A free-stream water wheel — sometimes called a stream wheel — is a vertical-axis paddle wheel partially submerged in a flowing river or tidal channel, extracting energy from the water's kinetic flow rather than from a head drop. A practical stream wheel runs at a tip speed of roughly 0.4 to 0.5 times the free-stream velocity, delivering 15 to 25% hydraulic efficiency at flows of 0.8 to 2.0 m/s. We see this form on tidal mills like Eling Tide Mill in Hampshire and on floating ship-mills historically moored on the Danube and the Tigris.
Water Wheel Form 4 Interactive Calculator
Vary stream speed, capture area, power coefficient, wheel diameter, and tip-speed ratio to see free-stream wheel power, speed, and torque.
Equation Used
The calculator uses the free-stream kinetic power equation from the article. Stream power is 0.5*rho*A*v^3, and the extracted shaft power is Cp times that value. Wheel RPM is estimated from the selected tip-speed ratio, where the paddle tip speed is lambda*v.
- Fresh water density is fixed at rho = 1000 kg/m3.
- Capture area A is the effective submerged projected paddle area in the stream.
- Cp represents overall kinetic power extraction coefficient.
- Wheel speed is set by tip-speed ratio and diameter.
How the Water Wheel (form 4) Actually Works
A free-stream wheel sits in the current with its lower paddles immersed by a fixed depth — typically 200 to 400 mm — and the moving water pushes those paddles backward. Unlike an undershot wheel running in a controlled millrace, this form has no headrace, no penstock, and no controlled tailwater. The wheel takes whatever the river or tide is doing and converts a fraction of the flow's kinetic energy into shaft torque. Because there is no head drop, output scales with the cube of velocity — drop the river speed from 1.5 m/s to 1.0 m/s and you lose roughly 70% of your power, not 33%.
The geometry is dictated by the tip speed ratio. If the paddle moves too slowly, water just piles up against it and spills around the sides without transferring momentum. If it moves too fast, the paddle outruns the water and the relative velocity collapses. The sweet spot sits near a tip speed ratio of 0.4 — slow enough that water actually impacts the paddle, fast enough that fresh water keeps reaching the next paddle in line. This is why you'll see free-stream wheels run at 4 to 8 RPM on a 3 m diameter wheel rather than the 12 to 15 RPM of a controlled overshot.
Get the immersion depth wrong and the whole system misbehaves. Too shallow — the bottom paddle barely engages the flow and torque drops off. Too deep — the upstream face of the paddle drags through stagnant water that the wheel itself has slowed down, robbing efficiency and adding cyclic load on the bearings. Bushings on heritage installations like the floating ship-mills of the lower Danube wore through in 18 months when builders pushed immersion past 1/3 of the paddle height, because the lower paddle was effectively churning a captured slug of water rather than meeting fresh stream.
Key Components
- Paddles (floats): Flat or slightly cambered boards bolted to the rim, typically 150 to 400 mm deep and spaced at 12 to 24 around the wheel. The paddle face must sit perpendicular to flow at the bottom-dead-centre position within ±3° — more skew than that and the paddle deflects water sideways instead of converting momentum to torque.
- Rim and arms: Two parallel rims connected by spokes carry the paddles. On larger wheels (above 2.5 m diameter) the arms are tensioned with iron tie rods because the cyclic loading from each paddle entry and exit fatigues unbraced timber spokes within a few seasons.
- Axle and bearings: A horizontal shaft running across the channel, supported on plain bronze or lignum vitae bushings in heritage builds and on sealed roller bearings in modern restorations. Shaft deflection must stay below 1 mm/m of span, otherwise the paddles run unevenly and one side of the wheel takes a disproportionate share of the load.
- Float pontoons or fixed mounts: On a free-floating ship-mill the wheel hangs between two boats so it can rise and fall with river level. On fixed installations like tidal mills, the wheel pit must accommodate the full tidal range — at Eling Tide Mill that's roughly 4 m of variation between springs and neaps.
- Headboards and side cheeks: Vertical boards either side of the wheel that prevent water spilling around the paddle ends. Gap between paddle tip and side cheek must be tight — 15 to 25 mm — or you lose 5 to 10% of your output through end leakage.
Where the Water Wheel (form 4) Is Used
Free-stream wheels make sense wherever you have moving water but no usable head — wide slow rivers, tidal channels, and irrigation canals. They lose to higher-head wheels on efficiency, but they win on simplicity and on sites where you cannot legally or practically build a weir.
- Heritage tidal milling: Eling Tide Mill in Hampshire uses a tidal pond that drains through undershot wheels twice per tide cycle, producing about 4 hours of grinding per tide.
- Floating ship mills: The reconstructed Murau ship-mill on the Mur in Austria — a twin-hulled boat with a paddle wheel between the hulls, milling grain while moored in the current.
