Potential Energy: Gravitational & Elastic

The crane from two modules ago lifted a 200 kg beam through 15 m at constant speed and did +30+30 kJ of work — yet the beam's kinetic energy never budged. The work–energy theorem's books balanced (+30+30 kJ cable, −30-30 kJ gravity, net zero), but a nagging question survives the audit: the crane burned real fuel. Did those 30,000 joules simply evaporate? Let the cable snap and you have your answer. The beam arrives at the ground doing 17 m/s, carrying — to the joule — 30 kJ of kinetic energy. The energy never left. It was banked, held not in the beam's motion but in its position, waiting for gravity to pay it back. That waiting energy is potential energy, and this module builds the two storage devices every exam draws on: height in a gravitational field, U=mghU = mgh, and deformation of a spring, U=12kx2U = \tfrac12 kx^2. Both are literal savings accounts — the deposit is the work done against gravity or against the spring, and the balance is refunded, in full, as kinetic energy on demand. You will also meet the strangest rule of the accounting: the zero of potential energy is yours to place. Only differences ever matter, which is why the same problem can quote U=+200U = +200 J or U=−300U = -300 J and predict identical physics. Master the two storage formulas, the sign discipline, and the freedom of the reference level, and the grand unification of the next modules — conservation of mechanical energy — becomes a single line.

Key takeaways
  • Potential energy is stored work: energy held in a system's configuration (how high, how stretched) rather than in its motion.
  • Near Earth's surface: Ug=mghU_g = mgh — the work needed to raise mass mm through vertical height hh, refundable in full on the way down. Only the vertical height enters, never the path length.
  • Elastic: Us=12kx2U_s = \tfrac12 kx^2 — the spring-triangle area, with xx the deformation from natural length. Compression and extension store identically (xx is squared).
  • The zero of potential energy is a free choice (floor, table, sea level, infinity). Physics only ever uses differences ΔU\Delta U — so negative potential energy is bookkeeping, not danger.
  • The defining exchange: potential energy change is the negative of the storing force's work, ΔU=−Wgrav/spring\Delta U = -W_{\rm grav/spring}. Gravity doing positive work means the account is being spent.
  • Potential energy belongs to the system (Earth + ball, spring + block), not to the object alone — the field and the deformation do the storing.
  • Key idea: potential energy is a refundable deposit — work banked against gravity (mghmgh) or a spring (12kx2\tfrac12 kx^2), with a freely chosen zero, repaid as kinetic energy whenever the system is released.
The core insight

Potential energy measures work stored in configuration — mghmgh for height in uniform gravity, 12kx2\tfrac12 kx^2 for a deformed spring — defined so that the storing force's work equals the account's withdrawal (ΔU=−W\Delta U = -W), and pinned to a reference level chosen purely for convenience, since only differences in potential energy ever touch the physics.

Where this lives today

The world's biggest batteries are still buckets of mghmgh. Pumped-storage hydroelectric plants — about 95% of all grid storage on Earth — pump water uphill when electricity is cheap and let it fall through turbines at peak demand; a single large plant banks around 101310^{13} J, the mghmgh of a small lake raised a few hundred metres. Gravity batteries hoisting concrete blocks in disused mine shafts run the same equation at startup scale. Meanwhile 12kx2\tfrac12 kx^2 earns its living everywhere metal bends: vehicle suspension and crash structures, the power springs of watches, archery's compound bows (engineered force–draw curves storing over 100 J), and prosthetic running blades, which bank a sprinter's landing energy and refund it into the next stride — a spring pretending, briefly, to be a tendon.

The physical picture

Kinetic energy is a current account: visible, spendable, riding with the motion. Potential energy is the fixed deposit. Lift a brick to a shelf and the work you did has not vanished — it sits there, silently, for a year if need be, and the day the brick slips, gravity refunds every joule as speed. Draw a bow and your muscles' work waits in the bent limbs; release, and the arrow collects. The two storage devices of this module — height and deformation — are the only ones the mechanics syllabus needs, and they share one defining habit: the stored amount depends only on the configuration now, not on how it was reached. Haul the brick up a ladder, a ramp, or a wild spiral staircase: same shelf, same mghmgh, same refund.

