Over-expanded nozzle — oblique shocks in plume, efficiency loss: Question bank

This is a rapid-fire trap bank for the parent topic. Every item is a question ::: answer reveal — cover the answer, commit out loud, then check. The point is not arithmetic (see D3/D4 for that) but catching the wrong-but-tempting picture before an exam does.

Before any trap can be judged, every symbol it uses must already have a plain-word meaning and a picture. So we build the whole vocabulary first, then hunt misconceptions.

The symbols, earned one at a time

The thrust equation appears in almost every trap, so unpack its three symbols before using it.

Definition The thrust equation and its parts

$F = \overset{m}{˙} v_{e} + (p_{e} - p_{a}) A_{e}$

$\overset{m}{˙}$ = mass flow rate — kilograms of exhaust leaving the nozzle each second. The dot is Newton's shorthand for a time derivative, so $\overset{m}{˙} = \frac{d m}{d t}$ : the rate at which mass $m$ leaves per unit time $t$ . More gas per second → more push.
$v_{e}$ = exit velocity — how fast (in m/s) that gas leaves. Faster exhaust → more push.
$A_{e}$ = exit area — the physical area of the nozzle mouth in m². The pressure mismatch $(p_{e} - p_{a})$ acts over this whole disc, so a bigger mouth multiplies both the gain (if $p_{e} > p_{a}$ ) and the penalty (if $p_{e} < p_{a}$ ).
$\overset{m}{˙} v_{e}$ is momentum thrust (the jet's recoil); $(p_{e} - p_{a}) A_{e}$ is pressure thrust (net push on the exit disc). The full picture is in the Thrust Equation note: $F$ is force, split into "stuff thrown" and "pressure imbalance on the mouth."

Now the flow-geometry symbols. Look at the figure below while reading each definition.

Figure — Over-expanded nozzle — oblique shocks in plume, efficiency loss

Definition Mach number, shock angle, deflection angle

$M$ = Mach number — flow speed divided by the local sound speed. $M > 1$ is supersonic; a nozzle exit is $M_{e} > 1$ . $M_{1}$ is the Mach number just upstream of a shock.
$β$ (beta) = shock angle — the tilt of the shock line relative to the incoming flow (the slanted violet line in the figure).
$θ$ (theta) = deflection angle — how much the streamlines bend as they cross that shock (the orange kink). Sign convention: in an over-expanded plume we take $θ > 0$ for a turn toward the centerline (inward), which is the direction the ambient squeeze always forces; a compression shock can only turn the flow into itself, never away, so $θ$ is positive throughout this page. For an attached oblique shock $β > θ$ always: the shock line is steeper than the turn it produces.
$M_{1 n} = M_{1} sin β$ = the normal Mach component, the part of the flow's speed pointing straight through the shock. Only this component gets compressed; the sideways part slides along unchanged. This is why an oblique shock is gentler than a head-on (normal) one — it only processes a fraction of the flow's speed.

Finally, stagnation pressure — the single quantity that measures loss.

Definition Stagnation pressure

$p_{0}$ and why it is the "quality gauge" Imagine slowing a moving stream of gas to a dead stop with no friction. The pressure it would then reach is its stagnation pressure $p_{0}$ — it bundles the static pressure plus all the kinetic energy converted back to pressure. Think of it as the flow's total "usable pressure budget."

When we cross one shock we label the value just upstream $p_{0, 1}$ and the value just downstream $p_{0, 2}$ ; the ratio $p_{0, 2} / p_{0, 1}$ is what survives.
In smooth (isentropic) flow $p_{0}$ is conserved — no budget is spent.
Across a shock $p_{0}$ drops ( $p_{0, 2} < p_{0, 1}$ ): some ordered energy became heat and can never be fully recovered. That drop is the loss.
The velocity coefficient $C_{v} = \frac{v _{e, actual}}{v _{e, ideal}} = \frac{p _{0, actual}}{p _{0, chamber}}$ compares the flow's surviving pressure budget to the chamber's original budget; since $p_{0}$ can only fall, $C_{v} \leq 1$ always.

