A Klondike mining machine is a gravity-fed gold-recovery device — typically a sluice, rocker box, or steam-thawed shaft hoist — used by placer miners during the 1896-1899 Yukon gold rush to separate fine gold from frozen creek gravel. George Carmack's August 1896 strike on Rabbit Creek (renamed Bonanza Creek) triggered the build-out of these machines along Yukon tributaries. Water washes loosened paydirt over angled riffles where heavy gold settles and lighter sand flushes downstream. A well-tuned 12 ft sluice could clean 3-5 cubic yards a day per two-man crew.
Klondike Mining Machine Interactive Calculator
Vary paydirt feed rate, gold grade, and sluice recovery to see recovered gold, lost gold, and tailings grade.
Equation Used
The calculator multiplies paydirt throughput by gold grade to estimate gold entering the sluice, then applies the selected recovery efficiency to split that gold into recovered concentrate and gold lost to tailings.
- Paydirt grade is uniform through the feed.
- Recovery percent represents total riffle and mat capture efficiency.
- Gold is calculated directly in grams per cubic yard of paydirt.
- Water grade, riffle condition, and feed size effects are represented by the recovery input.
How the Klondike Mining Machine Actually Works
The Klondike problem was simple to describe and brutal to solve — gold was locked in gravel buried under 2 to 15 m of permanently frozen muck above bedrock. You could not just dig it out. Miners thawed the ground with wood fires (the burning ground method) or later with steam points driven by portable boilers, then hoisted thawed paydirt up a shaft with a hand windlass or horse whim and stockpiled it through winter. Spring melt gave them the water they needed to actually run a sluice or rocker box and recover the gold.
The sluice itself is a gravity concentration device. Water carries the paydirt down a wooden trough at roughly a 1-inch-per-foot grade — too shallow and the gravel packs and stalls, too steep and gold rolls right over the riffles and into the tailings. The riffles are transverse wood bars or stones spaced about 75-100 mm apart on the floor. Each riffle creates a low-pressure eddy on its downstream face. Heavy gold (specific gravity 19.3) drops into that eddy and stays put. Lighter quartz sand and gravel (SG 2.6) tumbles past. A burlap or coco-mat liner under the riffles catches fine flour gold the riffles miss.
If the grade is wrong or the water flow is wrong, you lose gold to the tailings — full stop. Old-timers checked their tails by panning a shovelful from the bottom of the sluice. Anything more than a couple of colours per pan meant retune the sluice or slow the feed. The rocker box was the small-operator version — a hand-cranked cradle that let a single miner work a few yards a day without a continuous water supply, which mattered on dry benches above the creek.
Key Components
- Sluice trough: A wooden box typically 12 ft long, 12-16 inches wide, and 8-10 inches deep, set on a 1-in-12 grade. The grade is the single most-tuned parameter on the machine — drop it below 0.5-in-12 and gravel packs the riffles within an hour.
- Riffles: Transverse bars on the trough floor, spaced 75-100 mm apart and standing 25-40 mm proud. They generate the eddies where gold settles. Hungarian-style angled riffles outperform straight wood bars on fine gold by roughly 15-20% in recovery.
- Burlap or coco-mat liner: Sits under the riffles to catch flour gold below 100 mesh that the riffles cannot trap on their own. Pulled and rinsed into a tub during clean-up, typically once or twice per season on a small claim.
- Rocker box: A 4 ft-long hand-cradled box with a top hopper screen, a canvas apron, and a short riffle section. Rocked at roughly 60 cycles per minute by one miner while another shovels paydirt and pours water. Throughput around 1 cubic yard per day.
- Windlass or horse whim: Hoists thawed paydirt up the shaft from bedrock. A two-man windlass handles a 50 lb bucket from depths up to about 12 m. Below that depth a horse whim or steam hoist takes over — a single horse whim could pull from 30 m at 0.3 m/s.
- Steam point and boiler: Replaced wood-fire thawing after about 1898. A 6 ft hollow iron rod driven into frozen ground while live steam at roughly 30-40 psi flows through it. Thaws a 1 m radius of ground per point in 24 hours, vs 3-4 days for a wood fire.
Where the Klondike Mining Machine Is Used
The Klondike machine is not one device — it's a kit of gravity and thermal tools that get used in sequence. Modern placer operators on the same Yukon and Alaska creeks still run direct descendants of these machines, because the underlying physics of separating SG 19.3 gold from SG 2.6 gravel hasn't changed. Below are real operations and historic sites where these machines did the work.
- Placer gold mining: Original Klondike claims along Bonanza, Eldorado, and Hunker Creeks near Dawson City — over 1,000 sluice boxes operating by 1898.