Figure
Hero — two savings accounts: height in a field, deformation in a spring
deposit work now - withdraw speed later m h lift: +mgh banked drop: refund as KE U = mgh any path up, same balance m release x from natural length U = ½kx² compress or stretch - the square stores both

One habit of language keeps you honest for the rest of physics: the energy is stored in the system, not in the object. "The ball's gravitational potential energy" is convenient shorthand, but the storage device is the ball–Earth pair — remove the Earth and no amount of height means anything. Likewise the spring's energy lives in the spring's stretched bonds, not in the block attached to it. Exams rarely punish the shorthand, but conceptual questions ("where is the energy stored?") pay the precise answer. And a warning made friendly early: the zero is yours. Measuring heights from the floor, the tabletop, or the ceiling changes every UU value on the page and changes no physics whatsoever, because every prediction routes through ΔU\Delta U. Choose the zero once, write it down, never move it mid-problem.

The math: two storage formulas and one sign rule

The deposit rule defines the account. When you lift a body at constant speed, your lifting force does work +mgh+mgh while gravity does −mgh-mgh; the deposit is what gravity will later refund, so we define the change in gravitational potential energy as the negative of gravity's work:

ΔU=−Wgrav(1)\Delta U = -W_{\rm grav} \qquad (1)

where: WgravW_{\rm grav} — gravity's work over the move. Read the sign dial both ways: gravity works negatively on a rising ball (account filling, ΔU>0\Delta U > 0); positively on a falling one (account draining, ΔU<0\Delta U < 0). This single line is the master convention — every potential energy in physics, from springs to atoms, is defined by the same "minus the storing force's work." Since gravity near the surface bills by height alone (Wgrav=−mg ΔhW_{\rm grav} = -mg\,\Delta h, the ramp theorem), the account balance is

Ug=mgh(2)U_g = mgh \qquad (2)

where: mm — mass (kg); gg — 9.89.8 m/s² (CBSE/ISC) or 1010 (JEE/NEET unless stated); hh — vertical height above your chosen zero, in metres, sign included (a cellar sits at negative hh and negative UU). Read it physically: 2 kg raised 3 m banks about 60 J — and banks the same 60 J whether it went straight up or around a helical ramp, because gravity never billed the horizontal wandering. The spring's account comes straight from the variable-force module's triangle. Deforming a spring by xx from natural length takes work equal to the area under F=kxF = kx:

Us=12kx2(3)U_s = \tfrac12 kx^2 \qquad (3)

where: kk — spring constant (N/m); xx — deformation measured from natural length, never from an equilibrium-under-load position, never the spring's total length. The square makes compression and extension identical depositors, and it makes the account grow quadratically: between two stretches the stored energy changes by 12k(x22−x12)\tfrac12 k(x_2^2 - x_1^2) — the difference of squares, big triangle minus small, never 12k(x2−x1)2\tfrac12 k(x_2 - x_1)^2.

devicebalancezero conventionscaling
height (uniform gg)mghmghanywhere — choose and declarelinear in hh: each metre banks the same
spring12kx2\tfrac12 kx^2forced: natural lengthquadratic in xx: the second centimetre costs triple
Derivation AGravitational potential energy — why the account only reads the vertical

We derive Ug=mghU_g = mgh and its path independence from the definition ΔU=−Wgrav\Delta U = -W_{\rm grav}. The strategy: compute gravity's work on an arbitrary crooked climb by slicing it, watch the horizontal parts of every slice drop out, and pocket the height-only bill. The subtlety to watch: this is exactly the "ramp theorem" from the Work module, now running the account rather than the ledger.

