Third-body perturbations: Question bank — Physics deep dive

Definition Symbols used on this page (read first)

Every quantity below appears throughout the questions — pin them down before you start.

$r$ = the satellite's position as seen from Earth's center (the arrow from Earth to the satellite). Its length is $r = ∣ r ∣$ . This is a position vector magnitude, not the physical size of the spacecraft.
$d$ = the third body's position from Earth's center (arrow from Earth to Moon/Sun). Its length is $d = ∣ d ∣$ .
$\hat{d} = d / d$ = the unit vector pointing from Earth toward the third body (length 1, direction only).
$s = d - r$ = the arrow from the satellite to the third body; length $s = ∣ s ∣$ . This is the distance the naive "raw pull" $μ_{3} / s^{2}$ would use.
$μ_{3} = G m_{3}$ = the third body's gravitational parameter. For the specific bodies we write $μ_{M}$ (Moon) and $μ_{S}$ (Sun) — these are just $μ_{3}$ with the Moon's or Sun's mass plugged in. Similarly $d_{M}, d_{S}$ are the Earth–Moon and Earth–Sun distances.

Picture: Earth at the origin, $d$ a long arrow to the Moon, $r$ a short arrow to the satellite, and $s$ closing the triangle from satellite to Moon. Because $r ≪ d$ , that short arrow barely tilts the long one — which is exactly why only the gradient survives.

Intuition Where the tidal coefficients

$+ 2$ and $- 1$ come from (heuristic) Look at the figure above. Earth sits in the Moon's field; the satellite sits a small step $r$ away. The Moon's pull on the satellite differs from its pull on Earth by roughly how much the field changes over that step — a gradient. Two effects combine:

Magnitude change (along $\hat{d}$ ): step toward the Moon and the pull strengthens; step away and it weakens. A $1/ d^{2}$ field changes twice as fast as the field itself (that is where the derivative of $d^{- 2}$ , giving $- 2 d^{- 3}$ , comes from) → a stretch of coefficient $+ 2$ along the line.
Direction change (across $\hat{d}$ ): step sideways and the pull re-aims slightly inward toward the Moon's line → a squeeze of coefficient $- 1$ perpendicular.

The Taylor expansion in the parent note is just this bookkeeping made exact: $(1 - 2 \hat{d} \cdot r / d)^{- 3/2} \approx 1 + 3 \hat{d} \cdot r / d$ produces the $3 (\hat{d} \cdot r) \hat{d}$ term, and subtracting the leftover $r$ yields $3 (\hat{d} \cdot r) \hat{d} - r$ : +2 along, −1 across, trace $2 - 1 - 1 = 0$ .

True or false — justify

The Moon perturbs a satellite by pulling it with force

μ_{M} / s^{2}

toward the Moon (where

s = ∣ d - r ∣

is the satellite-to-Moon distance).

False. That raw pull

μ_{M} / s^{2}

is almost entirely cancelled because the Moon pulls the whole Earth-frame with nearly the same

μ_{M} / d^{2}

; only the tiny difference across the Earth–satellite gap survives as a perturbation.

A perfectly rigid, point-like satellite sitting exactly at Earth's center would feel zero third-body perturbation.

True. At

r = 0

the tidal term

\frac{μ _{3}}{d ^{3}} (3 (\hat{d} \cdot r) \hat{d} - r)

vanishes, and in the exact form

\frac{d - r}{∣ d - r ∣ ^{3}} = \frac{d}{d ^{3}}

, so the two gravitational pulls are identical and the difference is exactly zero.

The Sun's third-body effect on Earth satellites is larger than the Moon's because the Sun is vastly more massive.

False. Tidal strength scales as

μ_{3} / d^{3}

; the Sun's ~390× greater distance cubes to a ~

6 \times 1 0^{7}

suppression, leaving the Moon (

μ_{M} / d_{M}^{3}

) about twice as strong as the Sun (

μ_{S} / d_{S}^{3}

Third-body perturbations weaken with distance to the third body as

1/ d^{2}

, just like gravity.

