Why Your Turbocharger Uses a Radial Turbine and a Jet Engine Doesn't

August 21, 2026

Open a turbocharger and the turbine wheel takes gas in at its outer diameter and throws it out along the axis. Open a jet engine and the turbine takes gas in along the axis and passes it straight through, stage after stage.

Both are turbines. Both extract shaft work from expanding gas. They look nothing alike, and the reason is a term in an equation that almost nobody in the aftermarket has heard of.

The quantity that matters inside a rotor

Outside a rotor, the useful bookkeeping quantity is stagnation enthalpy: h₀ = h + V²/2. Static enthalpy plus kinetic energy. It is what you use to talk about what a turbine stage can extract.

Inside a spinning rotor, that is the wrong frame. The blade is moving, so the gas velocity you care about is the velocity relative to the blade, and the correct conserved quantity is rothalpy:

I = h + w²/2 − u²/2

where h is static enthalpy, w is gas velocity relative to the blade, and u is local blade speed, ω·r. For steady adiabatic flow through a rotor, rothalpy at the exit equals rothalpy at the inlet.

Note the minus sign on the u² term. That sign is the whole story.

Where the free work comes from

Take the Euler turbomachinery equation for specific work and decompose it into its three physical contributions:

w = ½(c₁² − c₂²) + ½(u₁² − u₂²) + ½(w₂² − w₁²)

Reading left to right: the change in absolute kinetic energy, the change in blade speed squared, and the change in relative velocity through the passage.

The first and third terms are ordinary aerodynamics. Every blade row in every turbomachine, axial or radial, works those two terms. Blade shape, incidence, camber, solidity — all of that is engineering effort spent on the first and third terms.

The middle term is different. It only exists if the gas changes radius while inside the rotating field. In a radial inflow turbine, gas enters at the outer diameter where u is large and exits near the axis where u is small, so u₁ > u₂ and the term is positive. It is a contribution to work extraction that no blade profile produced. You collect it for moving gas toward the axis.

In an axial turbine, u₁ ≈ u₂. The middle term is essentially zero. Every joule has to be won by the first and third terms — by aerodynamics alone.

Why that decides the geometry

An axial turbine that has to do all its work aerodynamically needs stages. Stages need length, blade rows, discs and shafting. That is fine in a jet engine, where the turbine runs at near-constant speed for hours and nobody cares how long it takes to accelerate. Steady-state efficiency is the objective and mass is amortised over an entire flight.

A turbocharger has the opposite problem. It spends its life accelerating and decelerating. What limits how fast it can respond is rotational inertia — and inertia scales badly, roughly with mass times radius squared, so every extra blade row and every millimetre of shaft costs you response.

The radial wheel banks the centrifugal term in a single short passage. One rotor, no stages, minimal shaft length, low inertia. It gives up some peak efficiency compared with a well-developed multi-stage axial machine, and it does not care, because peak efficiency is not what a turbocharger is being judged on.

Spool is a rotational inertia problem at least as much as a gas-flow problem. That sentence is the practical takeaway, and rothalpy is the reason the geometry that solves the work problem also happens to solve the inertia problem.

What this says about pulse energy

There is a second consequence worth drawing out.

A radial wheel is fed from its outer diameter through a volute that wraps the rotor. That geometry is unusually good at converting a pressure pulse into work, because the entire circumference is available and the gas is doing useful work from the instant it enters. It is part of why exhaust pulse energy matters so much on a turbocharged engine and why the design of what sits upstream — manifold length, collector geometry, whether the housing is divided — has such a large effect on low-speed response.

This is the mechanism underneath the twin-scroll argument. A divided housing keeps pulses separated all the way to the wheel, which preserves the peaks that the radial geometry is good at cashing in. Which cylinders share a passage, and whether they share one at all, is covered in twin-scroll vs single-scroll on the AMG 4.0L V8.

The adiabatic assumption, and why it breaks

Rothalpy is conserved for steady adiabatic flow. Real turbocharger rotors are not adiabatic, and on a turbocharger the violation is unusually large.

The turbine housing, bearing housing and compressor housing are bolted together in a short stack, with a temperature difference of many hundreds of degrees across a few centimetres. Heat flows continuously from the hot end to the cold end, including through the gas paths themselves.

The consequence shows up in the maps you are given. Turbine outlet temperature is measured lower than a true adiabatic expansion would produce, which inflates apparent turbine efficiency. Compressor outlet temperature is measured higher, which deflates apparent compressor efficiency. Both errors are worst at low speed and low mass flow, because the heat transfer rate barely changes while the work terms shrink — which is exactly the region where transient response lives.

That is why serious 1-D engine simulation carries an explicit thermal network for the turbocharger instead of trusting a published map. The map is a real measurement of a real machine; it is just not a measurement of the thing you think it is measuring. There is more on this in exhaust heat retention and turbo response.

And the point most vendors skip

Rothalpy is defined at the rotor. It is inlet-to-exit accounting across the wheel, in the rotating frame, and every term in it describes something happening between the volute and the exducer.

Nothing bolted downstream of the turbine housing appears anywhere in that expression.

Which is not an argument against exhaust work — a catless downpipe lowers the pressure the turbine expands into, and that raises the pressure ratio across the wheel, which is a real and useful effect on the boundary condition. It is an argument against the specific claim that what you attach after the turbine changes how the turbine itself behaves internally. It does not. It changes what the turbine is expanding against.

Precision about which of those two things you are buying is the difference between a modification that does what you expected and one that does not.

Item added to your cart