Step 1. Take any path from the floor to a shelf at height hh — ladder, ramp, spiral. Chop it into tiny straight steps; a generic step moves Δx\Delta x horizontally and Δh\Delta h vertically.

Step 2. Gravity is a constant force mgmg straight down, so its work on that step is −mg Δh-mg\,\Delta h: the horizontal part of the step is perpendicular to gravity and bills nothing.

Step 3. Sum the steps. The Δh\Delta h's add up to the total height gain hh regardless of the horizontal wanderings, so

Wgrav=−mghW_{\rm grav} = -mgh

Step 4. Apply the definition (1):

ΔU=−Wgrav=+mgh\Delta U = -W_{\rm grav} = +mgh

Setting the zero at the floor makes the shelf balance U=mghU = mgh. The deposit depends on endpoints only — which is what lets a single number per position play the role of stored energy at all. Why a friction-like force could never run an account: friction bills by path length, so two routes to the same shelf would demand two different "balances" for one configuration — the account book would contradict itself. Only path-blind forces get potential energies; that is the next module's whole subject. *Verified: 2 kg up a 5 m ramp to a 3 m shelf — along-ramp component method: mgsin⁡θ×5mg\sin\theta \times 5 with sin⁡θ=3/5\sin\theta = 3/5 gives 2×10×0.6×5=602 \times 10 \times 0.6 \times 5 = 60 J; height method: 2×10×3=602 \times 10 \times 3 = 60 J. Identical, path forgotten.*

Go deeper · Derivation B · 1 minSpring potential energy — the triangle, banked
Derivation BSpring potential energy — the triangle, banked

We derive Us=12kx2U_s = \tfrac12 kx^2 from the work stored while deforming a spring, and extract the between-two-stretches result that carries most of the marks. The strategy: the storing force here is the spring's own restoring pull −kx-kx; the work done against it is the area under the applied-force graph — the triangle from the variable-force module — and the definition ΔU=−Wspring\Delta U = -W_{\rm spring} banks it.

Step 1. Stretch a spring slowly from natural length to extension xx. The spring pulls back with kx′kx' at every intermediate stretch x′x'; the spring's own work is negative, and its magnitude is the area under the F=kx′F = kx' line: 12⋅x⋅kx=12kx2\tfrac12 \cdot x \cdot kx = \tfrac12 kx^2.

Step 2. By the definition, ΔU=−Wspring=+12kx2\Delta U = -W_{\rm spring} = +\tfrac12 kx^2. With the forced zero at natural length (x=0x = 0 stores nothing),

Us=12kx2U_s = \tfrac12 kx^2

Step 3. Compression: run the same argument leftward; the deformation enters squared, so compressing by xx banks the same +12kx2+\tfrac12 kx^2. "Negative extension" never means negative energy.

Step 4. Between two stretches x1→x2x_1 \to x_2:

ΔU=12k(x22−x12)\Delta U = \tfrac12 k\left(x_2^2 - x_1^2\right)

— big triangle minus small triangle. The pretender 12k(x2−x1)2\tfrac12 k(x_2-x_1)^2 is the energy a relaxed spring would store over that extra stretch; a pre-tensioned spring charges more, because you are pulling against an already-fighting force. Why the zero is forced here but free for gravity: the spring has a physically special configuration — natural length, where the restoring force vanishes — and the formula 12kx2\tfrac12 kx^2 hard-codes it. Uniform gravity has no special height; every shelf looks alike to gg, so the zero is yours. *Verified: k=200k = 200 N/m — stretch 0→2 cm: 12×200×0.0004=0.04\tfrac12 \times 200 \times 0.0004 = 0.04 J; 2→4 cm: 12×200×(0.0016−0.0004)=0.12\tfrac12 \times 200 \times (0.0016-0.0004) = 0.12 J, thrice as much for the same centimetres; total 0→4 cm: 0.160.16 J =0.04+0.12= 0.04 + 0.12 ✓.*

