Propulsive forces — thrust misalignment, gimbal angle: Question bank

Before you start, the symbols we'll reuse (all from the parent):

$T$ = thrust magnitude (how hard the engine pushes), in newtons.
$L$ = distance from the center of mass (CoM) to the engine's pivot point — the lever arm, in metres.
$δ$ = the tilt of the engine away from the rocket's long axis — the gimbal angle, in radians.
$τ$ = torque, the "twisting" push that changes how fast the rocket rotates, in newton-metres.
$I$ = moment of inertia about the pitch axis — how hard the rocket is to spin up, the rotational analogue of mass, in $kg\cdotm^{2}$ . It appears in the rotational Newton's law $τ = I \ddot{θ}$ , so $\ddot{θ} = τ / I$ . See Torque and Moment of Inertia.
$r$ = the position vector pointing from the CoM to the engine pivot (magnitude $L$ ), used in the torque cross product $τ = r \times F$ .
$d$ = a lateral offset: how far sideways (perpendicular to the rocket axis) the thrust line sits from the CoM, in metres — a sideways displacement, not an angle.

Recall Why the forward loss is

$\frac{1}{2} T δ^{2}$ (Taylor sketch) When we tilt by $δ$ , the forward-pushing part of thrust is $T cos δ$ , so the lost part is $T (1 - cos δ)$ . Near $δ = 0$ the cosine curve is nearly flat, and its Taylor expansion is $cos δ \approx 1 - \frac{δ ^{2}}{2} + \dots$ Substituting gives loss $\approx T (1 - (1 - \frac{δ ^{2}}{2})) = \frac{1}{2} T δ^{2}$ . The first-order term is absent — that is exactly why the loss is "second-order small" while the torque $L T sin δ \approx L T δ$ keeps its first-order term.

True or false — justify

An engine mounted exactly at the center of mass produces zero torque no matter how it tilts.

True in this idealized case — with

L = 0

the lever arm vanishes, so

τ = L T sin δ = 0

. But real engines sit behind the CoM, so

L \neq = 0

and this situation almost never occurs. See Center of Mass shift during burn.

Gimballing the engine to steer always costs a large fraction of forward thrust.

False. The forward loss is

T (1 - cos δ) \approx \frac{1}{2} T δ^{2}

, which is second-order (tiny) at the few-degree tilts used in practice — at

5^{\circ}

only about

0.4%

is lost.

A gimbal angle of

δ

and a lateral thrust offset

d

produce torque by the same formula.

False. A tilt gives

τ = L T sin δ

(and an axial loss); a pure offset gives

τ = d T

with thrust still fully axial. Different geometry, different formula.

Doubling the gimbal angle roughly doubles the steering torque.

True at small angles, because

τ \approx L T δ

is linear in

δ

. Doubling

δ

nearly doubles

τ

while the forward loss barely changes (it scales as

δ^{2}

A steady held gimbal angle produces a steady rotation rate.

False. A steady

δ

gives steady angular acceleration

\ddot{θ} = τ / I

(with

I

the pitch moment of inertia), so the rotation rate keeps growing. See Torque and Moment of Inertia.

0. 5^{\circ}

manufacturing misalignment is negligible because the angle is tiny.

False. It creates an always-on disturbance torque the controller must continuously fight, burning gimbal authority and propellant. Small angle, persistent consequence.

The forward thrust loss from tilting is a first-order effect in

δ

False. The loss is

T (1 - cos δ) \approx \frac{1}{2} T δ^{2}

, which is second-order (its first-order Taylor term is zero). Only the torque is first-order — that asymmetry is why TVC works. See Thrust Vector Control (TVC).

Radians versus degrees only matters for exact answers, not for understanding.

False. Using degrees inside

sin δ

or in

\ddot{θ} = τ / I

gives numerically wrong results (off by a factor of

\sim 57

), because the small-angle rule and rotational dynamics are defined in radians.

Spot the error

"Thrust pushes the whole rocket as a single point, so its direction can't cause spin."

Wrong: thrust is applied at the nozzle pivot, a specific point behind the CoM. Any component perpendicular to the CoM-pivot line acts on lever arm

L

and twists the vehicle.

"Since

τ = r \times F

, and the engine sits on the axis,

r

is zero."

