An arastra is a primitive ore mill that grinds gold or silver ore by dragging heavy stones in a circular path across a stone-paved basin. Spanish miners introduced the design to the Americas in the 1500s, and it spread through Mexico and the California and Nevada gold fields by the 1850s. A central vertical post drives one or more drag stones, powered by mules, water, or later small steam engines, pulverising ore against the bedded floor. The result is a flour-fine charge ready for mercury amalgamation, recovering free gold at small scales without the capital cost of a stamp mill.
Arastra Ore Mill Interactive Calculator
Vary sweep radius and RPM to see the drag stone track speed, lap time, and grinding distance.
Equation Used
The calculator uses the circular track speed of the drag stone: multiply the sweep circumference by revolutions per minute, then divide by 60 to convert to metres per second.
- Sweep radius is measured from the central post to the drag stone contact path.
- RPM is steady and the stone follows a circular track.
- Linear speed is the contact speed of the muller across the basin floor.
How the Arastra Ore Mill Works
The arastra is brutally simple. You build a shallow circular basin — typically 3 to 4.5 m across — paved with flat, hard bedrock or fitted granite blocks. A vertical post sits dead-centre, with horizontal sweep arms extending outward. From those arms hang the drag stones, usually 100 to 300 kg each, suspended on chains so they can ride up and over harder lumps of ore instead of jamming. A mule, water wheel, or small steam engine rotates the post, the stones drag across the bedded ore and water slurry, and the ore crushes by abrasion and impact between two stone surfaces.
Why build it this way? Because in 1550s Mexico or 1860s Arizona, you could build one with hand tools, local stone, and a mule. No castings, no precision parts. The chain suspension matters — if you bolt the drag stones rigidly to the sweep arm, a single quartz boulder will snap the arm or stall the mule. The chain lets the stone lift, ride over the obstruction, and drop back into contact. Get the chain length wrong — too short and the stone won't ride up; too long and the stone bounces and chips the basin floor — and you'll spend more time repairing the mill than crushing ore.
Failure modes are predictable. The basin floor wears unevenly, dishing out near the outer track where stone velocity is highest. Drag stones fracture along bedding planes if you use sedimentary rock instead of dense igneous stone — the muller stone needs to be granite, basalt, or porphyry for any reasonable life. The central post wears at the bearing and starts to wobble, which throws the sweep radius off and leaves an uncrushed inner ring. Most surviving arastras you see today in the Sierra Nevada or Sonora show all three failure patterns at once.
Key Components
- Stone-paved basin: The fixed grinding floor, 3 to 4.5 m diameter, paved with flat granite or fitted bedrock slabs. The pavement must be tight — gaps wider than 5 mm let amalgamated mercury and fine gold escape into the subsoil, which is why surviving arastra sites are often re-worked by modern prospectors.
- Drag stones (mullers): Heavy abrasive stones, 100 to 300 kg each, dragged in a circular path. Dense igneous rock — granite, basalt, or trachyte — gives 6 to 12 months of working life. Sedimentary mullers fracture within weeks.
- Central vertical post: The drive shaft, typically a hardwood timber 200 to 300 mm in diameter, set in a stone or iron footstep bearing. Wear at the footstep is the main mechanical failure point — once the post wobbles by more than ~20 mm at the sweep tip, grinding becomes uneven.
- Sweep arms: Horizontal timbers extending from the post, usually 2 to 3 per mill. They carry the chain attachment for each muller and also the harness point for the mule or the connection to a water-wheel crank.
- Suspension chains: Connect the drag stones to the sweep arms with 200 to 400 mm of slack. The slack lets the muller climb over hard lumps without jamming or shock-loading the sweep arm.
- Power source: Mule, ox, water wheel, or small steam engine, delivering 0.5 to 3 kW at 2 to 8 RPM at the sweep. Mule-powered arastras typically run at 3 to 5 RPM — faster than that and the animal won't sustain it for a 10-hour shift.
Industries That Rely on the Arastra Ore Mill
The arastra dominated small-scale precious-metal milling for almost 400 years because it solved one specific problem: how do you process a few tonnes of high-grade ore per week, in a remote location, with no industrial supply chain? Anywhere a prospector hit a vein but couldn't justify a 10-stamp mill, an arastra got built. They appear across the entire Spanish colonial mining belt and throughout the 19th-century American West, and a surprising number still see use today in artisanal mining operations in Latin America.
- Spanish Colonial Mining: Silver ore processing at the Real del Monte mines near Pachuca, Mexico, from the 1550s onwards — the arrastra (Spanish spelling) was the standard pre-amalgamation grinder before larger Chili mills displaced it.
- California Gold Rush: Hundreds of arastras built across the Sierra Nevada foothills between 1849 and 1870, including documented sites along the Mokelumne and Tuolumne rivers, processing quartz vein ore from small claims.
- Arizona Territory Mining: Arastras at the Vulture Mine and across the Bradshaw Mountains in the 1860s-1880s, often water-powered where seasonal creeks allowed, used to bulk-grind ore before shipping concentrates out.