Figure
The movable zero — three observers, three balances, one physics
same ball, same drop - the zero is bookkeeping ball, height 20 m lands here ground table zero (B) rooftop zero (C) balances (0.5 kg, g = 10) zero at ground: U 100 J → 0 J ΔU = −100 J zero at table: U 60 J → −40 J ΔU = −100 J zero at rooftop: U 0 J → −100 J ΔU = −100 J every zero → same ΔU → same speed predictions subtract two balances - the choice of zero always cancels

Limiting cases

caseresultmeaning
h=0h = 0 (at the chosen zero)Ug=0U_g = 0nothing special happens there — the zero is bookkeeping, not a floor of the world
h<0h < 0 (below the zero)Ug<0U_g < 0legal and routine; only ΔU\Delta U ever meets the physics
x→2xx \to 2x on a springUs→4UsU_s \to 4U_squadratic storage: pre-stretched springs are expensive to stretch further
same force FF into stiff vs soft springU=F2/2kU = F^2/2k: soft spring stores MOREat fixed force, deformation is what stores
same stretch xx into stiff vs softU=12kx2U = \tfrac12 kx^2: stiff stores moreat fixed deformation, stiffness is what stores
g→0g \to 0Ug→0U_g \to 0 everywhereno field, no storage — height means nothing in deep space

Worked examples

1 · Board level — the water-tank invoice (CBSE/ICSE pattern, 3 marks)

A pump lifts 500500 kg of water from a well 2020 m deep to a rooftop tank 1010 m above the ground. Take the ground as the zero and g=10g = 10 m/s². Find (a) the water's initial and final potential energies, (b) the minimum work the pump must do.

Step 1. (a) Initial: the water sits 2020 m below the zero: Ui=mghi=500×10×(−20)=−1.0×105 JU_i = mgh_i = 500 \times 10 \times (-20) = \boxed{-1.0 \times 10^5\ \mathrm{J}}. Final: Uf=500×10×10=+5.0×104 JU_f = 500 \times 10 \times 10 = \boxed{+5.0 \times 10^4\ \mathrm{J}}.

Step 2. (b) Minimum work (no speed gained, no friction losses) is the deposit: W=ΔU=Uf−Ui=5.0×104−(−1.0×105)=1.5×105 JW = \Delta U = U_f - U_i = 5.0\times 10^4 - (-1.0\times10^5) = \boxed{1.5 \times 10^5\ \mathrm{J}}.

Step 3. Audit with the total lift: the water rose 3030 m in all, and mg×30=1.5×105mg \times 30 = 1.5\times10^5 J ✓ — the negative initial balance was doing honest work, not decoration.

2 · Board level — the pre-stretched spring (ICSE/ISC pattern)

A spring of constant k=400k = 400 N/m is already stretched 5.05.0 cm. How much additional work is needed to stretch it to 10.010.0 cm? Compare with the work that took it from natural length to 5.05.0 cm.

Step 1. First leg, 0 → 5 cm: U1=12×400×(0.05)2=0.50 JU_1 = \tfrac12 \times 400 \times (0.05)^2 = \boxed{0.50\ \mathrm{J}}.

Step 2. Second leg, 5 → 10 cm — difference of squares, never the square of the difference: ΔU=12×400×(0.102−0.052)=200×0.0075=1.5 J\Delta U = \tfrac12 \times 400 \times \left(0.10^2 - 0.05^2\right) = 200 \times 0.0075 = \boxed{1.5\ \mathrm{J}}.

Step 3. Compare: the second 5 cm costs 3×3\times the first — the quadratic account in action. (The wrong formula 12k(0.05)2\tfrac12 k(0.05)^2 returns 0.50 J again and forfeits the comparison mark.)