Look at the plume-structure figure to see where all of this lives inside a real over-expanded jet.

In that figure the navy bell on the left is the nozzle; the magenta curves are the plume boundary being pinched inward (the navy arrows show $p_{a}$ pushing in, because $p_{e} < p_{a}$ ); each violet line is one oblique shock like the $β$ -line of the first figure; the orange bar on the axis is the Mach disk — a small normal shock where opposite oblique shocks meet. One bright "diamond" cell is one full shock-then-expand cycle, and the whole train repeats down the axis.

The last figure shows how the loss accumulates — the stagnation-pressure ratio across one oblique shock as a function of $M_{1 n}$ , and how three such cuts multiply. Refer to it whenever a trap mentions "compounding losses."

Formula Stagnation-pressure loss across one oblique shock

Apply the normal-shock loss law to the normal component $M_{1 n} = M_{1} sin β$ , where $p_{0, 1}$ is the stagnation pressure just upstream of the shock and $p_{0, 2}$ just downstream: $\frac{p _{0, 2}}{p _{0, 1}} = (\frac{( γ + 1 ) M _{1 n}^{2}}{2 + ( γ - 1 ) M _{1 n}^{2}})^{\frac{γ}{γ - 1}} (\frac{γ + 1}{2 γ M _{1 n}^{2} - ( γ - 1 )})^{\frac{1}{γ - 1}}$ Here $γ$ is the ratio of specific heats of the gas. At $M_{1 n} = 1$ each of the two factors separately equals 1: the first factor's numerator and denominator both become $(γ + 1)$ , and the second factor's denominator $2 γ - (γ - 1) = γ + 1$ cancels its numerator — so their product is $1$ (a Mach wave, zero loss). As $M_{1 n}$ grows past 1 the product falls below 1. Let $N$ be the number of oblique shocks the flow crosses in one reflection pattern of the diamond train; the survivors multiply as $(p_{0, 2} / p_{0, 1})^{N}$ — fractions below 1 multiplied together only shrink, so losses stack, never cancel.

With every symbol now anchored to words, a picture, and a formula, the traps below are self-contained.

True or false — justify

Over-expansion means the gas has too little energy to escape the nozzle.

False — the gas has plenty of energy; over-expansion means it expanded to a pressure below ambient (

p_{e} < p_{a}

), not that it lacks speed. It is a pressure-matching problem, not an energy shortage.

p_{e} = p_{a}

the pressure-thrust term is exactly zero.

True — thrust is

F = \overset{m}{˙} v_{e} + (p_{e} - p_{a}) A_{e}

, and

p_{e} - p_{a} = 0

kills the second term, leaving only momentum thrust

\overset{m}{˙} v_{e}

. This is the perfectly-matched, maximum-efficiency condition.

An over-expanded nozzle always produces less total thrust than a perfectly matched one at the same altitude.

True at that altitude — the pressure term

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

is negative because

p_{e} < p_{a}

, so it subtracts from the momentum thrust. Matching would restore

p_{e} = p_{a}

and recover that lost term.

The shocks in an over-expanded plume are isentropic.

False — shocks are inherently irreversible and raise entropy, so they drop

p_{0}

. That entropy rise is exactly why over-expansion destroys usable energy, unlike the reversible expansion fans of under-expansion.

Over-expansion and under-expansion are symmetric off-design conditions with equal efficiency penalties.

False — under-expansion relieves pressure through expansion fans, which are a spread-out family of infinitely weak Mach waves: each raises entropy by a negligible amount and the total is isentropic (reversible), so

p_{0}

is preserved. Over-expansion instead forces compression through finite shocks, which cut

p_{0}

irreversibly. Compression in supersonic flow is always lossy; smooth expansion is essentially free.

A normal shock and an oblique shock at the same upstream Mach number cause the same stagnation-pressure loss.

False — the loss law depends on

M_{1 n} = M_{1} sin β

, and an oblique shock (

β < 90°

) has a smaller

M_{1 n}

than a normal shock (

β = 90°

, so

M_{1 n} = M_{1}

). Smaller normal component → smaller

p_{0}

drop, which is why nature prefers oblique shocks to compress.

