Over - under expanded nozzle flows: Question bank

Before we start, one shared vocabulary reminder so nothing below is used unexplained:

Over-expanded $= p_{e} < p_{b}$ (expanded too much → too low → shocks). Under-expanded $= p_{e} > p_{b}$ (not expanded enough → too high → fans). Perfectly expanded $= p_{e} = p_{b}$ .

Where the exit-pressure formula comes from (the missing "why")

The traps below lean on one master formula, so let us earn it rather than quote it.

WHY does this hold? Two physical facts, chained:

Energy is conserved in adiabatic flow: the "total energy per kilogram" $h_{0} = h + \frac{1}{2} V^{2}$ stays constant. As the gas speeds up ( $V$ grows), its thermal part $h$ (hence its temperature) must drop. Written in Mach number this becomes $\frac{T _{0}}{T} = 1 + \frac{γ - 1}{2} M^{2}$ — the faster you go, the colder you get.
The process is isentropic (smooth, no shocks, no friction), so pressure and temperature are locked by $p / T^{γ / (γ - 1)} = const$ . Feed fact 1 into this and pressure inherits the same bracket raised to the power $γ / (γ - 1)$ .

That is the whole "why": going faster cools the gas (energy), and cooling drops the pressure (isentropy) — see Isentropic Flow Relations. So a bigger $M_{e}$ always means a smaller $p_{e}$ , and $M_{e}$ itself is set purely by the area ratio $A_{e} / A^{*}$ via the Area-Mach Number Relation. Look at the curve:

Figure — Over - under expanded nozzle flows

True or false — justify

A supersonic nozzle can be redesigned for its back pressure by opening a valve downstream.

False. Once the diverging section is supersonic, pressure signals (moving at the local speed of sound) cannot travel upstream against the faster flow, so nothing downstream can change

p_{e}

; only geometry and

p_{0}

set it.

Over-expanded flow means the exit pressure is higher than ambient.

False. Over-expanded means expanded too much →

p_{e} < p_{b}

; the exit pressure is below ambient, which is why compression (shocks) is needed to raise it back up.

At the exact design point the pressure term in the thrust equation is zero.

True. Perfectly expanded means

p_{e} = p_{b}

, so

(p_{e} - p_{b}) A_{e} = 0

and all thrust comes from the momentum term

\overset{m}{˙} V_{e}

Under-expanded flow always produces more thrust than perfectly expanded flow.

False in the sense that matters. The pressure term

(p_{e} - p_{b}) A_{e}

is positive and adds thrust, but you sacrificed exit velocity

V_{e}

by not expanding fully, so overall the engine is below its optimum efficiency.

A rocket that is over-expanded at sea level will still be over-expanded in space.

False.

p_{e}

is fixed by geometry, but

p_{b}

falls as altitude rises; once

p_{b}

drops below

p_{e}

the same nozzle becomes under-expanded, which is exactly what happens as a rocket climbs.

Shocks in an over-expanded jet appear because the gas is moving too fast.

False. They appear because the pressure must rise from

p_{e}

up to

p_{b}

; compression demands shocks. Speed enables shocks to exist (they need supersonic flow) but the driver is the pressure direction.

Expansion fans compress the gas.

False. Prandtl–Meyer fans (see Prandtl-Meyer Expansion Fans) lower pressure and accelerate the flow; they appear in under-expanded jets precisely because the gas still has excess pressure to shed.

Raising the chamber pressure

p_{0}

while keeping the nozzle fixed pushes the flow toward under-expansion.

True. In

p_{e} = p_{0} / (1 + \frac{γ - 1}{2} M_{e}^{2})^{γ / (γ - 1)}

the exit Mach

M_{e}

is fixed by geometry, so

p_{e}

scales directly with

p_{0}

; higher

p_{0}

raises

p_{e}

, and if

p_{b}

is unchanged

p_{e}

can climb above

p_{b}

→ under-expanded.

Spot the error

"The nozzle is over-expanded, so we should make the diverging section shorter to fix it."

The verdict is right but only if you can show why. A less-flared exit means a smaller

A_{e} / A^{*}

; by the Area-Mach Number Relation a smaller area ratio gives a smaller supersonic

M_{e}

; and in the boxed formula a smaller

M_{e}

shrinks the denominator, so

p_{e} = p_{0} / (1 + \frac{γ - 1}{2} M_{e}^{2})^{γ / (γ - 1)}

rises toward

p_{b}

. The chain

A_{e} / A^{*} ↓\Rightarrow M_{e} ↓\Rightarrow p_{e} ↑

is the real reason it works.

"Since

p_{e}

depends only on geometry, thrust is independent of altitude."

Error:

p_{e}

is fixed, but thrust contains

(p_{e} - p_{b}) A_{e}

, and

p_{b}

changes with altitude. So thrust rises as the rocket climbs even with

p_{e}

locked — that pressure term is the whole reason altitude matters.

"A Mach disk is just a big oblique shock."

Error: a Mach disk is a normal (perpendicular) shock across the core of a strongly over-expanded jet, giving a large pressure jump; oblique shocks are weaker and inclined. Confusing them under-predicts the pressure rise and the entropy loss.

"To find

M_{e}

from the area ratio, just take the subsonic root of the area–Mach relation."

