clean residual iron silicates out of the final concentrate.
- Industrial Minerals: Imerys kaolin operations use WHIMS to remove iron-bearing impurities from clay slurry, hitting the brightness specs needed for paper coating.
The Formula Behind the Magnetic Ore Separator
The working number for any Magnetic Ore Separator is the magnetic force per unit volume acting on a particle at the shell surface. This is what determines whether a given mineral will be picked up and held against the centrifugal force of the rotating drum. At the low end of the operating range — say 600 gauss with a shallow gradient — you'll recover liberated magnetite but lose finely disseminated material. In the nominal LIMS range around 1000-1200 gauss the force is high enough to pull magnetite cleanly while letting silica gangue fall away. Push into WHIMS territory at 10000+ gauss and you start grabbing weakly paramagnetic minerals like hematite and ilmenite, but you also start dragging silicates and bearing iron into the concentrate, which kills selectivity.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Fm | Magnetic force on the particle | N | lbf |
| χ | Volumetric magnetic susceptibility of the particle (dimensionless) | — | — |
| V | Particle volume | m³ | in³ |
| B | Magnetic flux density at the shell surface | T (tesla) | gauss (1 T = 10000 G) |
| dB/dx | Magnetic field gradient normal to the shell | T/m | G/in |
| μ0 | Permeability of free space | 4π × 10-7 T·m/A | — |
Worked Example: Magnetic Ore Separator in a taconite cobber drum
A Mesabi Range taconite operation in northern Minnesota is sizing a LIMS wet drum cobber to handle 250 t/h of ground magnetite ore at a P80 of 250 µm. The team needs to confirm the drum will hold a 100 µm magnetite particle against centrifugal force at the discharge arc, and they want to understand how the picking force changes if shell wear pushes the surface field from the design 1200 gauss down toward 800 gauss, or if a future rare-earth retrofit lifts it to 3500 gauss.
Given
- χ (magnetite) = 1.0 SI volume susceptibility (approx, near saturation)
- Particle diameter = 100 µm
- Bnom = 0.12 T (1200 gauss)
- dB/dx = 8 T/m
- Drum diameter = 1.2 m
- Drum speed = 25 RPM
Solution
Step 1 — compute the particle volume for a 100 µm sphere:
Step 2 — at the nominal 1200 gauss (0.12 T) shell field, compute the magnetic force:
Step 3 — check this against the centrifugal force at the discharge arc, with ω = 2π × 25 / 60 = 2.62 rad/s and r = 0.6 m:
The magnetic force is roughly 36× the centrifugal force at nominal field. The particle stays glued to the shell through the pickup arc with comfortable margin.
Step 4 — at the low end, after shell wear and partial demagnetisation drop B to 0.08 T (800 gauss), force scales linearly with B:
Still well above centrifugal demand for a 100 µm particle, but recovery on 30 µm fines (which scale with particle volume, so 27× lower force) collapses to roughly 1.0 × 10-8 N — the same order as centrifugal. That's where you start seeing fines report to tails.
Step 5 — at the high end, a rare-earth retrofit at 0.35 T (3500 gauss) with a sharper 25 T/m gradient gives:
Nine times the nominal force. The drum will now grab weakly magnetic middlings — hematite-bearing composites and iron-stained silicates — which lifts recovery but drops concentrate grade. That's why rare-earth retrofits are usually paired with a second cleaner stage at lower field to reject the middlings.
Result
At nominal 1200 gauss the drum exerts 4. 0 × 10-7 N on a 100 µm magnetite particle, comfortably 36× greater than the centrifugal force trying to fling it off. At 800 gauss (worn shell, demagnetised blocks) the force drops to 2.7 × 10-7 N — fine for coarse magnetite but inadequate for the sub-30 µm fraction, which is why aged drums lose 2-4 points of iron recovery before the operator notices anything in the concentrate assay. At 3500 gauss with a rare-earth retrofit the force jumps to 3.6 × 10-6 N, dragging middlings into the concentrate and dropping grade. If your measured recovery sits below prediction, check three things in this order: (1) feed pulp density drifting above 50% solids, which thickens the slurry and physically blocks fines from migrating to the shell; (2) a misaligned tailings splitter dumping recoverable concentrate over the wrong lip — a 3 mm offset on a 1.5 m drum costs about 1% recovery; (3) magnetised tramp build-up bridging across the drum-trough gap, which short-circuits the field locally.
