Intuition What this page is for
The vis-viva equation v 2 = GM ( r 2 − a 1 ) looks like one formula, but the shape of the orbit (a positive, infinite, or negative) makes it behave in wildly different ways. This page hunts down every kind of orbit and every trap the formula can spring, and works each one out from scratch. By the end you will never meet a vis-viva scenario you haven't already seen.
Before we start, three plain-language reminders so no symbol is unearned:
Definition The three symbols we reuse everywhere
r = your current distance from the center of the big body M (changes moment to moment).
a = the semi-major axis : half the longest width of the whole orbit — a fixed number for the orbit.
GM = the standard gravitational parameter μ : how strong the pull is. For Earth G M ⊕ = 398600 km 3 / s 2 .
We will always work in km and seconds , so GM in km 3 / s 2 keeps every answer in km/s. Mixing metres and km is the #1 silent error — we avoid it by picking km once and never switching.
Every vis-viva problem lives in one of these cells. The examples below are tagged with the cell they cover.
Cell
Orbit type
Sign / value of a
What's special
Example
C1
Circular
a = r (finite > 0 )
r 2 − a 1 = r 1 ; speed constant
Ex 1
C2
Elliptical, at perigee
a > 0 , r = r p < a
fastest point
Ex 2
C3
Elliptical, at apogee
a > 0 , r = r a > a
slowest point
Ex 2
C4
Elliptical, general point
a > 0 , r p < r < r a
in-between speed
Ex 3
C5
Parabolic (escape boundary)
a → ∞ , a 1 → 0
ε = 0 exactly
Ex 4
C6
Hyperbolic (unbound)
a < 0
v 2 > 2 GM / r , excess speed
Ex 5
C7
Solve backwards for a
given v , r → find orbit size
classifies the orbit
Ex 6
C8
Real-world word problem
mission Δ v
Hohmann transfer
Ex 7
C9
Exam twist / degenerate
r → r at a → r limit, sign trap
catch the mistake
Ex 8
Worked example Ex 1 — Circular orbit (Cell C1)
A GPS satellite circles Earth at radius r = 26560 km. How fast is it going?
Forecast: Guess first — will it be faster or slower than the ∼ 7.7 km/s of a low satellite? (Higher orbit → slower.)
Step 1. Recognise circular ⇒ a = r .
Why this step? A circle is an orbit whose "longest width" is its diameter, so half of that, a , equals the radius r . This is the only case where r and a are the same number.
Step 2. Substitute a = r into vis-viva:
v 2 = GM ( r 2 − r 1 ) = r GM .
Why this step? The two fractions collapse because they share the same denominator — that's the algebraic fingerprint of "circular means constant speed."
Step 3. Plug numbers:
v = 26560 398600 = 15.008 ≈ 3.874 km/s .
Verify: Units: km 3 / s 2 ÷ km = km 2 / s 2 = km/s ✓. And 3.87 < 7.7 km/s — slower than LEO, exactly as a bigger orbit should be. ✓
Worked example Ex 2 — Ellipse at both apsides (Cells C2 & C3)
An elliptical orbit has perigee r p = 7000 km and apogee r a = 42000 km. Find the speed at perigee (closest) and at apogee (farthest).
Forecast: Which will be bigger, v p or v a ? (Whip-fast when close, drift-slow when far — so v p > v a .)
Step 1. Find a from the apsides.
Why this step? Vis-viva always needs a , and the two apsides sit at opposite ends of the long axis, so their distances add to the full length 2 a :
a = 2 r p + r a = 2 7000 + 42000 = 24500 km .
Step 2. Speed at perigee, r = r p = 7000 km:
v p 2 = 398600 ( 7000 2 − 24500 1 ) = 398600 ( 2.8571 × 1 0 − 4 − 4.0816 × 1 0 − 5 ) .
Why this step? We just drop r = r p and the fixed a into the formula — nothing else changes.
v p 2 = 398600 × 2.4490 × 1 0 − 4 = 97.62 ⇒ v p ≈ 9.88 km/s .
Step 3. Speed at apogee, r = r a = 42000 km, same a :
v a 2 = 398600 ( 42000 2 − 24500 1 ) = 398600 ( 4.7619 × 1 0 − 5 − 4.0816 × 1 0 − 5 ) .
v a 2 = 398600 × 6.803 × 1 0 − 6 = 2.712 ⇒ v a ≈ 1.647 km/s .
