Optimal staging — equal mass ratios (for same Isp): Question bank

Before we start, pin down every symbol these items reuse:

Definition The cast of characters

$R_{i}$ — mass ratio of stage $i$ : mass at ignition ÷ mass at burnout. Always $R_{i} \geq 1$ (you can only lose mass by burning). It is the "leverage" of the stage.
$R^{*}$ — the common optimal mass ratio: when the theorem applies, all stages share one best value, written $R_{1} = R_{2} = \dots = R^{*}$ .
$ϵ$ — structural fraction: structure ÷ (structure + fuel) for a stage. The dead weight you are forced to carry per kilogram of tankage. See Structural fraction.
$v_{e}$ — exhaust velocity; $Δ v = v_{e} ln R$ per stage (that is Tsiolkovsky). Same $v_{e}$ = same engine tech for all stages.
$m_{p}$ — payload mass: the useful cargo you actually want to deliver (top of the stack). See Payload fraction.
$M_{0}$ — initial total mass: the whole rocket on the pad — payload plus every stage's structure and fuel.
$m_{s i}, m_{f i}$ — the ==structure mass and fuel mass of stage $i$ == (so $ϵ = \frac{m _{s i}}{m _{s i} + m _{f i}}$ ).
$n$ — number of stages.
$k$ — shorthand for the per-stage constraint factor $k \equiv (\frac{m _{p}}{M _{0}})^{1/ n}$ ; it is the $n$ -th root of the payload fraction, and the boxed result reads $R^{*} = \frac{1}{ϵ + ( 1 - ϵ ) k}$ .

Recall Where does the Lagrange condition come from?

We maximize $f = \sum_{i} ln R_{i}$ (total $Δ v / v_{e}$ ) subject to the constraint $g = \sum_{i} ln \frac{1 - ϵ R _{i}}{( 1 - ϵ ) R _{i}} = ln \frac{m _{p}}{M _{0}}$ . Lagrange multipliers says at the optimum $\nabla f = λ \nabla g$ , i.e. $\partial f / \partial R_{i} = λ \partial g / \partial R_{i}$ for every $i$ . Differentiating: $\frac{\partial f}{\partial R _{i}} = \frac{1}{R _{i}}$ and $\frac{\partial g}{\partial R _{i}} = \frac{- ϵ}{1 - ϵ R _{i}} - \frac{1}{R _{i}}$ , giving the stated condition. See the parent note's [!formula] box for the full walk-through.

True or false — justify

The theorem says equal mass ratios are optimal for stages that share the SAME

ϵ

and SAME

v_{e}

True — the Lagrange condition

\frac{1}{R _{i}} = λ (\frac{- ϵ}{1 - ϵ R _{i}} - \frac{1}{R _{i}})

is a function of

R_{i}

alone only because

ϵ

is common; that forces every

R_{i}

to solve the same equation.

If two stages have different

ϵ

, optimal staging still gives equal mass ratios.

False — the per-stage Lagrange equation now depends on

ϵ_{i}

, so the ratios differ. The stage with the smaller

ϵ

(less dead weight) is favoured with a larger

R

. See Variable specific impulse staging for the analogous coupling.

For

ϵ \to 0

the total

Δ v

is the same no matter how you split the fixed product

\prod R_{i}

True — the objective is

v_{e} ln (\prod R_{i})

and only the product is fixed, so every split gives identical

Δ v

. Equal ratios become strictly best only once

ϵ > 0

Adding more stages always strictly increases the achievable

Δ v

False — in pure theory (

ϵ \to 0

)

Δ v = v_{e} ln (M_{0} / m_{p})

is exactly independent of

n

, so more stages neither help nor hurt; and in practice real staging hardware adds structure, so beyond 3–4 stages the added dead weight makes

Δ v

eventually fall.

A single stage can always match

n

optimal stages if you just make its

R

big enough.

False — one stage's ratio is capped at

R_{ma x} = 1/ ϵ

(from

R = \frac{1}{ϵ + ( 1 - ϵ ) k}

k \to 0

). For

ϵ = 0.1

that ceiling is

10

, far below the

\sim 30

orbit needs.

The constraint linking payload and initial mass is simply

\prod R_{i} = M_{0} / m_{p}

False in general — that holds only at

ϵ \to 0

. The true constraint is

\frac{m _{p}}{M _{0}} = \prod \frac{1 - ϵ R _{i}}{( 1 - ϵ ) R _{i}}

, which is nonlinear in each

R_{i}

Because

ln

is concave, plain AM–GM proves equal ratios are optimal for any

ϵ

False — AM–GM works only when the product is fixed (

ϵ \to 0

). For

ϵ > 0

the constraint is not a fixed product, so you genuinely need the Lagrange multipliers argument.

Making

R_{1}

(first stage) huge and

R_{2}

tiny wastes performance when

ϵ > 0

True — an oversized lower stage carries structure it barely needs while the upper stage is starved; the nonlinear constraint punishes the imbalance, lowering

Δ v

below the equal-ratio value.

The payload fraction

m_{p} / M_{0}

shrinks as you demand more

Δ v

True — more

Δ v

needs larger

R^{*}

, and

\frac{m _{p}}{M _{0}} = (\frac{1 - ϵ R ^{*}}{( 1 - ϵ ) R ^{*}})^{n}

falls as

R^{*}

grows. See Payload fraction.

Spot the error

Δ v_{t o t a l} = v_{e} ln (R_{1} R_{2})

, and

R_{1} R_{2}

is fixed, so staging is pointless."

The product is fixed only when

ϵ \to 0

; with real

ϵ

the constraint is nonlinear and staging lets you shed structure mid-flight, raising the effective ratio past a single stage's ceiling.

