A gold separator is a gravity-concentration device that splits free gold from lighter host rock and sand by exploiting the density difference between gold (SG 15-19) and quartz or silicate gangue (SG 2.6-2.7). It solves the core problem of recovering fine and coarse free gold from a high-volume slurry without chemicals. The device uses water, riffles, pulsation, or centrifugal force to settle dense particles while flushing lights to tailings. A modern Knelson CVD or Falcon SB concentrator routinely recovers 95%+ of free gold above 30 µm.
Gold Separator Interactive Calculator
Vary heavy mineral, gangue, and fluid specific gravity to see the concentration criterion and riffle-pocket separation behavior.
Equation Used
The concentration criterion compares how strongly the heavy mineral separates from the fluid versus the light gangue. For the worked example, gold SG 19.3, quartz SG 2.65, and water SG 1.0 give CC = (19.3 - 1.0) / (2.65 - 1.0) = 11.1, which is well above the easy-separation threshold of 2.5.
- Specific gravity values are used because density ratios cancel in the formula.
- The light mineral SG is greater than the fluid SG.
- CC > 2.5 indicates easy gravity separation.
Operating Principle of the Gold Separator
Every gold separator runs on the same physics — Stokes' settling combined with hindered settling in a fluidised bed. Gold is roughly 7 times denser than quartz, so under the right slurry conditions it sinks faster, sticks in a riffle pocket, or migrates to a high-G concentrate ring while the gangue washes off. The mechanism is purely mechanical specific-gravity separation, and that's exactly why it's still the front end of almost every gold flowsheet — it pulls free gold out cheaply before cyanide or flotation has to deal with it.
The details depend on the machine. A sluice box uses gravity flow over riffles, where eddies behind each riffle bar create a low-velocity zone that traps heavies. A jig concentrator pulses water vertically through a screen bed so the bed dilates on the up-stroke and compacts on the down-stroke, letting dense particles work their way to the bottom. A centrifugal concentrator like the Knelson or Falcon spins the slurry at 60-200 G, packing gold into fluidised retention rings while water and lights overflow the lip. A shaking table combines an asymmetric reciprocating motion with a riffled deck and a cross-flow of wash water, fanning particles into density-banded streams.
Tolerances matter more than people expect. If your sluice slope is 1.25 in/ft instead of 1 in/ft, the flow velocity climbs and fine gold below 100 mesh blows straight off the deck. If a Knelson's fluidising water pressure drops below the spec range (typically 3-5 psi for a KC-MD3), the concentrate ring packs solid and starts losing gold to the tails. Riffle wear, deck levelness within 1 mm, and feed solids percentage (target 30-45% by weight for most centrifugal units) are the three things that quietly destroy recovery if you stop watching them.
Key Components
- Riffles or capture surface: Physical traps that create low-velocity zones where dense particles settle out of the flow. On a sluice, riffle height is typically 12-25 mm with spacing equal to 2-3× the height. Worn riffles below 8 mm height lose roughly half their fine-gold catch.
- Fluidisation system: On centrifugal and jig concentrators, water injected through a perforated cone or screen keeps the bed mobile so dense particles can migrate through it. Fluidising water on a Knelson KC-MD3 runs 3-5 psi at 30-40 L/min — drop below 25 L/min and the bed compacts.
- Drive or motion source: Provides the energy that creates separation. A shaking table uses an eccentric head delivering 250-300 strokes/min at a 10-25 mm stroke length. A centrifugal concentrator uses a VFD-driven motor maintaining 60-200 G at the bowl wall, depending on duty.
- Feed distributor: Delivers a uniform slurry across the working surface. Uneven feed is the single most common cause of poor recovery — a sluice fed off-centre will channel and lose gold from the starved side. Target feed solids 30-45% for most gravity units.
- Concentrate discharge: Removes the heavy fraction without disturbing the separation. On a batch Knelson it's a periodic flush cycle every 30-120 minutes; on a continuous unit like a Falcon C or a Knelson CVD, valves bleed concentrate out under pressure during operation.
- Tailings launder: Carries the lights away. The launder must handle the full volumetric throughput plus a safety margin — undersized tailings discharge backs up the bed and kills recovery instantly.
Where the Gold Separator Is Used
Gold separators show up everywhere gold is produced, from artisanal panners in Ghana to 30,000 tpd hard-rock mills in Nevada. The choice of separator depends on feed grain size, throughput, gold liberation, and how clean a final concentrate you need. Coarse free gold above 1 mm is easy — almost anything catches it. Fine gold below 75 µm is where machine choice starts to matter, and below 25 µm gravity concentration alone usually can't compete with cyanidation or flotation.
- Hard-rock gold mining: Knelson KC-XD48 centrifugal concentrators installed in the gravity circuit at Newmont's Boddington mine in Western Australia, recovering coarse free gold from grinding circuit cyclone underflow ahead of cyanide leach.