Verify: Angular momentum must match: r p v p = 7000 × 9.88 = 6.92 × 1 0 4 and r a v a = 42000 × 1.647 = 6.92 × 1 0 4 . Equal ✓ (this is Conservation of Angular Momentum in action). And v p > v a as forecast. ✓
Worked example Ex 3 — Ellipse at a general in-between point (Cell C4)
Same orbit as Ex 2 (a = 24500 km). What is the speed when the satellite is at r = 20000 km — somewhere between perigee and apogee?
Forecast: 20000 km is between 7000 and 42000 , so expect a speed between 9.88 and 1.65 km/s.
Step 1. Reuse the same a = 24500 km.
Why this step? a describes the whole orbit; it does not change just because we moved. Only r changes.
Step 2. Substitute r = 20000 :
v 2 = 398600 ( 20000 2 − 24500 1 ) = 398600 ( 1.0000 × 1 0 − 4 − 4.0816 × 1 0 − 5 ) .
v 2 = 398600 × 5.918 × 1 0 − 5 = 23.59 ⇒ v ≈ 4.857 km/s .
Verify: 1.65 < 4.86 < 9.88 — inside the perigee/apogee range as forecast ✓. Units km/s ✓.
Worked example Ex 4 — Parabolic escape, the
a → ∞ limit (Cell C5)
A probe sits at r = 7000 km. What speed does it need to just barely escape Earth (reach infinity with zero speed left)?
Forecast: Escape speed is famously 2 times circular speed — so guess about 1.41 × the circular speed at this radius.
Step 1. "Just barely escape" means the orbit is parabolic , i.e. a → ∞ .
Why this step? An escaping-but-not-overshooting path has total specific energy ε = 0 (the exact boundary between bound and free). Since ε = − GM / ( 2 a ) , setting ε = 0 forces a → ∞ , so a 1 → 0 .
Step 2. Drop the a 1 term:
v esc 2 = GM ( r 2 − 0 ) = r 2 GM = 7000 2 × 398600 = 113.89.
v esc ≈ 10.67 km/s .
Verify: Circular speed here is 398600/7000 = 56.94 = 7.546 km/s, and 2 × 7.546 = 10.67 km/s ✓ — matches Escape Velocity exactly.
Worked example Ex 5 — Hyperbolic orbit, the
a < 0 trap (Cell C6)
An interstellar object flies past the Sun on a hyperbolic orbit with a = − 50 × 1 0 6 km (negative — it is unbound). At r = 150 × 1 0 6 km (1 AU), how fast is it moving? Use G M ⊙ = 1.327 × 1 0 11 km 3 / s 2 .
Forecast: For a hyperbola v 2 > 2 GM / r (escape speed and then some). So expect a speed above the local escape speed.
Step 1. Keep a negative — do not flip its sign.
Why this step? The convention "a < 0 for hyperbolas" is exactly what makes − a 1 turn into + ∣ a ∣ 1 , adding energy so the speed exceeds escape. If you "fix" the sign you destroy the physics.
Step 2. Substitute with a 1 = − 50 × 1 0 6 1 = − 2.0 × 1 0 − 8 :
v 2 = 1.327 × 1 0 11 ( 150 × 1 0 6 2 − ( − 2.0 × 1 0 − 8 ) ) .
= 1.327 × 1 0 11 ( 1.3333 × 1 0 − 8 + 2.0 × 1 0 − 8 ) = 1.327 × 1 0 11 × 3.3333 × 1 0 − 8 .
v 2 = 4423.3 ⇒ v ≈ 66.51 km/s .
Verify: Local escape speed is 2 GM / r = 2 × 1.327 × 1 0 11 / ( 150 × 1 0 6 ) = 1769.3 = 42.06 km/s. Our 66.5 > 42.06 ✓ — genuinely hyperbolic, as forecast.
Worked example Ex 6 — Solve backwards for
a (Cell C7)
A spacecraft at r = 8000 km is measured moving at v = 9.0 km/s. What is the semi-major axis a , and is the orbit bound or unbound?
Forecast: 9.0 km/s — is that above or below escape at 8000 km? Escape is 2 × 398600/8000 = 9.98 km/s. We're below it, so the orbit should be bound (positive a ).
Step 1. Rearrange vis-viva for a 1 :
a 1 = r 2 − GM v 2 .
Why this step? We know v and r ; the only unknown is a , so isolate it. This is the "diagnose the orbit" move.
Step 2. Plug in:
a 1 = 8000 2 − 398600 9. 0 2 = 2.5000 × 1 0 − 4 − 2.0321 × 1 0 − 4 = 4.679 × 1 0 − 5 .
a = 4.679 × 1 0 − 5 1 ≈ 21372 km .
