Mechanical Power: Water, Wind, and Gearing

Difficulty: moderate — it's carpentry and geometry, not chemistry. Prerequisites: woodworking tools, ideally iron for bearings, gudgeons, and axles.

Why it matters

Every other guide in this repo is bottlenecked by labor. Grinding pozzolan for concrete, crushing ore for iron, beating pulp for paper, milling cane for sugar, spinning thread for cloth, pumping the bellows for every furnace — all of it is someone turning a crank until their arms give out.

A single medieval watermill did the grain-grinding work of roughly 30 to 60 people. It never got tired, never ate, and ran day and night. Mechanical power is the multiplier that makes everything else on this list economically possible rather than merely technically possible.

Illustrative power comparison at suitable sites (actual output can be much lower or higher):

Water power

Choosing a site

You need net head (usable vertical drop after waterway losses) and flow through the wheel. A useful estimate is

P (kW) ≈ 9.81 × Q × H × η

where Q is flow in cubic metres per second, H is net head in metres, and η is the combined efficiency of wheel and drive. Measure both head and usable flow rather than choosing from the table above, and measure flow in dry and wet seasons before sizing the works. A small stream with a large drop can outperform a broad but slow river.

Look for: a stream you can dam or divert, a natural fall or steep reach, a bank you can cut a leat (mill race) into, and a location that won't flood out the wheel pit. A mill pond upstream is enormously valuable — it stores water so you can run the mill hard for part of the day even on a modest stream.

Wheel types

Undershot — water flows past the bottom of the wheel and pushes paddles. It can work with low head or a sufficiently fast current, drawing mainly on the water's kinetic energy, but it does not obtain useful power from a motionless flat reach. Traditional examples are often in the ~20–30% efficiency range, but output and efficiency are highly site- and design-dependent. It is simple to build.

Breastshot — water enters at axle height into curved buckets. Needs a few metres of head. Efficiency ~50–60%.

Overshot — water is carried by a flume over the top of the wheel and falls into buckets; the weight of the water turns it. Needs head at least equal to the wheel diameter. Efficiency 60–85% — the best of the traditional wheels, and the difference from undershot is dramatic. Note it turns the opposite direction to what people expect.

Horizontal ("Norse" or "Greek") mill — a horizontal wheel on a vertical shaft driving the millstone directly with no gearing at all. Crude and inefficient, but you can build one in a weekend with an axe. If you need something running tomorrow, build this.

Water turbine — the modern descendant. A properly designed reaction turbine in a casing hits 90%+ and is far more compact. Achievable with good iron casting and careful blade geometry, and worth aiming at eventually.

Construction points that decide success

• Bearings are the whole game. A wooden axle in a wooden socket wears out and wastes power in friction. Use iron gudgeons (axle ends) turning in replaceable bearing blocks: lignum vitae or another hard bearing wood, bronze where available, or suitably dressed stone in some installations. Align them carefully and lubricate them as their material requires. Bad bearings are a common reason a homebuilt mill underperforms.
• Seal the buckets. Overshot buckets that leak spill water before it does work. Fit them tightly and shape them so water stays in until the bottom.
• Control the water. A sluice gate on the leat lets you start, stop, and regulate the mill. Without one you cannot stop it, which is dangerous.
• Build the wheel true. An out-of-balance wheel destroys its own bearings.

Wind power

Use where there's no fall of water — flat coastal plains, dry country.

• Post mill — the entire mill body pivots on a massive vertical post so the sails can be turned into the wind. Medieval European standard.
• Tower mill — a fixed masonry tower with only the cap and sails rotating. Bigger, sturdier, more expensive.
• Vertical-axis (Persian) mill — walled panels shield half the rotor so the wind pushes only one side. Doesn't need turning into the wind at all. Used for grinding and pumping in Iran from at least the 9th century. Much easier to build, considerably less efficient.

Sail design matters more than size. Flat boards catch wind badly; a sail with twist — steeper angle near the hub, shallower at the tip — works far better, because the tip is moving much faster through the air than the root. This is the same principle as a propeller or a modern turbine blade, and getting it right is a large free improvement over a naive flat sail.

Add a fantail (a small rotor at right angles to the main sails, geared to turn the cap) and the mill turns itself into the wind automatically. Elegant, and a real quality-of-life upgrade.

Transmitting and transforming power

This is where the leverage really is — the same wheel can drive completely different machines.

Gearing

• Lantern pinion and cogwheel — the classic wooden gear pair: a wheel with pegged wooden teeth driving a "lantern" of parallel rods between two discs. Tolerant of imprecision, easy to make, easy to repair (replace one tooth, not the wheel). Use hardwood teeth in a softer wheel so the cheap part wears.
• Gear ratio = tooth count ratio. Trading speed for torque, or the reverse, is the fundamental operation. A slow, powerful waterwheel geared up drives a fast millstone; geared down it drives a heavy trip hammer.
• Bevel gears turn the drive through 90°. A conventional vertical waterwheel rotates a horizontal axle; bevel or crown gearing can drive the vertical spindle about which a horizontal millstone rotates. Essential and buildable in wood.
• Make tooth counts relatively prime (e.g. 31 and 8, not 32 and 8) so the same teeth don't repeatedly meet — wear distributes evenly. A cheap trick with real effect on service life.

Converting rotation to other motions

The cam is the most useful single mechanism here. Pegs on a rotating shaft lift heavy hammers and let them fall. That one arrangement gives you: forge hammers for iron, ore stamps for silver and copper, fulling stocks for cloth, and pulp stampers for paper. Build one shaft, retool the hammers, and you've mechanized four industries.