Magnetic Ore Separator vs Alternatives
Magnetic Ore Separators come in several variants — wet drum LIMS, dry drum, induced roll, rare-earth roll, and WHIMS — and the choice depends on mineral susceptibility, particle size, and whether the feed is wet or dry. Magnetic Ore Separators compete on real engineering dimensions: field strength, throughput per metre of drum width, capital cost, and how finely they discriminate between magnetic susceptibilities. Below is a comparison against the two most common alternatives a beneficiation engineer evaluates.
| Property | LIMS Wet Drum Magnetic Ore Separator | WHIMS (High Intensity) | Induced Roll Separator (Dry) |
|---|---|---|---|
| Field strength at working surface | 600-1500 gauss | 8000-20000 gauss | 12000-22000 gauss |
| Suitable mineral susceptibility | Strongly ferromagnetic (magnetite, pyrrhotite) | Paramagnetic (hematite, ilmenite, wolframite) | Paramagnetic, dry feed only |
| Throughput per metre drum width | 80-150 t/h | 20-50 t/h | 1-5 t/h |
| Particle size range | 10 µm to 6 mm slurry | 10-500 µm slurry | 75 µm to 2 mm dry |
| Capital cost (relative) | 1.0× (baseline) | 3-5× | 2-3× |
| Maintenance interval (shell wear) | 12-24 months | 6-12 months (matrix replacement) | 24-36 months |
| Best application fit | Magnetite cobbing, dense medium recovery | Hematite finishing, kaolin purification | Mineral sands, tungsten cleaning |
Frequently Asked Questions About Magnetic Ore Separator
The shell is wearing thinner on the leading edge of the magnetic arc, which actually pulls the field closer to the surface and increases pickup of weakly magnetic middlings — silicate composites with locked magnetite inclusions. Recovery looks good because everything sticks, but the concentrate is now carrying gangue that used to fall off.
Check shell thickness at the 10 o'clock position with an ultrasonic gauge. If you've lost more than 1 mm from the original 4-5 mm, the drum is past its grade-optimal life even if it's still mechanically sound. Most operators replace shells on grade slip, not on wear-through.
Two stages almost always win on metallurgy. A single drum has one chance to grab a particle and one splitter to make the concentrate-tails decision. Two drums in rougher-cleaner configuration let the first drum run aggressively for recovery and the second run conservatively for grade, with the cleaner tails recycling back to rougher feed.
The capital premium is roughly 60-70% over a single drum of equivalent throughput, but you typically gain 3-5 points of concentrate grade and 1-2 points of recovery. On a 250 t/h taconite circuit that pays back inside a year on smelter penalty avoidance alone.
Two effects compound. Slurry viscosity drops with temperature, which sounds helpful but actually shortens the residence time in the matrix — particles wash through before the field can deflect them. At the same time, paramagnetic susceptibility of hematite has a slight inverse temperature dependence, so the magnetic force per particle weakens by a few percent over a 20°C swing.
If you see seasonal recovery drift, check feed temperature first and either chill the feed or drop the feed rate by 10-15% during warm months. Operators in the Pilbara routinely de-rate WHIMS circuits in summer for exactly this reason.
Mechanically yes, metallurgically often no. NdFeB blocks deliver 3-4× the field at the shell, but the existing splitter geometry, drum speed, and trough gap were all designed around the ferrite field profile. Drop rare-earth blocks in without redesigning the discharge arc and you'll pull middlings into concentrate and grade will collapse by 4-8 points.
If you commit to the retrofit, plan on adjusting drum RPM up by 30-50% to throw off the middlings via centrifugal force, and expect to recalibrate the splitter weekly for the first month while the circuit settles.
Surface moisture on dry feed creates particle agglomerates — fines stick to coarse particles and travel as a clump. The clump's bulk magnetic susceptibility is dominated by the largest particle in it, so non-magnetic fines piggyback into the concentrate and magnetic fines get carried into tails attached to gangue particles.
The fix is feed conditioning, not the separator itself. Most operations either run an upstream fluid-bed dryer to hold feed moisture below 0.5%, or accept seasonal de-rating. If you're seeing a 5%+ recovery swing between dry and humid days, moisture is your variable.
The practical floor for a LIMS wet drum is around 10-15 µm for liberated magnetite. Below that, magnetic force scales with particle volume (cubed with diameter) while drag force scales with surface area (squared with diameter), so drag wins and particles wash through to tails regardless of field strength.
For ultrafines below 10 µm you need either WHIMS with a fine matrix or a magnetic flocculation circuit ahead of the separator, where small magnetic particles are first agglomerated under a low field into pickable flocs. Edison's original Ogden ore-milling work in the 1890s ran into exactly this lower-size limit and it remains the binding constraint on magnetite recovery today.
References & Further Reading
- Wikipedia contributors. Magnetic separation. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