"Stage 2's payload is just

m_{p}

No — for stage 1, stage 2 (its structure and remaining fuel) is part of the payload it must lift. Stage 2 alone treats

m_{p} + m_{s 2}

as its burnout mass.

"Structure mass

m_{s i} = ϵ m_{f i}

Wrong scaling —

ϵ = \frac{m _{s i}}{m _{s i} + m _{f i}}

, so

m_{s i} = \frac{ϵ}{1 - ϵ} m_{f i}

. For

ϵ = 0.1

that is

\frac{1}{9} m_{f i}

, not

0.1 m_{f i}

"Since

R^{*} = 1/ (ϵ + (1 - ϵ) k)

with

k = (m_{p} / M_{0})^{1/ n}

, letting

ϵ \to 1

gives a big ratio."

The opposite — as

ϵ \to 1

R^{*} \to 1

(all structure, no fuel), giving zero

Δ v

. Small

ϵ

is what allows large ratios.

"The mass handed to the next stage is

M_{i} = M_{i - 1} / R_{i}

That ignores dropping the spent structure. Correct:

M_{i} = M_{i - 1} \frac{1 - ϵ R _{i}}{( 1 - ϵ ) R _{i}}

, which is smaller than

M_{i - 1} / R_{i}

because structure is jettisoned too.

Δ v_{1} = v_{e} ln R_{1}

uses burnout mass

= m_{p} + m_{s 2}

No — stage 1 burns out to

m_{p} + m_{s 2} + m_{f 2} + m_{s 1}

(stage 2 still full, stage 1 structure still attached). Stage 1's structure is dropped after its burnout.

"Equal ratios means equal fuel mass in each stage."

Equal ratios, not equal masses — the lower stage moves a heavier vehicle, so

R_{1} = R_{2}

requires much more fuel in stage 1 than stage 2.

Why questions

Why does the same

ϵ

for all stages force equal

R_{i}

The Lagrange stationarity condition for each stage depends only on that stage's own

R_{i}

and the shared

ϵ

; identical equations have identical solutions, so all

R_{i}

coincide at

R^{*}

Why is

ln R

the right measure of a stage's "bang for the buck"?

Tsiolkovsky gives

Δ v = v_{e} ln R

, so velocity adds linearly in

ln R

; comparing

ln

values compares actual speed contributions, and

ln

's concavity is what rewards balancing.

Why can't a mass ratio ever be less than 1?

R =

ignition mass ÷ burnout mass, and burning fuel only removes mass, so the numerator is at least the denominator —

R \geq 1

, with equality meaning no fuel burned.

Why do total-

Δ v

formulas prefer the natural log over any other function?

Because momentum conservation on a variable-mass body integrates to

\int v_{e} d m / m = v_{e} ln (mass ratio)

— the log falls out of the physics, it is not a chosen convenience.

Why does staging beat a monolithic rocket even though the ideal

Δ v

is the same?

The monolithic rocket's

R

is capped at

1/ ϵ

by structure; staging shifts to a fresh smaller vehicle after jettison, so effective ratios compound (

R^{n}

) far past that ceiling.

Why does the advantage of equal ratios vanish at

ϵ \to 0

With no structure the constraint reduces to a fixed product

\prod R_{i}

, and

v_{e} ln (\prod R_{i})

then ignores the split entirely — every partition ties.

Edge cases

ϵ \to 0

: what does

R^{*}

become?

R^{*} = (M_{0} / m_{p})^{1/ n}

and

Δ v = v_{e} ln (M_{0} / m_{p})

, independent of

n

— the idealized "free lunch" limit where staging gives no ideal penalty or bonus.

ϵ \to 1

: what happens to a stage?

The stage is pure structure with no fuel;

R^{*} \to 1

, it contributes zero

Δ v

, and staging cannot help — it is dead weight only.

n = 1

(single stage): does the general formula still hold?

Yes —

\frac{m _{p}}{M _{0}} = \frac{1 - ϵ R ^{*}}{( 1 - ϵ ) R ^{*}}

, giving the ceiling

R_{ma x} = 1/ ϵ

m_{p} / M_{0} \to 0

; the multi-stage boxed result reduces to this at

n = 1

What if the required

R^{*}

from constraint

(⋆)

comes out below 1?

It signals an impossible mission — you are asking for more payload than the mass budget allows; no physical fuel split reaches it, since

R < 1

is forbidden.

m_{p} \to 0

(throw-away probe, no payload): what is

R^{*}

k = (m_{p} / M_{0})^{1/ n} \to 0

, so

R^{*} \to 1/ ϵ

per stage — each stage hits its structural ceiling, the maximum a real stage can deliver.

m_{p} \to M_{0}

(payload nearly the whole rocket):

k \to 1

, so

R^{*} \to 1

and

Δ v \to 0

— almost no fuel, almost no acceleration, the degenerate "already at your mass budget" case.

Two stages with

R_{1} = 8, R_{2} = 2

R_{1} = R_{2} = 4

ϵ = 0

: which wins?

Neither — both give

v_{e} ln 16

identically, because at

ϵ = 0

only the product

R_{1} R_{2} = 16

matters, not the split.

Recall Quick self-test

The single common quantity that must match across stages for equal ratios to be optimal ::: the structural fraction $ϵ$ (with the same $v_{e}$ ). The real constraint that replaces " $\prod R_{i}$ fixed" when $ϵ > 0$ ::: $\frac{m _{p}}{M _{0}} = \prod \frac{1 - ϵ R _{i}}{( 1 - ϵ ) R _{i}}$ . The definition of the constraint factor $k$ ::: $k = (m_{p} / M_{0})^{1/ n}$ , the $n$ -th root of the payload fraction. The hard ceiling on a single stage's mass ratio ::: $R_{ma x} = 1/ ϵ$ .