False. Gravity itself is

1/ d^{2}

, but the differential (tidal) effect is a gradient of that field, so it falls off as

1/ d^{3}

If you keep only the direct term

μ_{3} (d - r) /∣ d - r ∣^{3}

and drop the indirect term

- μ_{3} d / d^{3}

, you still get the correct perturbation for small satellites.

False. Without the indirect term the "perturbation" would not vanish for a satellite co-moving with Earth and would misbehave as

r \to 0

; the subtraction is what turns a raw field into a gradient.

The tidal acceleration always points toward the third body.

False. Along the line

\hat{d}

it stretches (points away on the far side, toward on the near side); perpendicular to that line it squeezes inward. Only the component structure, not a single direction, describes it.

LEO satellites experience no third-body force at all.

False. They experience one, it is just numerically swamped (~

1 0^{- 6} m/s^{2}

) by Earth's monopole (~~9 m/s²) and $J_{2}$ (~~

1 0^{- 2} m/s^{2}

), so it is negligible, not absent. See J2 perturbation and oblateness, where the same "compare accelerations" bookkeeping shows

J_{2}

outweighs luni-solar effects at low altitude.

Being in free-fall is the reason astronauts don't feel the Moon's pull, and the same reason the Moon's common pull doesn't perturb the orbit.

True. In a free-falling frame the uniform part of an external field is unfelt; both the astronaut and the perturbation analysis only "feel" the variation of that field across space.

Spot the error

"Since

a_{3 b} = μ_{3} (\frac{d - r}{∣ d - r ∣ ^{3}} - \frac{d}{d ^{3}})

and both terms are positive, the perturbation is always larger than either term alone."

The two terms are vectors that nearly cancel, not positive scalars that add. Their difference is much smaller than either — that near-cancellation is the whole point of "differential force".

"The tidal approximation

\frac{μ _{3}}{d ^{3}} (3 (\hat{d} \cdot r) \hat{d} - r)

is exact because it came from Newton's law."

It came from a first-order Taylor expansion in

r / d

; it drops

O ((r / d)^{2})

terms and is only accurate when

r ≪ d

"Expanding

(1 - 2 \frac{d \cdot r}{d ^{2}} + \frac{r ^{2}}{d ^{2}})^{- 3/2} \approx 1 - \frac{3}{2} \cdot \frac{r ^{2}}{d ^{2}}

The dominant small quantity is the linear term

- 2 \frac{d \cdot r}{d ^{2}}

, giving

+ 3 \frac{d \cdot r}{d ^{2}}

; the

r^{2} / d^{2}

term is second order and is discarded, not kept.

"The stretch coefficient along the third-body line is

+ 3 μ_{3} r / d^{3}

because of the factor 3 in

3 (\hat{d} \cdot r) \hat{d}

Along

\hat{d}

you get

3 (\hat{d} \cdot r) \hat{d} - r = 3 r_{∥} - r_{∥} = 2 r_{∥}

, so the coefficient is

+ 2 μ_{3} / d^{3}

, not

+ 3

"Because the Moon dominates, we can ignore the Sun entirely in stationkeeping."

The Sun is roughly half the Moon's effect — comparable, not negligible. Luni-solar perturbations together drive GEO stationkeeping North–South budgets; dropping the Sun mis-sizes the fuel.

"The perpendicular squeeze and the along-line stretch have equal magnitude."

They differ by a factor of 2: stretch is

+ 2 μ_{3} r / d^{3}

, squeeze is

- μ_{3} r / d^{3}

. The trace

2 - 1 - 1 = 0

of the tidal tensor forces exactly this 2:1 split. See Tidal forces, where the same 2:1 ratio explains why oceans raise two bulges, not four.

Why questions

Why is the perturbation a difference of accelerations rather than a single force?

Because we track the satellite in an Earth-centered frame that is itself accelerating toward the third body; subtracting Earth's acceleration removes the common pull and leaves the gradient.

Why does the inverse-cube law let a small nearby Moon beat a giant distant Sun?