The flywheel is the most underrated. A heavy rotating mass carries a crank through its dead points and smooths pulsed power into steady power. Without one, reciprocating machines stall.

The machines worth building first

Ranked by how much labor they free up:

1. Grain mill. Grinding grain by hand quern is hours of daily work per household, almost always done by women. Mechanizing it returns an enormous number of person-hours to a community immediately, and it's the easiest mill to build. Start here.
2. Water-powered bellows. Your furnaces are currently limited by how long a person can pump. Mechanical bellows run all day and enabled the European blast furnace. This directly unlocks better iron.
3. Trip hammer. Forging by hand limits you to what a smith's arm can move. A water-powered hammer works blooms and forges bar stock at a scale hand forging cannot touch.
4. Sawmill. A crank-driven reciprocating saw turns logs into planks without pit-sawing, which is brutal, slow work.
5. Stamp mill. Crushing ore for any metal, and grinding pozzolan for concrete.
6. Pump. Draining mines (which is what limits mine depth), irrigation, water supply. An Archimedes screw lifts large volumes a short distance; a chain pump or piston pump lifts smaller volumes much higher.
7. Fulling stocks and, eventually, spinning machinery. See textiles.

Simple machines and the ideas behind them

Worth stating explicitly, because they're the conceptual toolkit:

• Lever, pulley, wedge, screw, wheel and axle, inclined plane — all trade force for distance. You never get more work out than you put in; you get it in a more useful form.
• Block and tackle — each additional supporting rope divides the force needed. Four ropes, quarter the effort, four times the pulling distance.
• Treadwheel crane — people walking inside a large wheel lift enormous weights. This is how cathedrals were built.
• Windlass and capstan — rotary force multiplication for hauling.
• Gear trains for measurement, not just power — the Antikythera mechanism (~100 BCE) was a geared astronomical calculator. Gears compute ratios; that's the seed of clockwork and of mechanical calculation.

What about steam?

Achievable but demanding, and it's the wrong first target.

A Newcomen atmospheric engine (1712) is genuinely within reach of good pre-industrial technology: a large cylinder, a piston, a boiler, and condensing steam with a water spray to pull a vacuum that drives the piston down. It's extremely fuel-hungry, but low boiler pressure and an atmospheric power stroke make it more tolerant of leakage than later efficient engines. It still needs an adequately round, straight large cylinder, a workable piston seal, sound valves, and safe boiler construction. It became economically useful especially at coal mines where low-cost fuel and drainage value outweighed poor fuel economy.

Watt's separate condenser (conceived in 1765) greatly reduces repeated heating and cooling of the power cylinder by condensing the steam in a separate vessel. It can cut fuel consumption sharply, but the gain is not one universal multiplier, and implementation adds a condenser, air pump, valves, piping, maintenance, and workmanship demands.

Accurate cylinder boring is a major scaling barrier. Newcomen engines already worked with rougher bores and forgiving seals, but leakage and irregular geometry limited size and efficiency. Wilkinson's 1770s boring machine supplied much more accurate cylinders to Boulton and Watt, enabling larger and more economical engines. If you want steam, invest early in an accurate lathe, gauges, and a boring mill; they are as important as the thermodynamic design. See precision and toolmaking, which covers how to develop reference geometry and calibrated measurement.

Do water power first where the site supports it. It is simpler, needs no fuel, and will often beat an early steam engine on operating cost and maintenance; seasonality, location, and storage still matter.

Fuller treatment — atmospheric engines, the separate condenser, the centrifugal governor, and the boiler-safety discipline that any pressurised vessel demands — is in steam and heat engines.

Failure modes

Safety

Powered machinery in a pre-industrial setting is genuinely dangerous. Fit fixed guards over accessible shafts, gears, belts, couplings, and flywheels, and build a sluice gate and brake that can be reached quickly. Those controls stop a machine; they do not make maintenance safe. Before clearing a jam, positively isolate and secure the water or wind input, wait for all motion to stop, chock or release stored energy, and verify that nobody can restart it. Keep long hair and loose sleeves away from shafts. Millstones running dry can throw sparks, and suspended grain dust can explode: control leaks, remove settled dust without raising a cloud, exclude flames and sparks, and ventilate dusty enclosures.

Sources and limits

• The U.S. Department of Energy's guide to assessing hydropower resources derives output from flow, net head, and efficiency. Wheel ranges in this guide are orientation figures, not substitutes for a site survey.
• The Science Museum Group documents Watt's separate-condenser model and John Wilkinson's precision-boring work.
• OSHA's guidance on hazardous-energy control distinguishes an emergency stop from positive isolation before servicing; powered shafts, gears, belts, and flywheels also require guarding.
• Power, efficiency, and fuel savings depend on water conditions, scale, load, leakage, workmanship, and transmission losses; treat historical percentages as ranges rather than machine guarantees.

Era notes

Waterwheels appear around the 1st century BCE in the Greek and Roman world; the Barbegal complex in Roman Gaul ran 16 overshot wheels together. Windmills come from Persia around the 9th century. The medieval period saw water power applied to fulling, sawing, forging, and ore crushing — the Domesday Book records over 6,000 mills in England alone.

Your knowledge advantages: matching wheel type to measured flow and head, using an overshot design where sufficient head permits, twisted windmill sails, relatively prime gear ratios, the flywheel, proper bearings, and — most importantly — the habit of asking "what else can this shaft drive?" about every mill you build.