Making the nozzle area ratio larger always increases thrust.

False — it is true only in vacuum. The Nozzle Area Ratio fixes how far the gas expands, hence

p_{e}

: a larger ratio lowers

p_{e}

. At low altitude that drops

p_{e}

further below

p_{a}

, deepening over-expansion and its shock losses; there is an optimum ratio for each altitude (see Altitude Compensation).

Shock diamonds are decorative and harmless.

False — each bright diamond marks one shock/expansion cycle, and every shock (especially the central Mach disk, a small normal shock on the axis) bleeds

p_{0}

and converts ordered kinetic energy into heat.

Flow separation is guaranteed the instant

p_{e}

drops even slightly below

p_{a}

False — mild over-expansion is tolerated. The thin, slow boundary layer on the wall can only climb a limited adverse pressure jump before it detaches; empirical criteria (e.g. Summerfield) put detachment near

p_{e} / p_{a} \approx 0.3 - 0.4

because that is where the required pressure rise exceeds what the near-wall low-momentum gas can push against. Above that ratio the layer stays attached.

The velocity coefficient

C_{v}

can exceed 1 for a severely over-expanded nozzle.

False —

C_{v} = p_{0, actual} / p_{0, chamber}

and stagnation pressure can only fall or stay equal through real flow, never rise, so

C_{v} \leq 1

always. Over-expansion pushes it down toward 0.8 or worse.

Spot the error

p_{e} < p_{a}

means the nozzle throat is choked too hard, so we should widen the throat."

The throat is unrelated to the over-expansion; over-expansion is set by the diverging section's area ratio, which fixed

p_{e}

too low. Widening the throat changes mass flow

\overset{m}{˙}

, not the exit-to-ambient pressure match.

"The oblique shocks bend the flow outward, spreading the plume."

They bend it inward. The ambient air squeezes from the sides, so the compression turns the streamlines toward the centerline — that inward turning is exactly the deflection angle

θ

in the figure.

"Since the pressure-thrust term is negative, the rocket experiences net backward thrust."

No — the momentum term

\overset{m}{˙} v_{e}

still dominates for any real engine, so net thrust

F

stays forward. The negative pressure term merely reduces the positive total, it does not reverse it.

"Because shocks raise pressure, they raise stagnation pressure too, so

p_{0}

increases."

Static pressure

p

rises but stagnation pressure

p_{0}

falls across every shock — some ordered energy became heat. The

p_{0}

drop is the true measure of the irreversible loss; confusing

p

with

p_{0}

is the classic trap.

"A vacuum engine tested at sea level just makes a bit less thrust — no big deal."

The pressure-thrust penalty

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

can be tens of kilonewtons, and stacked shock losses can gut

C_{v}

down toward 0.78. Severe over-expansion can also separate the flow and physically shake the nozzle apart.

"We use oblique-shock relations because the flow is subsonic and turning."

Shocks only exist in supersonic flow (

M_{e} > 1

); a subsonic exit adjusts smoothly with no shock at all. The whole shock discussion presupposes

M_{e} > 1

from Isentropic Flow through the diverging cone.

"The shock angle

β

is larger than the deflection angle

θ

only by coincidence."

It is structural, not coincidental: the shock must be steep enough to compress the flow, yet the streamlines turn less than the shock line, so

β > θ

always for an attached oblique shock. The geometry is fixed by Gas Dynamics, not luck.

Why questions

Why does the atmosphere compress an over-expanded plume instead of the plume just pushing the air aside?

Because

p_{a} > p_{e}

: the higher outside pressure wins the force balance at the plume boundary and pushes inward, and a supersonic flow can only respond to that inward push through shock waves.

Why must the compression happen through a shock rather than a smooth pressure rise?

In supersonic flow the gas moves faster than pressure signals (which travel at the sound speed), so information cannot propagate upstream to "warn" the flow — the only way to raise pressure is a discontinuous shock.