Error: the area–Mach relation gives two roots for a given

A_{e} / A^{*}

; a supersonic nozzle takes the supersonic root. The subsonic root describes a venturi that decelerates back down, not a de Laval exit.

"Because the flow is choked at the throat, the exit must also be at

M = 1

Error: choking fixes

M = 1

only at the throat. Downstream in the diverging section the supersonic branch accelerates past

M = 1

, so

M_{e} > 1

and grows with area ratio.

"The jet boundary of an over-expanded jet is a solid wall, so shocks reflect like off a mirror."

Error: the jet edge is a free (constant-pressure) boundary, not a wall. A shock hitting it reflects as an expansion fan (and vice versa), which is what builds the alternating diamond/shock-cell pattern.

Why questions

Why can't the back pressure "reach into" a supersonic exit and change

p_{e}

Pressure disturbances travel at the local sound speed; in supersonic flow the gas moves faster than that, so upstream-directed signals are swept downstream and never arrive at the exit plane.

Why is "over-expanded" such a confusing name?

It names the action (the gas was expanded too much) not the resulting pressure. Too much expansion drives pressure below ambient, so "over" pairs with a surprisingly low

p_{e}

Why does an over-expanded nozzle risk flow separation but an under-expanded one does not?

Over-expansion needs a pressure rise against the flow direction; a strong enough adverse gradient stalls the boundary layer off the wall inside the nozzle (see Flow Separation in Nozzles). Under-expansion needs a pressure drop, which the flow accepts happily and does entirely outside the nozzle.

Why does the pressure-thrust term

(p_{e} - p_{b}) A_{e}

appear at all in the thrust equation?

The control-volume momentum balance includes surface pressure forces on the exit plane:

p_{e}

pushes the gas out, atmospheric

p_{b}

pushes in over the same area

A_{e}

, and their net is that term.

Why is perfect expansion the thrust optimum rather than maximum-velocity expansion?

Expanding further raises

V_{e}

but drops

p_{e}

below

p_{b}

, turning the pressure term negative (a drag-like loss). The gain in momentum thrust is more than eaten by the pressure penalty, so the true peak sits at

p_{e} = p_{b}

(see Rocket Nozzle Design & Thrust Optimization).

Why do under-expanded jets form visible "shock diamonds" if the first waves are expansion fans?

The fans over-expand the jet below

p_{b}

; they reflect off the free boundary as compression waves that coalesce into oblique shocks, which then reflect back as fans — this repeating over/under cycle is the diamond pattern.

Why does a single fixed nozzle inevitably run off-design most of the time?

It is perfectly expanded for exactly one

p_{b}

; as ambient pressure changes (altitude, throttling)

p_{b}

drifts away from the fixed

p_{e}

, so real engines spend their flight over- or under-expanded, matched only momentarily.

Edge cases

What happens as

p_{b} \to p_{e}

from either side?

The external waves weaken toward vanishing; at

p_{b} = p_{e}

the jet leaves parallel and wave-free (perfect expansion). Approaching from above gives weakening shocks, from below gives weakening fans.

What is the flow like when

p_{b}

is very high — nearly equal to

p_{0}

The nozzle may not even reach supersonic exit; a normal shock sits inside the diverging section (or the flow is entirely subsonic), so the "supersonic-exit, geometry-locked

p_{e}

" story breaks down and back pressure does control the exit again.

What happens in the extreme case

p_{b} > p_{0}

(ambient exceeds even the chamber pressure)?

There is no longer any pressure drop to push gas out — the driving pressure difference reverses. The nozzle cannot establish forward supersonic flow at all; it either stops flowing or the atmosphere pushes gas backward into the chamber. Every over/under-expanded concept assumes

p_{0} > p_{b}

, so this case sits entirely outside the regime map.

What is the limiting under-expanded case as

p_{b} \to 0

(firing into vacuum)?

p_{e} > p_{b}

always, so the jet is maximally under-expanded; the exhaust plume expands enormously outward with strong fans, and the pressure-thrust term

(p_{e} - 0) A_{e} = p_{e} A_{e}

reaches its largest positive value.

At the throat itself, is the flow ever over- or under-expanded?

No. The throat is choked at

M = 1

regardless of regime; over/under-expansion is a statement about the exit pressure versus ambient, decided further downstream in the diverging section.

A_{e} / A^{*} = 1

, what regime can occur?

Then the "diverging" section is absent and

M_{e} = 1

(sonic exit). There is no supersonic expansion, so the classic over/under-expanded shock-and-fan behaviour doesn't develop; the exit is just choked and sonic.

What happens to an over-expanded oblique-shock pattern if

p_{b}

keeps rising?

The required pressure jump grows until oblique shocks can no longer supply it; a stronger normal shock / Mach disk forms, and with further increase the shock marches upstream into the nozzle, triggering wall separation (see Oblique Shocks and Flow Separation in Nozzles).

Recall One-line summary to lock it in

Geometry sets $p_{e}$ ; the sky sets $p_{b}$ ; the mismatch $p_{e}$ vs $p_{b}$ decides shocks (over, $p_{e} < p_{b}$ ) or fans (under, $p_{e} > p_{b}$ ), and $(p_{e} - p_{b}) A_{e}$ is where it hits your thrust.