- Heritage and tourism: Dredge No. 4 National Historic Site on Bonanza Creek preserves the bucket-line dredge that replaced hand-sluicing after 1913.
- Modern small-scale placer: Operators on Eureka Creek near Chicken, Alaska still use sluice boxes essentially identical to 1898 designs for cleanup of hand-stacked tailings.
- Education and demonstration: MacBride Museum in Whitehorse runs working rocker box demonstrations using paydirt trucked in from active claims.
- Reality television production: Discovery's Gold Rush series films modern wash plants on Indian River and Quartz Creek that use the same sluice-and-riffle principle on a 100 yd³/hour scale.
- Hobby and prospecting clubs: GPAA (Gold Prospectors Association of America) outings use 6 ft hand-fed sluices and rocker boxes on permitted claims across the Yukon and BC.
The Formula Behind the Klondike Mining Machine
The number every Klondike operator cared about was daily yield in dollars — but that breaks down into a recovery equation you can actually tune. The recovery rate depends on three things: how much paydirt you can feed through, the grade of that paydirt in grams of gold per cubic yard, and the recovery efficiency of the sluice itself. At the low end of the practical operating range — a single rocker box at 1 yd³/day on lean 2 g/yd³ ground — you're scratching out a living. At the high end — a 12 ft sluice running 5 yd³/day on rich Eldorado-grade 30 g/yd³ paydirt — you're paying off the trip in a week. The sweet spot for a two-man hand operation sits around 3 yd³/day at 8-12 g/yd³, which is what most upper-Bonanza claims actually returned.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Yday | Daily gold yield recovered to the riffles | g/day | oz/day |
| Vday | Paydirt volume processed per day | m³/day | yd³/day |
| Gpaydirt | Gold grade of the paydirt | g/m³ | oz/yd³ |
| ηsluice | Sluice recovery efficiency (fraction of contained gold actually trapped) | dimensionless | dimensionless |
Worked Example: Klondike Mining Machine in a 12 ft hand-fed sluice on Bonanza Creek
A two-man crew is working a claim on upper Bonanza Creek in summer 1898. They have a 12 ft wooden sluice set at 1-in-12 grade with Hungarian riffles, fed by a small ditch off the creek. Test pans show paydirt grading around 10 g of gold per cubic yard. Sluice recovery on this gravel runs about 0.85 — the riffles miss some flour gold despite the burlap liner. They want to predict daily yield so they know whether to keep working this drift or move the operation downstream.
Given
- Vday = 3 yd³/day
- Gpaydirt = 10 g/yd³
- ηsluice = 0.85 dimensionless
Solution
Step 1 — at the nominal operating point of 3 yd³/day, 10 g/yd³, and 0.85 recovery, multiply the three terms:
Convert to troy ounces — 25.5 g ÷ 31.1 g/oz ≈ 0.82 oz/day. At the 1898 gold price of roughly $18.90/oz, that's about $15.50/day per crew, or near $7.75 per man. For context, a labourer in Seattle that summer made $2/day. The crew is making roughly four times outside wages — solid Klondike money but well short of the Eldorado bonanzas.
Step 2 — at the low end of the typical range, the same crew working leaner 4 g/yd³ ground at only 2 yd³/day with a poorly tuned sluice running 0.70 recovery:
That's 0.18 oz/day, about $3.40 — barely above outside wages once you back out food, supplies, and the 50% royalty the Crown took on rich ground. Most claims that ran below this number were abandoned by 1899.
Step 3 — at the high end of the practical range, a well-located Eldorado claim at 25 g/yd³ with a tuned sluice at 0.90 recovery and a hired-on third man pushing 4 yd³/day:
That's nearly 2.9 oz/day, around $55 — the kind of number that built mansions in Dawson City. The famous Berry claim on Eldorado reportedly produced over $130,000 in a single season at numbers in this range.
Result
The nominal yield is 25. 5 g/day, or about 0.82 oz/day. In practical terms that's a thumbnail-sized pile of dust and small nuggets at the end of a 12-hour shift, weighed on a brass balance and stored in a leather poke. The range from 5.6 g/day on lean ground up to 90 g/day on Eldorado paydirt explains why claim location mattered more than equipment quality during the rush — the sweet spot for sustainable hand operations sat around 20-30 g/day. If your measured yield comes in below predicted, the three most common causes on a hand sluice are: (1) sluice grade off — anything below 0.75-in-12 packs the riffles within hours and recovery collapses to under 0.6, (2) feed rate too high so a slug of unwashed gravel buries the upper riffles and gold rides over the top, and (3) flour gold loss because the burlap liner is saturated and needs rinsing — check the tailings with a pan and if you see more than 2-3 colours per pan you're losing fines.