The cube in

μ_{3} / d^{3}

punishes distance three times over; the Sun's ~390× farther reach costs a factor

\sim 6 \times 1 0^{7}

, which its ~

2.7 \times 1 0^{7}

-times-larger

μ

cannot fully recover.

Why does the tidal pattern stretch along the line to the body but squeeze across it?

The near side is pulled harder than the center and the far side less, stretching the two apart; sideways points are pulled slightly inward toward the center line, squeezing — the same geometry that raises two ocean bulges.

Why do high or eccentric orbits (GEO, GTO, Molniya) care about third-body effects far more than LEO?

The tidal term scales with

r

, so large satellite–Earth distances amplify it, while at the same time Earth's monopole (

\propto 1/ r^{2}

) and

J_{2}

have weakened — the competition tips toward the third body.

Why can't we just use the Two-body solution and call it a day for a GEO satellite over decades?

The tiny (

\sim 1 0^{- 6} m/s^{2}

) perturbation acts continuously and coherently, secularly precessing the orbit plane and drifting eccentricity; small accelerations integrated over years become large orbit changes, handled via Gauss and Lagrange planetary equations which turn this acceleration into slow drifts of the orbital elements.

Why does the indirect term

- μ_{3} d / d^{3}

appear even though it does not depend on where the satellite is?

It encodes the acceleration of the Earth frame itself; every object measured relative to Earth inherits this subtraction regardless of its own position.

Why is the full restricted three-body treatment sometimes needed instead of this tidal formula?

When

r

is not small compared to

d

(e.g. lunar transfers or near a Lagrange point) the

r / d

expansion breaks down and you must keep the exact geometry — see Restricted three-body problem.

Edge cases

What happens to the exact perturbation

a_{3 b}

as the satellite's Earth-distance

r = ∣ r ∣ \to 0

(the position magnitude, not the craft's physical size)?

It smoothly goes to zero:

\frac{d - r}{∣ d - r ∣ ^{3}} \to \frac{d}{d ^{3}}

, so the difference vanishes — a satellite at Earth's center shares Earth's exact acceleration.

What happens if the satellite lies exactly on the Earth–third-body line (so

\hat{d} \cdot r = \pm r

The tidal term becomes

\frac{μ _{3}}{d ^{3}} (3 r \hat{d} - r \hat{d}) = + 2 \frac{μ _{3}}{d ^{3}} r

: pure maximum stretch, no perpendicular component.

What happens if the satellite is exactly perpendicular to the third-body line (

\hat{d} \cdot r = 0

The tidal term is

- \frac{μ _{3}}{d ^{3}} r

: pure squeeze toward the center line, the perpendicular minimum.

What if

r

becomes comparable to

d

(a near-lunar-encounter trajectory)?

The first-order expansion is invalid; you must revert to the exact

a_{3 b}

or a full Restricted three-body problem model, since

O ((r / d)^{2})

terms are no longer small.

What is the third-body perturbation on a satellite when the Moon is at first quarter versus new moon — does it change?

Yes;

d

changes direction, so the stretch axis

\hat{d}

rotates and the whole tidal pattern reorients with the Moon's position, producing time-varying (partly periodic) perturbations.

What happens to the Sun-versus-Moon dominance for a spacecraft in heliocentric orbit far from Earth?

The comparison flips entirely — the framing assumes an Earth-centered frame; around the Sun the "third body" roles change and the

μ_{3} / d^{3}

bookkeeping must be redone for the new primary.

Recall One-line summary to keep

Third-body perturbation = difference of pulls (a gradient), scaling as == $μ_{3} / d^{3}$ ==; along the body-line $\hat{d}$ it stretches ( $+ 2 μ_{3} r / d^{3}$ , sign $+$ = outward from Earth), across it squeezes ( $- μ_{3} r / d^{3}$ , sign $-$ = inward toward the center line). See the figure above for the two arrows. Moon beats Sun ~2×; matters for high/eccentric orbits, negligible in LEO.