Wrong:

r

(defined above) points from the CoM to the pivot and has magnitude

L \neq = 0

. It is the lateral thrust component

T sin δ

(perpendicular to

r

) that survives the cross product, giving

τ = L T sin δ

"The small-angle form

τ \approx L T δ

works because

sin

and the angle are literally equal."

Wrong: they are not equal, only approximately so for small

δ

in radians (

sin δ

and

δ

agree to under

0.6%

1 0^{\circ}

). It's a controlled approximation that keeps control math linear.

"To get more torque without wasting thrust, mount the engine at a big fixed offset

d

Wrong reasoning: a fixed offset

d

(a sideways displacement of the thrust line) gives a constant unwanted torque

d T

that the controller must always fight — that's a disturbance, not steering. Steering needs a commanded, adjustable gimbal, not a permanent offset.

"As fuel burns,

L

stays fixed since the engine doesn't move."

Wrong: the engine pivot is fixed to the structure, but the CoM drifts forward as propellant depletes, so

L

(CoM-to-pivot) grows. Control authority changes through the burn. See Center of Mass shift during burn.

"The minus sign in

τ = - L T sin δ

means the torque is somehow negative energy."

Wrong: the sign is just a rotation direction convention (which way it pitches) in the chosen coordinate frame, not a magnitude or an energy statement.

Why questions

Why do we deliberately tilt the engine instead of using fins or reaction wheels alone at liftoff?

Near liftoff the air is thick and the rocket slow, so fins have little grip; gimballing the main engine gives large, immediate control torque exactly when it's needed. See Attitude Control & Stability.

Why is gimballing described as "turning a bug into a feature"?

Misalignment (a bug) is unwanted thrust-angle torque; gimballing uses that same physics on purpose (a feature) by commanding the tilt to steer instead of suffering it.

Why does a first-order torque with second-order loss make TVC so efficient?

You gain torque proportional to

δ

but pay only a

δ^{2}

forward-thrust penalty — so at small angles you buy a lot of steering for almost no performance cost.

Why must the guidance computer keep changing

δ (t)

rather than set it once?

A fixed

δ

gives constant angular acceleration, which would spin the rocket ever faster; continuous adjustment lets it null out the rate and hold or track the desired attitude.

Why does Newton's Third Law appear at all here?

Thrust exists because the rocket throws mass backward and the exhaust pushes the rocket forward — the direction of that reaction force is what tilting the nozzle redirects. See Newton's Third Law.

Why can a bigger moment of inertia

I

actually be helpful for stability?

For a given torque,

\ddot{θ} = τ / I

is smaller when

I

(the pitch moment of inertia) is larger, so disturbances build up angular rate more slowly, giving the controller more time to respond.

Edge cases

What is the torque and thrust loss when

δ = 0

Zero torque (

sin 0 = 0

) and zero loss (

cos 0 = 1

): the engine points straight, all thrust is axial, no steering — the "flying straight" baseline.

What happens to torque and loss at

δ = 9 0^{\circ}

(engine fully sideways)?

Torque is maximal (

sin 9 0^{\circ} = 1

τ = L T

) but axial thrust is zero (

cos 9 0^{\circ} = 0

) — pure turning, no forward push. Physically impossible for a real gimbal but the mathematical extreme of the trade-off.

T = 0

(engine off), can any gimbal angle produce torque?

No:

τ = L T sin δ = 0

when

T = 0

. With no thrust there is no force to act on the lever arm, so a tilted-but-dead engine steers nothing.

For a negative gimbal angle (tilt the other way), how does the torque respond?

It reverses sign:

sin (- δ) = - sin δ

, so

τ

flips direction while its magnitude and the (even) thrust loss

\frac{1}{2} T δ^{2}

stay the same — this is how the rocket steers both ways.

In the limit

δ \to 0

, which effect dominates: the torque or the thrust loss?

The torque, because it shrinks linearly (as

δ

) while the loss shrinks quadratically (as

δ^{2}

); for very small angles the loss becomes utterly negligible next to the still-useful torque.

If the CoM ever ended up behind the pivot (unusual), what happens to the sign of

L

The lever-arm vector

r

reverses, flipping the sign of the torque for the same gimbal direction — the same commanded tilt would now pitch the rocket the opposite way, a genuine control-inversion hazard.

Recall One-line takeaway

Torque is linear in the tilt, thrust loss is squared in the tilt, and thrust always acts at the pivot a distance $L$ behind the CoM — every trap on this page is a consequence of those three facts.