- Modern Artisanal Mining: Small-scale gold operations in Colombia, Peru, and the Philippines still use motorised arastras (often called 'Chilean mills' locally even when geometry is true arastra) for amalgamation circuits processing 1-5 tonnes per day.
- Heritage and Living History: Operating reproduction arastras at Columbia State Historic Park in California and Tombstone Courthouse State Historic Park in Arizona, run as demonstration mills for visitors.
- Tailings Reprocessing: Modern small operators in Nevada and Sonora reworking historical arastra sites — the gappy stone basins lost 10-30% of free gold to the subsoil, making the floor itself an ore body.
The Formula Behind the Arastra Ore Mill
The single number that determines whether your arastra grinds ore or just polishes it is the linear velocity of the drag stone across the basin floor. That velocity sets the abrasion rate. Run too slow at the low end of typical operating range — 2 RPM with a 2 m sweep — and you grind 0.4 m/s, barely enough to break quartz. Run nominal at 4 RPM you sit at 0.8 m/s, the historical sweet spot where mules sustain the load and ore reduces from 25 mm down to flour size in 6-8 hours. Push to the high end at 8 RPM and you theoretically hit 1.7 m/s, but the muller starts skipping and chipping the basin floor instead of grinding, and no mule will hold that pace. The formula below gives you the tip speed for any sweep radius and rotation rate.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| vtip | Linear velocity of the drag stone at the sweep radius | m/s | ft/s |
| Rsweep | Radius from the central post to the centre of the drag stone | m | ft |
| N | Rotation rate of the central post | RPM | RPM |
| π | Mathematical constant (≈3.14159) | dimensionless | dimensionless |
Worked Example: Arastra Ore Mill in a mule-powered arastra at a small Sonora gold claim
A small operator in Sonora is rebuilding a historical arastra to process about 2 tonnes per day of quartz-hosted gold ore. The basin is 3.6 m in diameter, the drag stones ride at a sweep radius of 1.5 m, and a single mule turns the central post. The operator wants to know the muller tip speed at three operating points: a slow walk at 2 RPM, the working pace at 4 RPM, and a forced trot at 8 RPM, to decide whether to run one mule or rig a two-mule sweep arm.
Given
- Rsweep = 1.5 m
- Nnominal = 4 RPM
- Nlow = 2 RPM
- Nhigh = 8 RPM
Solution
Step 1 — at the nominal mule walking pace of 4 RPM, compute tip speed:
This is the historical sweet spot. At 0.628 m/s the muller drags hard enough to fracture quartz, but slowly enough that it stays in continuous contact with the floor and doesn't chip the basin pavement. A 250 kg muller at this speed will reduce 25 mm ore to flour-fine in roughly 6 to 8 hours of continuous milling.
Step 2 — at the low end of the typical operating range, 2 RPM (a tired mule near end-of-shift):
At 0.314 m/s grinding nearly stops. The muller has enough velocity to polish the ore but not enough kinetic energy to break the quartz crystal lattice on impact with each lump. You'll see the charge stay coarse for hours and the operator will wonder why the mill 'isn't working' — it's working, just at maybe 30% of nominal throughput.
Step 3 — at the high end, a forced trot at 8 RPM:
On paper this doubles throughput. In practice it doesn't work. Above roughly 1.0 m/s the chain-suspended muller starts bouncing rather than dragging, which chips the basin floor and lets fine gold escape into the cracks. No single mule will hold 8 RPM for more than a few minutes — this is why historical accounts describe two-mule or water-powered sweeps once operators tried to push past 5 RPM.
Result
Nominal muller tip speed at 4 RPM and 1. 5 m sweep radius is 0.628 m/s. That's the speed at which the arastra actually grinds — fast enough to fracture quartz, slow enough that the muller stays in contact with the floor. The low end at 2 RPM gives 0.314 m/s where grinding effectively stalls, and the high end at 8 RPM reaches 1.257 m/s in theory but causes the muller to skip and chip the basin instead of grind. If your measured throughput is well below the predicted rate at nominal RPM, check three things: (1) chain slack on the mullers — chains shorter than 200 mm prevent the stone from riding over hard lumps, jamming progress; (2) basin floor flatness — a dished outer track means the muller only contacts ore on the inner edge, halving effective grinding area; (3) muller density — if someone substituted sandstone for granite, the stones are abrading away faster than the ore is and the charge looks like sand-stone slurry rather than ore flour.