Why do the shocks appear oblique (angled) rather than normal (perpendicular)?

The compression comes from the sides (a lateral squeeze at the nozzle lip), and an angled shock only needs to process the flow's normal component

M_{1 n} = M_{1} sin β

, which is nature's lower-entropy, cheaper compression route.

Why does a Mach disk form on the centerline of the plume?

The oblique shocks from opposite lips travel inward and meet at the axis; where they cross, a single continued oblique turn would over-constrain the flow, so a small normal shock (the Mach disk) resolves it, resetting the flow to subsonic locally.

Why does over-expansion get worse as a rocket climbs early then better higher up?

At sea level

p_{a}

is highest, so a fixed low

p_{e}

is furthest below ambient — deepest over-expansion. As altitude rises,

p_{a}

falls toward

p_{e}

, the mismatch shrinks, and eventually

p_{e} = p_{a}

(matched) then

p_{e} > p_{a}

(under-expanded).

Why do multiple shocks compound the loss rather than average out (in terms of

p_{0}

Because

p_{0}

ratios are multiplicative across successive shocks, not additive; each shock removes a fixed fraction of what remains, so

N

shocks give

(p_{0, 2} / p_{0, 1})^{N}

, an ever-shrinking product that can never recover.

Why is under-expansion the "safer" off-design failure for engine hardware?

Under-expansion produces expansion fans (reversible, no separation) and the plume simply bulges outward, whereas over-expansion can separate the flow inside the bell, causing asymmetric side-loads that can crack the nozzle. This is why altitude-compensating designs (see Altitude Compensation) are prized.

Why does firing a vacuum-optimized engine at sea level punish you twice?

Once through the negative pressure-thrust term

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

(raw force lost to the pressure mismatch), and again through the stacked shock losses that drag

C_{v}

down — the momentum thrust itself is degraded on top of the pressure penalty.

Edge cases

What is

p_{e} - p_{a}

when the engine fires in a perfect vacuum?

Then

p_{a} = 0

, so the pressure term becomes

+ p_{e} A_{e}

, always positive — a vacuum-designed nozzle is never over-expanded in space, which is exactly what it was built for.

Can a nozzle be over-expanded at one altitude and under-expanded at another during a single ascent?

Yes — a fixed-geometry nozzle has fixed

p_{e}

, so as

p_{a}

falls with altitude it passes from

p_{e} < p_{a}

(over) through

p_{e} = p_{a}

(matched) to

p_{e} > p_{a}

(under). One design point cannot be optimal everywhere, motivating Rocket Staging and altitude compensation.

At exactly

p_{e} = p_{a}

, do any oblique shocks form at the lip?

No — with perfect matching there is no pressure jump to accommodate, so the plume leaves as a clean parallel column with neither shocks nor expansion fans. This is the ideal, loss-free boundary between the two off-design regimes.

What happens in the limit of extreme over-expansion,

p_{e} / p_{a} \to 0

The adverse pressure ratio becomes so violent that the boundary layer cannot stay attached; the flow separates inside the nozzle, the effective exit shifts upstream, and the internal shock structure and side-loads can destroy the hardware.

If the exit flow were subsonic (

M_{e} < 1

), what would over-expansion look like?

There would be no shocks at all — subsonic flow adjusts its pressure smoothly to ambient before leaving. Shock-based over-expansion losses are unique to supersonic exits, so the whole plume-shock story only applies when

M_{e} > 1

For a single oblique shock, what is the limiting deflection

θ

β

shrinks toward the Mach angle?

β \to arcsin (1/ M_{1})

the normal component

M_{1} sin β \to 1

, the shock weakens to a Mach wave, and the deflection

θ \to 0

— an infinitely weak shock turns the flow by nothing and causes essentially no loss.

Recall One-line self-test before you close this page

Name the single inequality that defines over-expansion, and the single physical process that makes it lossy. Answer ::: $p_{e} < p_{a}$ defines it; irreversible shock compression ( $p_{0}$ drop, entropy rise) makes it lossy.