- Placer mining: Trommel-fed sluice boxes on bucket-line dredges working the Yukon River drainage near Dawson City, processing gravel at 50-150 yd³/hour with riffle-and-matting recovery.
- Artisanal and small-scale mining: Manual rocker boxes and pan-and-puddle setups used across the Madre de Dios region of Peru for alluvial gold, often the first step before a Cleangoldm sluice mat.
- Tailings reprocessing: Falcon SB concentrators retrofitted at the Mponeng mine in South Africa to recover residual gold from old flotation tailings before disposal.
- Gold dredging: Gold Cube and Le Trap-style finishing sluices used downstream of suction-dredge primary recovery on the Klamath River in northern California.
- Lab and exploration: Wilfley shaking tables in assay labs running gravity recoverable gold (GRG) tests per the Knelson/Steve McAlister GRG protocol on drill-core composites.
The Formula Behind the Gold Separator
Recovery and concentrate-ratio prediction starts with the concentration criterion (CC) — a single number telling you whether gravity separation will work at all on a given mineral pair, and how hard it'll be. At the low end of the typical range (CC near 1.25), separation is borderline and you'll need a finely-tuned shaking table with skilled operators. Around CC = 2.5 (the sweet spot for gold-quartz) almost any gravity device works well. Above CC = 4 the separation is so easy a kid with a pan can do it. The formula is your first sanity check before you ever buy equipment.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| CC | Concentration criterion, dimensionless. >2.5 is easy, 1.75-2.5 is workable, <1.25 is impractical. | dimensionless | dimensionless |
| ρh | Density of the heavy mineral (gold) | kg/m³ | lb/ft³ |
| ρl | Density of the light mineral (gangue, typically quartz) | kg/m³ | lb/ft³ |
| ρf | Density of the fluid medium (water) | kg/m³ | lb/ft³ |
Worked Example: Gold Separator in a Yukon placer wash plant
An operator on Hunker Creek near Dawson City is sizing a 100 yd³/hour wash plant feeding a Gold Cube finishing stage off a primary 6-foot sluice. Feed is alluvial gravel with quartz-dominant gangue, water from the creek at roughly 4°C, and target gold grain size from 50 µm flood gold up to 2 mm pickers. The crew wants to confirm gravity separation is the right call before committing to the build, and to understand what happens to recovery at the fine end of the size range.
Given
- ρh = 19300 kg/m³ (native gold, allowing for some silver alloy bringing it down from 19320 pure)
- ρl = 2650 kg/m³ (quartz)
- ρf = 1000 kg/m³ (water)
- Feed rate = 100 yd³/hr
- Sluice slope = 1 inch per foot
Solution
Step 1 — at the nominal case, plug gold-against-quartz-in-water into the concentration criterion:
That's a huge number. Anything above 2.5 is easy gravity work, and 11.1 means the density contrast is so favourable a sluice with proper riffles will catch nearly everything coarse and middling. This is why placer gold has been recovered with pans and rockers since the 1850s — the physics is on your side.
Step 2 — at the low end of the practical recovery range, look at very fine gold near 50 µm. Stokes' settling velocity for a 50 µm gold particle in 4°C water is roughly:
16 mm/s is slow. On a sluice running 0.6-0.9 m/s surface velocity, a 50 µm flake has under a second to settle into a riffle pocket before it's swept off the deck. Field recovery on flood gold below 100 mesh on a standard expanded-metal sluice typically falls to 40-60%, which is why the crew added the Gold Cube finishing stage.
Step 3 — at the high end, a 2 mm picker:
A picker drops out of suspension almost the instant it hits the head of the sluice. Recovery on coarse gold above 1 mm sits at 95-99% on a properly slope-matched sluice. The sweet spot for the primary sluice is gold between 200 µm and 1 mm — fast enough to settle, big enough to lodge behind a riffle.
Result
Concentration criterion is 11. 1 — gravity separation is unambiguously the right tool for this feed. In practice that means the operator should expect 95%+ recovery on coarse and medium gold from the primary sluice, dropping to 40-60% on flood gold below 100 mesh, with the Gold Cube finishing stage clawing back another 20-30% of the fines that the primary loses. The sweet spot for the whole plant sits in the 200 µm to 1 mm size band where both stages run near-perfect. If the operator measures recovery 15-20% below predicted, the three failure modes to check first are: (1) sluice slope drift — a deck that's settled to 1.5 in/ft over a season runs water too fast and blows fines off, (2) packed riffles where heavy black sand has filled the eddies behind each bar and the riffles are now hydraulically smooth, killing the catch zone, and (3) feed-end overload above 120 yd³/hr where bed depth exceeds riffle height and gold rolls right over the top.