- Irrigation lifting: Noria-style stream wheels at Hama in Syria lift water 20 m from the Orontes river using attached pots on a 21 m diameter wheel, no external power required.
- Micro-hydro on slow rivers: Hydrokinetic stream-wheel installations on the Mississippi for off-grid riverbank cabins, generating 200 to 500 W continuous from 1.0 to 1.5 m/s flow.
- Living-history demonstration: The reconstructed Roman ship-mill at the Mainz Museum of Ancient Seafaring, used as an operating educational exhibit on Roman hydraulic engineering.
- Aquaculture aeration: Slow-turning stream wheels driving paddle aerators in flow-through trout raceways, where the same flow that oxygenates the fish powers the agitator.
The Formula Behind the Water Wheel (form 4)
The useful output of a free-stream wheel is set by how much kinetic energy passes through its swept area and what fraction the wheel can capture. At the low end of typical river speeds — around 0.8 m/s — power is so marginal that bearing drag and gear losses eat most of it, and you'll struggle to do useful mechanical work. At a nominal 1.5 m/s the wheel hits its design sweet spot, where tip speed ratio, immersion depth, and paddle count all line up. Push beyond 2.5 m/s and you start fighting cavitation behind the paddles and structural fatigue in the arms — so the practical envelope is narrower than the cubic curve suggests.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| P | Hydraulic power captured at the shaft | W | ft·lb/s |
| Cp | Power coefficient (efficiency) of the wheel, typically 0.15 to 0.25 for free-stream wheels | dimensionless | dimensionless |
| ρ | Density of water, ~1000 kg/m³ fresh, ~1025 kg/m³ seawater | kg/m³ | lb/ft³ |
| A | Submerged frontal area of paddles in contact with flow (paddle depth × wheel width) | m² | ft² |
| v | Free-stream water velocity upstream of the wheel | m/s | ft/s |
Worked Example: Water Wheel (form 4) in a tidal-channel stream wheel for an oyster hatchery
You are sizing a free-stream water wheel to drive a seawater circulation pump for an oyster hatchery on a tidal channel in Galway Bay. The channel runs 1.5 m/s on the mid-tide ebb, dropping to 0.8 m/s near slack and peaking at 2.2 m/s on spring tides. The wheel is 2.4 m diameter with 0.6 m wide paddles immersed 0.30 m, giving a submerged frontal area of 0.30 × 0.60 = 0.18 m². Assume Cp = 0.20 (a realistic figure for a well-built timber stream wheel) and seawater density ρ = 1025 kg/m³.
Given
- Cp = 0.20 dimensionless
- ρ = 1025 kg/m³
- A = 0.18 m²
- vnom = 1.5 m/s
- vlow = 0.8 m/s
- vhigh = 2.2 m/s
Solution
Step 1 — compute power at nominal mid-ebb velocity of 1.5 m/s:
62 W is enough to drive a small magnetic-coupled circulation pump moving roughly 30 L/min through the hatchery's settling tanks. The wheel spins at about 6 RPM at this flow — slow enough you can count paddle entries by eye, fast enough that water doesn't pile up.
Step 2 — at the low end near slack tide, 0.8 m/s:
9 W barely overcomes bearing drag and the mechanical seal on the pump. In practice the wheel will turn but deliver almost no useful flow for the 90 minutes either side of slack — you need a small battery buffer or accept dead time twice per tide.
Step 3 — at the high end on spring tides, 2.2 m/s:
Theoretical 197 W, but in practice Cp drops at this speed because the wheel cannot accelerate fast enough to maintain the 0.4 tip speed ratio — the paddles get hammered rather than gently loaded. Expect maybe 140 W at the shaft, plus structural cyclic loading that will fatigue any unbraced timber arms within a season.
Result
The nominal output is 62 W of shaft power at 1. 5 m/s mid-ebb flow — useful, modest, and well-matched to a hatchery circulation duty. The full operating envelope swings from 9 W near slack to a theoretical 197 W on spring tides, with the design sweet spot sitting between 1.3 and 1.8 m/s where the wheel runs efficiently and structurally calm. If you measure significantly less than the predicted 62 W, check three things in order: (1) immersion depth — if the channel bed has silted up and your paddles are now sitting 100 mm deeper than design, you're churning a captured slug of water and losing 20 to 30% to drag; (2) paddle-to-cheek gap — gaps over 30 mm let 10%+ of the flow bypass the paddle entirely; (3) shaft alignment — if the bearing pedestals have shifted and the wheel is running with end-thrust against a flange, you can lose 15 W or more to friction at this scale.