Arastra Ore Mill vs Alternatives
The arastra competed historically with two main alternatives: the stamp mill (the dominant industrial choice from 1860 onwards) and the Chilean mill (the rolling-stone variant). Each makes different tradeoffs on capital cost, throughput, ore size handled, and labour intensity. The right choice depends entirely on tonnage and capital available.
| Property | Arastra | Stamp Mill | Chilean Mill (Trapiche) |
|---|---|---|---|
| Throughput (tonnes/day) | 1-5 | 20-200+ | 5-30 |
| Capital cost (relative) | Very low — built from local stone and timber | High — heavy castings and steam plant | Medium — iron rolling stones and frame |
| Operating speed | 2-8 RPM | 60-100 drops/min per stamp | 10-20 RPM |
| Maximum ore feed size | ~50 mm (hand-broken first) | ~150 mm directly from crusher | ~75 mm |
| Power source | Mule, water wheel, or small steam (0.5-3 kW) | Steam engine or large water wheel (20-200 kW) | Steam, water, or motor (5-30 kW) |
| Typical lifespan of grinding surface | 6-12 months for muller, years for basin | Stamp shoes: 3-6 months | Rolling stones: 1-2 years |
| Best application fit | Small claims, remote sites, artisanal scale | Industrial mining operations | Mid-scale operations or wet grinding for amalgamation |
| Gold loss to floor/escape | 10-30% historically (gappy basins) | <2% with mercury plates | 5-15% |
Frequently Asked Questions About Arastra Ore Mill
The drag stones only cover an annulus — they ride at the sweep radius, not at the centre. If you have one muller at 1.5 m and the basin is 1.8 m in radius, you grind a band roughly 300-400 mm wide centred on 1.5 m. Anything inside about 1.2 m radius and outside about 1.7 m radius gets minimal contact.
Two fixes work. Either rake the charge inward every hour or two during the run so fresh ore moves into the muller's track, or hang two mullers at different radii (say 1.0 m and 1.6 m) so their tracks overlap. Most well-built historical arastras used two or three mullers at staggered radii precisely for this reason.
Not really, and this is the single biggest selection mistake. The arastra works on quartz and quartz-hosted ore because quartz fractures readily under dragging abrasion at moderate velocities. Sulfide ore is tougher and tends to smear rather than fracture, and dense mafic rock like gabbro will outlast your muller stones. You'll see the muller wear down faster than the charge reduces.
For anything harder than quartz, jump to a stamp mill or a ball mill. The arastra's capital advantage disappears the moment you have to replace mullers monthly.
The deciding factor is feed top size and water availability. If your ore comes in already broken to under 25 mm and you have water for slurry, an arastra is cheaper to build and easier to repair — you can replace a cracked muller in an afternoon. If your feed is 50-75 mm or you want continuous feed through a hopper, the Chilean mill (rolling iron stones rather than dragged stones) handles the larger feed and runs faster.
At 3 tonnes/day you're right at the crossover. If you have iron-working capability locally, build the Chilean. If you only have stone, timber, and a welder, build the arastra.
This is almost always one of two things, and neither is the mercury chemistry. First check the muller chain length. As chains stretch and pin holes elongate over months of operation, the muller starts to bounce slightly with each rotation, which throws fines and mercury droplets out of the basin onto the surrounding apron. You're losing gold-laden mercury before it can settle.
Second, check the basin pavement joints. If a slab has shifted by even 3-5 mm, mercury runs into that gap and stays there. Lift suspect slabs and you'll often find a silver bead of amalgam underneath. Re-bed the slabs with clay-mortar and recovery typically returns to baseline within one or two charges.
The footstep bearing — where the bottom of the post sits in a cup of stone or iron at the basin floor — takes the entire vertical load of the post, sweep arms, and harness, plus side load from the mule pulling. With wood on stone, you get steady wear at maybe 1-2 mm per week of operation. Iron-on-iron lasts longer but the wear pattern is the same.
Once tip wobble at the muller exceeds about 20 mm of lateral runout, the muller traces an oval rather than a circle and the basin starts dishing unevenly. Re-cup the footstep, replace the post bottom, or fit an iron pivot pin. Don't ignore wobble — every millimetre of lateral runout magnifies into uneven floor wear that you'll spend weeks dressing later.
The animal sets the limit. A working mule sustains roughly 1 m/s at the harness point on a multi-hour shift. The harness sits at maybe 2.0-2.5 m from the post on a typical sweep. Solve backwards: at 4 RPM the harness moves at 0.84-1.05 m/s, which is exactly where mules historically settled.
For the muller itself, you want sweep radius between 1.2 and 1.7 m. Smaller and you don't get enough tip speed at sustainable RPM. Larger and the geometry pushes the harness too far out, the mule walks too fast, and you tire the animal in three hours instead of eight. The historical convergence on roughly 3-4 m basin diameter wasn't aesthetic — it's the optimisation of mule biomechanics against muller velocity.
It can be, but not by running the arastra again. The economic ore at a historical site is the basin floor itself and the surrounding apron — those gappy stone joints captured 10-30% of the original gold as amalgam over decades of operation. Strip the basin slabs, screen and pan the underlying clay, and you often recover more gold per cubic metre than the original ore body produced.
Running the arastra itself for new ore rarely pencils out at modern labour rates. A small motorised hammer mill or a used ball mill handles the same throughput at lower labour cost. The historical site's value is archaeological and as a tailings resource, not as a production mill.
References & Further Reading
- Wikipedia contributors. Arrastra. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