Gold Separator vs Alternatives
There's no universal best gold separator — the right machine depends on feed size, throughput, capital budget, and whether you're chasing nuggets or 30 µm flood gold. Here's how the main families compare on the dimensions that actually matter when you're picking equipment.
| Property | Centrifugal concentrator (Knelson/Falcon) | Sluice box | Shaking table |
|---|---|---|---|
| Throughput per unit | 1-1000 tph dry feed (KC-MD3 to KC-XD70) | 5-200 yd³/hr depending on width | 0.5-2 tph per deck |
| Recovery on −75 µm gold | 85-95% | 30-60% | 70-85% |
| Recovery on +1 mm gold | 95-99% | 95-99% | 90-95% |
| Concentrate ratio (mass reduction) | 1000:1 to 2000:1 | 20:1 to 100:1 | 50:1 to 500:1 |
| Capital cost | $30k-$500k+ | $500-$50k | $15k-$80k per deck |
| Operator skill required | Moderate — VFD and fluidising water tuning | Low — clean-up and slope checks | High — deck tilt and stroke tuning by feel |
| Water consumption | 20-100 L/min fluidising | High — full slurry flow 500-2000 L/min | Moderate — 30-80 L/min wash water |
| Best application fit | Hard-rock grinding circuits, fine gold | Placer and alluvial, coarse to medium gold | Final cleanup, lab GRG testing, small batches |
Frequently Asked Questions About Gold Separator
The classic cause is a compacted bed — fluidising water dropped below spec or the cone got plugged by oversize tramp material. When the retention zone packs solid, incoming gold can't displace the trapped pyrite and magnetite already in the ring, so it skates over the top and out to tails.
Check fluidising water pressure first (3-5 psi at the manifold for a KC-MD3, more for larger units), then pull the cone and inspect for plugged fluidising holes. A 5% drop in fluidising flow can cost you 20% of your recovery on fine gold. Screen the feed at 6 mm minimum if you're not already.
Yes, but you'll be working for it. CC between 1.75 and 2.5 means separation is possible only with careful feed preparation and a fine-tuned shaking table or centrifugal unit — a sluice will struggle. The pyrite (SG 5.0) is competing with the gold for every riffle pocket, so you'll get a high-mass concentrate that still needs cleaning.
The standard fix is to run a primary gravity stage to reject quartz, then upgrade the gravity concentrate on a Wilfley table or a magnetic separator to drop pyrrhotite and magnetite. Don't expect a single-stage finished product when CC is below 2.5.
For raw placer feed at that scale, a sluice is almost always the right primary — it's cheap, handles oversize and clay better, and gold above 200 µm is easy. A centrifugal unit only earns its keep if your gold is fine (below 100 µm) and you've already screened and deslimed the feed.
The common build is a trommel into a 4-6 ft primary sluice for the easy gold, then send the fines (−2 mm undersize) to a small centrifugal unit or Gold Cube to clean up the flood gold. Don't put raw gravel into a Knelson — you'll plug the cone and damage the bowl.
Three things to check in order: deck transverse tilt, stroke length, and wash water volume. If the bands are crowding toward the concentrate end, the deck slope is too flat or the stroke is too short — heavies aren't getting fanned out. If they're spreading too far toward the tailings end, the wash water is too high or the longitudinal tilt is steep.
The diagnostic trick is to add a pinch of red lead or magnetite tracer to the feed and watch where the band lands. Adjust stroke first (typical range 10-25 mm), then wash water, then tilt. Change one variable at a time or you'll chase your tail for an hour.
Surface velocity is too high for the fine particles to settle into the riffle eddies. A 100-mesh (150 µm) gold flake has a Stokes' settling velocity around 50 mm/s in cold water — if your sluice is running 0.8 m/s with a 1.5 in/ft slope, the flake travels 16 riffle-widths before it can drop. Coarse gold (1 mm+) settles at over 1 m/s and lodges instantly.
Drop your slope to 1 in/ft, add a Cleangoldm or miner's-moss matting section at the tail end, and consider a finishing stage like a Gold Cube fed off the sluice tailings. Don't try to catch flood gold on expanded-metal riffles alone — the fluid mechanics is against you.
Feed solids drives bed density inside the bowl. Below 20% solids the bed is too dilute and middlings (pyrite, magnetite) displace gold from the retention rings. Above 50% solids the slurry is too viscous to fluidise properly and you lose fines to bypass.
Target is 30-45% solids by weight for most concentrators. If you're feeding from a cyclone underflow that swings between 55% and 65% solids, dilute it down with a controlled water addition before the feed inlet — don't trust the cyclone to deliver a clean number. A simple density gauge on the feed line pays for itself in a month.
It varies massively. A coarse-gold ore like the Macassa mine in Ontario can run 70-90% GRG, meaning gravity alone can pull most of the value before cyanide ever sees it. A refractory or fine-gold ore like Carlin-trend material in Nevada might be under 10% GRG — gravity is barely worth installing.
The standard test is a 3-stage Knelson lab procedure (the Steve McAlister / John Laplante protocol) on a 25-50 kg drill-core composite, sequentially regrinding and concentrating to liberate locked gold. Run this before you size a gravity circuit — designing without a GRG number is how mills end up with $400k of equipment recovering 12% of the gold.
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