When to Use a Klondike Mining Machine and When Not To
The Klondike kit gave a placer miner three machines that all do the same job — separate gold from gravel — but at very different scales and with very different water and labour requirements. Choosing between them came down to claim size, water availability, and crew count. Here's how the three classic machines compare on the dimensions that actually mattered to a 1898 operator and still matter today.
| Property | Sluice box (12 ft) | Rocker box | Long tom |
|---|---|---|---|
| Daily throughput | 3-5 yd³/day with two men | 1 yd³/day with one man | 4-6 yd³/day with three men |
| Water requirement | Continuous flow, ~50 gal/min | Bucket-fed, ~5 gal/min equivalent | Continuous flow, ~30 gal/min |
| Recovery efficiency on fine gold | 0.80-0.90 with burlap liner | 0.70-0.85 with apron and screen | 0.75-0.85 |
| Build cost (1898 Dawson prices) | ~$40 in lumber and nails | ~$15 in lumber, canvas, screen | ~$60 — larger and more iron |
| Best application fit | Established creek claim with ditch water | Dry bench, prospecting, lone miner | Mid-size claim with crew |
| Setup time | 1-2 days to grade and build | 2-3 hours, fully portable | 2-3 days, semi-permanent |
| Crew requirement | 2 men sustained | 1 man, intermittent | 3 men sustained |
Frequently Asked Questions About Klondike Mining Machine
Nine times out of ten the issue is feed inconsistency, not the sluice itself. If you shovel in a heavy slug of paydirt all at once, the upper riffles get buried in gravel and the eddy that traps gold collapses for several seconds — gold rides right over the buried section and out the bottom. A steady, metered feed of one shovel every 15-20 seconds keeps the riffles working as designed.
The other common cause is water flow that's too high for the gravel size you're running. If the water is moving fast enough to lift small pebbles off the riffle face, it's also lifting fine gold. Throttle the head gate until pebbles sit between the riffles instead of skipping over them.
Run a test pan from the deepest point you've reached and from the proposed new location. On Klondike-style creeks, gold concentrates in the lower 30-60 cm of gravel directly above bedrock — the pay streak. If your test pans from current depth show 3+ colours per pan and bedrock is within 2 m, keep digging. The grade typically doubles or triples in that last metre.
If you've hit bedrock and grade is still under 5 g/yd³, move. The pay streak meanders, and 50 m laterally on the same creek can mean a 5x change in grade. The original Carmack claim and the famously rich Berry claim were less than 1 km apart on the same creek.
The rocker's screen at the top rejects oversize, but its short riffle run — typically only 600-900 mm — gives gold less distance to settle than a 12 ft sluice. On coarse gold the difference is small, maybe 5%. On fine flour gold the rocker can lose 15-20% more than a properly tuned sluice with a burlap liner.
The fix is rocking cadence. Too fast and gold doesn't have time to drop behind the riffles between strokes — aim for around 60 cycles per minute, slow enough that you can see water clear the apron between each rock. Too slow and the canvas apron clogs and gold washes off the top.
Around 12 m. A two-man windlass with a 50 lb bucket can cycle roughly once every 90 seconds at that depth — about 30 buckets per hour, or 1.5 yd³/day of hoisted paydirt. Below 12 m the rope weight itself starts eating into useful payload and cycle time stretches past 2 minutes per bucket.
That's why every successful claim past about 10 m depth on Bonanza converted to a horse whim or steam hoist by 1898. A horse whim triples the hoist rate and a small steam hoist quintuples it — the capital cost paid back in 3-4 weeks of operation on rich ground.
Single test pans lie. Gold distribution in placer gravel is patchy — one pan can show 10 colours and the next pan from 50 cm away shows nothing. The standard practitioner's check is to pan 10 spaced samples across the working face, weigh the total, and divide by the total volume panned. Anything less than 10 samples and your error bar is wider than the grade itself.
If your sluice yield consistently runs 30%+ below your pan-based grade prediction, the test pans are biased high — likely you were panning visible-gold spots rather than representative gravel. Re-sample on a strict grid pattern, not by eye.
Wood fires thaw a roughly cone-shaped 1 m radius of ground per fire over 3-4 days, and they consume 1-2 cords of wood per cone. By 1898 the creeks within 10 km of Dawson were stripped of timber, and wood was selling for $40 a cord — uneconomic for thawing.
A steam point driven by a portable boiler thaws the same 1 m radius in 24 hours and runs on whatever fuel you have, including the lower-grade scrub the wood-fire era couldn't use. The throughput tripled and the fuel cost halved. Within two seasons every serious operation on Bonanza had a boiler.
References & Further Reading
- Wikipedia contributors. Klondike Gold Rush. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
