The Tesla Turbine: A Bladeless Turbine Driven by Friction

The mechanism: momentum diffuses sideways through the boundary layer

In every fluid there is a no-slip condition: the layer of fluid touching a solid surface moves at exactly the surface's speed. Between the free stream and the wall lies the boundary layer, where velocity is sheared. Shearing a viscous fluid costs momentum, and that momentum is deposited into the wall as a tangential force. The Tesla turbine is built entirely around harvesting that force.

Fluid enters through a nozzle at the rim, tangential to the disc pack, moving faster than the discs. As it spirals inward it drags on both faces of every gap. The shear stress on each disc face is set by Newton's law of viscosity:

τ = µ · (∂u/∂y)

where µ is dynamic viscosity, u the tangential fluid velocity, and y the distance across the gap. Integrate that stress over the whole disc area and you get a torque on the shaft. Because the flow slows as it gives up momentum, its angular momentum falls; the difference — τ_shaft = ṁ · (r_in·V_θ,in − r_out·V_θ,out), the Euler turbomachine equation — is exactly what the discs collect. Unlike a bladed rotor, no pressure difference across a blade does the work; it is pure friction, applied over a large wetted area.

Why the gap must be about twice the boundary-layer thickness

The single most important design number is the inter-disc gap b. If the gap is much larger than the boundary layer, the fluid in the middle of the channel slides through untouched — it never couples to the disc, so its energy is wasted as it exits still moving. If the gap is much smaller, viscous drag chokes the flow and windage losses dominate. The sweet spot is when the two boundary layers (one from each disc) just fill the channel, so essentially the entire flow is being sheared:

b ≈ 2δ, where δ ≈ √(µ·L / (ρ·V))

Here δ is the laminar boundary-layer thickness, ρ density, V flow speed, and L a flow-path length. For steam or air at working speeds this makes δ only a few tenths of a millimetre, so practical gaps run 0.2–1 mm — and hundreds of discs may be stacked to get useful throughput. This is why the turbine is intrinsically a laminar, low-Reynolds-number machine: the physics rewards keeping the flow between the discs smooth and viscous, not turbulent.

A worked spec: Tesla's own steam prototype

Around 1910–1911, Tesla tested machines at the New York Edison Waterside Station. The figures are historically documented:

• Rotor: a stack of flat steel discs about 18 in (≈ 457 mm) in diameter, spaced roughly 0.8 mm apart.
• Steam supply: ≈ 125 psi (≈ 8.6 bar).
• Output: up to ~200 hp (≈ 150 kW) from a single stage.
• Speed: ≈ 9,000 rpm — very high, because the disc tips must run near the fluid speed to keep the shear favourable.

A later, more compact demonstration unit fit two 25 cm rotors on a base the size of a hat and was claimed to make similar power. The catch was always efficiency: Tesla's turbine converted heat to work at a small fraction of the ~85% a good bladed steam stage of the era achieved. Decades later, Warren Rice at Arizona State University built careful multi-disc rotors and measured a rotor efficiency around 40–55% in ideal conditions — respectable, but only for the rotor in isolation, and only in the laminar regime.

Trade-offs, failure modes, and the efficiency ceiling

The turbine's great virtues are directly tied to its great weakness.

• Ruggedness: with no thin trailing edges to erode, it shrugs off dirty steam, sand-laden water, cavitation and even two-phase (wet) flow. Nothing on it stalls.
• Cheapness: the rotor is just flat discs — no investment-cast airfoils, no 5-axis machining. That is its enduring appeal.
• The ceiling: to get high rotor efficiency the fluid must give up nearly all its tangential speed to the discs, which happens only over a long spiral path at high rpm. But push mass flow up and the channel goes turbulent, drag collapses, and efficiency falls off a cliff. One modern small-scale air prototype tested at around 3 bar peaked at only ~15% overall.
• Disc dynamics: sub-millimetre gaps at 9,000+ rpm demand tight balance. Thin discs bow, flutter, and rub; thicker discs blunt the inlet and one study found efficiency dropping ~45% when disc thickness went from 1 mm → 2 mm. Getting the exit right (holes near the hub) is also delicate — the flow must leave with almost no residual swirl, or that energy is lost.

Where it actually earns its keep

The Tesla turbine never displaced the steam turbine, but the same disc-pack principle is genuinely useful where blades fail:

• Disc (viscous-drag) pumps: run the machine in reverse and it becomes a superb pump for slurries, polymers, sewage, and shear-sensitive fluids (live cells, foams) precisely because there are no blades to clog, chop, or cavitate. Discflo Corporation has sold bladeless disc pumps commercially for decades on exactly this logic.
• Micro-scale power: at chip and MEMS scale, viscosity dominates and Reynolds numbers are naturally low — the Tesla geometry is being studied for waste-heat organic-Rankine-cycle (ORC) micro-turbines and small compressed-air expanders.
• Flow meters, blowers and vacuum pumps using smooth-disc rotors for the same fouling-resistance reasons.

In short: whenever the fluid is thick, dirty, or the flow is small, the friction turbine's weaknesses stop mattering and its ruggedness starts to pay.

The common misconception

The persistent myth is that Tesla's turbine was a suppressed 95%-efficient miracle that only failed for lack of good metallurgy. That is not what the record or the physics says. Tesla did claim very high potential efficiency, but measured units never approached it, and the reason is fundamental, not metallurgical: the machine relies on viscous shear, and viscous shear is a dissipative process. You are deliberately using friction to transmit power — some of that friction inevitably heats the fluid instead of turning the shaft. A bladed turbine turns nearly all the flow's momentum into blade force with little internal shearing; the Tesla turbine cannot, by construction, avoid the entropy generation in its boundary layers. The subtle engineering pitfall follows from this: engineers instinctively try to increase flow to get more power, but for a disc turbine more flow makes it worse once the gaps go turbulent. It is a machine that rewards patience — thin gaps, laminar flow, high speed, low mass rate — not brute force.

Tesla (bladeless) turbine vs. a conventional bladed impulse/reaction turbine

Property	Tesla disc turbine	Bladed turbine (e.g. Pelton/steam stage)
Drive mechanism	Viscous shear on flat discs (boundary layer)	Pressure/momentum on curved blades
Rotor cost & manufacture	Very low — flat discs, no airfoils, forgiving	High — precision-cast/machined 3D blades
Peak isentropic efficiency	~40–55% (rotor), often much lower	85–92% for a well-designed stage
Tolerance to dirt / cavitation / two-phase	Excellent — nothing to erode or foul	Poor — blades pit, erode, cavitate
Best operating regime	Low flow, high viscosity, small scale	High mass-flow, clean single-phase fluid
Speed at good efficiency	Very high rpm (rotor tip ≈ fluid speed)	Moderate, geared to load