Choosing the Water Wheel (form 4): Pros and Cons
Free-stream wheels are not the most efficient water wheel — they're the most permissive. You pick this form when you cannot build a head, not when you want maximum output per litre of flow. Compared to a controlled overshot or a modern hydrokinetic turbine, the stream wheel trades efficiency for buildability and regulatory simplicity.
| Property | Free-stream water wheel | Overshot water wheel | Hydrokinetic turbine |
|---|---|---|---|
| Hydraulic efficiency (Cp) | 0.15 to 0.25 | 0.65 to 0.85 | 0.35 to 0.45 |
| Operating speed | 4 to 10 RPM | 8 to 15 RPM | 60 to 200 RPM |
| Site requirement | Any moving water, no head | Requires 2 m+ head drop | Any moving water, no head |
| Build cost (small scale) | Low — timber and bronze | Medium — timber, sluice, headrace | High — composite blades, generator, control |
| Maintenance interval | Annual paddle inspection | Annual rim and bucket inspection | Bearing service every 2 to 3 years |
| Lifespan (typical timber/steel build) | 20 to 40 years | 30 to 80 years | 10 to 20 years |
| Power per m² of swept area at 1.5 m/s | ~340 W/m² | Not comparable (head-based) | ~600 W/m² |
| Permitting complexity | Low — no impoundment | High — weir/dam consents | Low to medium |
Frequently Asked Questions About Water Wheel (form 4)
You're running above the optimal tip speed ratio. The wheel is freewheeling — paddles enter and exit the water without taking a proper bite, so they generate little drag against the flow. This usually happens when the load on the shaft is too light: a small pump, a slipping belt, or a generator that hasn't engaged.
Add load until you pull the wheel down to roughly 0.4 × v / r RPM (so for v = 1.5 m/s on a 1.2 m radius wheel, that's about 5 RPM). You should see torque climb sharply as the paddles start meeting water with proper relative velocity rather than skating across the surface.
Width is almost always cheaper and structurally safer. Power scales linearly with both diameter and width through the submerged area term, but diameter also drives bending moment on the shaft and squared cyclic stress on the arms. Doubling the wheel width doubles power and doubles shaft torque. Doubling diameter doubles power but quadruples the bending stress in the spokes.
Practical rule: stretch width up to roughly 1.5 × diameter before going wider becomes awkward (paddle deflection, end leakage at the cheeks). Beyond that, diameter increase is the only path forward, and you'll need to engineer the spoke bracing properly.
You're measuring at a load-limited operating point, not a flow-limited one. If the pump or generator on the shaft has a fixed torque demand below what the wheel can deliver at full flow, the wheel just spins slower against that fixed load and shaft power tracks the load curve, not the flow curve.
To verify, briefly remove the load and let the wheel free-spin. Measure the no-load RPM at both flow conditions. If no-load RPM scales roughly linearly with velocity (it should), the wheel is fine — your load is masking the flow change. The cube law is real but only fully visible when the load matches the available power across the full flow range.
Flat. Curved buckets are directional — they capture flow well in one rotation direction and badly in the other. On a reversing tidal stream you'd need an articulating paddle or a clutched freewheel to handle both ebb and flood, which adds expense and failure modes salt water punishes.
Flat paddles work symmetrically. The wheel just reverses direction with the tide, which is fine if your pump or generator accepts bidirectional input (a centrifugal pump won't; a paddle aerator or a rectified DC alternator will).
You're hitting paddle-entry resonance. Each paddle slamming into the water is a discrete impulse, and at a particular RPM the impulse frequency matches a natural frequency of the wheel-and-shaft assembly. With 16 paddles at 6 RPM you're putting 1.6 impulses per second into the structure — easily inside the bending mode of a 3 m timber arm assembly.
Two fixes: stiffen the arms with iron tie rods (raises the natural frequency above the operating range) or change paddle count to shift the excitation frequency. Going from 16 to 12 paddles drops impulse rate by 25% — often enough to clear the resonance band.
You need 250:1 reduction, which is too much for a single stage. Use a two-stage compound: a 1:15 belt or chain primary off the wheel shaft (so an intermediate jackshaft at 90 RPM), then a 1:17 belt or planetary into the generator. Each stage at 95% efficiency gives 90% overall — losing 10% rather than the 25 to 30% you'd see from a single overloaded worm box.
Avoid worm gears here. They're tempting because they hit big ratios in one stage, but they typically run 60 to 70% efficient at low input speed, which is catastrophic when you only have 60 W to start with. Belts, chains, and planetary stages all comfortably exceed 95% per stage at these powers.
You sized for mean load, not peak. A paddle entering the water at full immersion sees an impact load 3 to 5 times the steady-state force, especially if the wheel is running slightly fast and the paddle slaps the surface rather than sliding in. Add wave action or floating debris and peak loads can hit 8× mean.
Rule of thumb: design paddle fasteners for at least 6× the steady force from the formula, and use through-bolts with backing plates rather than wood screws or lag bolts. Stainless A4 (316) M8 minimum on anything larger than a 1.5 m wheel in salt or brackish water — A2 (304) pits and fails in tidal service within two seasons.
References & Further Reading
- Wikipedia contributors. Water wheel. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