Step 3. Classify: a > 0 and finite ⇒ bound ellipse .
Verify: Recompute v from this a : v 2 = 398600 ( 2/8000 − 1/21372 ) = 398600 ( 2.5 × 1 0 − 4 − 4.679 × 1 0 − 5 ) = 398600 × 2.032 × 1 0 − 4 = 81.00 , so v = 9.00 km/s ✓ round-trips. And a > 0 matches the bound forecast. See Orbital Elements .
Worked example Ex 7 — Real-world word problem: Hohmann
Δ v (Cell C8)
A satellite is in a circular orbit at r 1 = 7000 km. We want to raise it to a circular orbit at r 2 = 42000 km using a Hohmann Transfer Orbit . Find the first burn Δ v 1 — the speed boost needed to leave the low circle and enter the transfer ellipse.
Forecast: The transfer ellipse's perigee sits on the low circle, and its perigee speed (from Ex 2, 9.88 km/s) is faster than the low circular speed. So Δ v 1 is a modest positive number.
Step 1. Low circular speed (Cell C1 machinery) at r 1 = 7000 :
v c 1 = 7000 398600 = 56.943 = 7.546 km/s .
Why this step? Before the burn the satellite moves at circular speed; that's our starting point.
Step 2. Transfer-ellipse perigee speed. The transfer ellipse has r p = 7000 , r a = 42000 ⇒ a = 24500 km — identical to Ex 2. So its perigee speed is v p = 9.883 km/s (from Ex 2).
Why this step? The burn happens at perigee of the transfer ellipse, which coincides with the low circle, so we need that ellipse's perigee speed.
Step 3. The burn just makes up the difference:
Δ v 1 = v p − v c 1 = 9.883 − 7.546 = 2.337 km/s .
Verify: Δ v 1 > 0 (a speed-up to climb higher) ✓, and it's a realistic few-km/s figure for such a transfer ✓. Units km/s ✓.
Worked example Ex 8 — Exam twist: the degenerate / sign trap (Cell C9)
"A student computes the speed of a satellite at r = 30000 km on an orbit with a = 24500 km and gets v 2 = 398600 ( 2/30000 − 1/24500 ) . They panic because the two fractions are close. Then a second student, given a hyperbolic orbit with a = 24500 km (they forgot the sign), computes the same way. What's wrong, and what's the correct speed in the first (legitimate) case?"
Forecast: The first calculation is fine (bound ellipse, r < r a = ? ). The second is the classic sign mistake. Guess: legit answer is a couple of km/s.
Step 1. First student — check the orbit is real. Here a = 24500 is the Ex 2 orbit, whose apogee is r a = 2 a − r p = 2 ( 24500 ) − 7000 = 42000 km. Since 7000 ≤ 30000 ≤ 42000 , the point r = 30000 actually exists on the orbit. Good.
Why this step? Vis-viva silently returns a value even for an r the orbit never reaches; you must confirm r p ≤ r ≤ r a first, or the "speed" is fictitious.
Step 2. Compute:
v 2 = 398600 ( 30000 2 − 24500 1 ) = 398600 ( 6.6667 × 1 0 − 5 − 4.0816 × 1 0 − 5 ) = 398600 × 2.585 × 1 0 − 5 .
v 2 = 10.305 ⇒ v ≈ 3.210 km/s .
Step 3. Second student's error: for a hyperbola a must be negative . Using a = + 24500 describes a bound ellipse , not the unbound path they meant — the sign is not cosmetic, it flips − a 1 from adding energy to removing it. The fix is a = − 24500 , which would give a much larger v 2 = 398600 ( 2/30000 + 1/24500 ) .
Verify: First case: 3.21 km/s sits between perigee 9.88 and apogee 1.65 ? No — wait, 3.21 < 1.65 is false; 1.65 < 3.21 < 9.88 ✓ (it's between, closer to apogee since 30000 is near the far side). Consistent ✓.
Recall Which cell am I in? (quick reflex)
Given an orbit, first name its type ::: circle→a = r ; ellipse→0 < a finite; parabola→a → ∞ ; hyperbola→a < 0 .
Given v and r , how do I find the type ::: compute 1/ a = 2/ r − v 2 / GM ; sign of a tells you (+ bound, ∞ parabolic, − hyperbolic).
Before trusting a computed v on an ellipse, what must I check ::: that r p ≤ r ≤ r a , else the point isn't on the orbit.
a is the shape of the orbit
Plus = penned in (ellipse, bound). Infinity = the edge (parabola, escape). Minus = missile (hyperbola, flies off). Never "fix" a negative a — it's telling you the truth.