01
Measuring the site before designing anything
Two numbers determine everything, and they can both be measured with simple equipment in an afternoon.
Head is the vertical drop available between where you take the water and where you return it. Measure it with a level and a staff, or with a length of clear tube filled with water, working down the slope in stages. Head is what gives the water its energy, and a surprising amount is frequently available over a short distance.
Flow is the volume passing per unit of time. The bucket method works on a small stream: divert the whole flow into a container of known volume and time how long it takes to fill. On a larger one, measure the cross-section and the surface velocity by timing a float over a known distance, then apply a correction because water moves slower at the bed and banks than at the surface.
Measure both at the worst time of year, not the best. A wheel sized on spring flow stands idle through the dry season, and the useful figure is what's reliably there rather than what's occasionally there.
Then the arithmetic: available power rises with head multiplied by flow. That tells you whether the site will do what you want before you spend anything, and it frequently reveals that a modest continuous output is available where a large one is not.
02
Choosing the wheel for the site
The wheel type follows from the head, and picking the wrong one is the main reason installations disappoint.
An undershot wheel sits in the flow and is driven by the water pushing its paddles. It needs almost no head, works in a stream that simply runs past, and is the least efficient because it captures the water's motion instead of its weight.
An overshot wheel takes water onto the top and is driven mainly by the weight of water in its buckets as they descend. It needs head roughly equal to the diameter of the wheel, and it's much more efficient because gravity does the work over the full descent. Where the head exists, this is usually the answer.
A breastshot wheel takes water at about axle height and works between the two, using both impulse and weight. It suits medium heads and is the traditional compromise.
Turbines are the modern development, more efficient again and far more demanding to make. For a workshop-built installation on a modest site, a well-made overshot wheel is frequently the better engineering decision.
03
The leat, the headrace and the water you actually get
Most of the civil work in a water power installation is not the wheel, and most of the failures are here.
A leat is the channel that carries water from the intake to the wheel, and its job is to lose as little head as possible over the distance. It runs at a very gentle gradient, because too steep and the water arrives fast and low, and too flat and it silts up and stops flowing.
The intake is what needs the most thought. It has to take water reliably at low flow, survive floods without being destroyed, and keep out leaves, gravel and debris that would otherwise reach the wheel. Screening, a settling section and an overflow that dumps excess before it reaches the wheel are the standard answers.
The tailrace carries water away, and if it backs up, the wheel runs in its own discharge and loses power. It needs to be clear, adequately sized and free-draining even in high flow.
And there needs to be a way to shut the water off, so that maintenance is possible without waiting for a dry season. A sluice at the intake is the traditional and correct answer.
⚠ Diverting, damming or abstracting from a watercourse is regulated nearly everywhere, and the rules differ by country, by state and sometimes by valley. This is a check-locally matter before it is an engineering one.
04
Building the wheel
A wheel is a large structure that lives permanently wet, turning slowly under a heavy and unbalanced load, and that's the design brief.
Timber has been the material for most of the history of the technology and remains a sound choice. Species matter: durability in permanent wet is the property to select for, and heartwood, not sapwood. Some timbers last far longer submerged than exposed to alternating wet and dry, which is why the parts that are permanently wet and the parts that are not may sensibly be different species.
The structure has to resist a load that's always uneven, since water is only in the buckets on one side. That means a rim that keeps its shape, spokes or arms that carry the load into the hub, and a hub and shaft that take the bending. Traditional construction used a clasp arrangement of arms around a square shaft in place of mortising into it, because mortising weakens the shaft exactly where the load is.
Bearings carry the whole weight while running slowly in a wet environment. Traditional practice used hardwood or stone bearings that could be replaced, with the shaft ends protected, and the modern equivalent is a sealed bearing chosen for the load and the conditions.
Bucket shape and spacing on an overshot wheel decide how much water is held and how long, and getting the water in cleanly at the top and out cleanly at the bottom is what separates a good wheel from a wet one. Water carried past the bottom is lifted weight doing no work.
05
Gearing, and what to drive
A wheel turns slowly with a great deal of torque, and almost nothing you want to drive wants that combination directly.
Gearing up trades that torque for speed. Traditional mills used large wooden gear trains, frequently with wooden teeth in an iron wheel deliberately, because wooden teeth are the sacrificial part and are far easier to replace than a broken casting. Belts and pulleys do the same job more simply and slip under overload, which is a protection instead of a fault.
For direct mechanical work the low speed is frequently an advantage: milling, pumping, sawing, hammering, grinding and driving a line shaft in a workshop all suit it, and no conversion is needed at all.
For electricity a generator wants high speed, so the gearing ratio becomes large. That's entirely feasible and it adds complexity, losses and maintenance to a system that would otherwise be almost maintenance-free.
The decision worth making deliberately is whether you need electricity at all, or whether the loads you actually have can be driven mechanically. A wheel driving a workshop directly is a simpler and more durable machine than one generating power to run motors.