3 · Competitive level — the movable zero, audited (JEE-Main pattern, ~2.5 minutes)

A 0.500.50 kg ball is thrown from a 2020 m rooftop and lands on the ground. Student A puts the zero at the ground; student B at the rooftop. For EACH student, write the ball's initial and final potential energies and the change ΔU\Delta U. (g=10g = 10 m/s².)

Step 1. Student A (zero at ground): Ui=0.5×10×20=+100U_i = 0.5 \times 10 \times 20 = +100 J; Uf=0U_f = 0. Change: ΔU=−100 J\Delta U = \boxed{-100\ \mathrm{J}}.

Step 2. Student B (zero at rooftop): Ui=0U_i = 0; Uf=0.5×10×(−20)=−100U_f = 0.5 \times 10 \times (-20) = -100 J. Change: ΔU=−100 J\Delta U = \boxed{-100\ \mathrm{J}}.

Step 3. The balances disagree everywhere; the change agrees exactly — and the change is all that gravity's refund (the ball's kinetic-energy gain) ever reads. Both students predict the same landing speed.

Step 4. MCQ armour: any option that lets a physical outcome (speed, height reached, tension) depend on where the zero sits is wrong by structure. Timing note: ~2.5 min, almost all of it careful sign work. Write hh WITH its sign relative to your zero and the table fills itself. The designed error is giving B's ΔU\Delta U as +100+100 J by subtracting in the wrong order — it is always final minus initial.

4 · Competitive level — same force versus same stretch (JEE-Main/NEET conceptual, ~3 minutes)

Two springs have constants k1=100k_1 = 100 N/m and k2=400k_2 = 400 N/m. (a) Each is stretched by the same force F=20F = 20 N. Which stores more energy, and how much each? (b) Each is stretched by the same extension x=10x = 10 cm. Which stores more now?

Step 1. (a) Same force: each spring settles at x=F/kx = F/k, so x1=0.20x_1 = 0.20 m, x2=0.05x_2 = 0.05 m. Energy: U=12kx2=12k(F/k)2=F22kU = \tfrac12 kx^2 = \tfrac12 k (F/k)^2 = \dfrac{F^2}{2k} — inverse in kk. U1=400200=2.0 JU_1 = \dfrac{400}{200} = \boxed{2.0\ \mathrm{J}}, U2=400800=0.5 JU_2 = \dfrac{400}{800} = \boxed{0.5\ \mathrm{J}}. The soft spring stores more: it yields further, and distance is where work lives.

Step 2. (b) Same stretch: U=12kx2U = \tfrac12 kx^2 directly — linear in kk. U1=12×100×0.01=0.5 JU_1 = \tfrac12 \times 100 \times 0.01 = \boxed{0.5\ \mathrm{J}}, U2=12×400×0.01=2.0 JU_2 = \tfrac12 \times 400 \times 0.01 = \boxed{2.0\ \mathrm{J}}. Now the stiff spring wins.

Step 3. The inversion is the entire question, and it is pure formula hygiene: at fixed FF use F2/2kF^2/2k; at fixed xx use 12kx2\tfrac12 kx^2. Identify which variable the problem holds equal before comparing. Timing note: ~3 min. The NEET phrasing is often "work done in stretching" rather than "energy stored" — same number. Options bait the unconverted with the other case's winner; note the elegant symmetry that the winners swap and the winning margins (4×4\times) match.

Exam traps

Path length for height. hh in mghmgh is the vertical rise, never the ramp length. A 5 m ramp to a 3 m shelf banks mg×3mg \times 3. If an incline appears, multiply by sin⁡θ\sin\theta once — and only once.

Stretch measured from the wrong place. The xx in 12kx2\tfrac12 kx^2 runs from natural length. A spring hanging with a load at equilibrium is already deformed; further motion questions must track total deformation from natural length, not from the loaded rest point. (When energy methods and this trap collide, the mass-dropped-on-spring archetype settles it — ahead in this unit.)

The difference-of-squares evasion. Between stretches: 12k(x22−x12)\tfrac12 k(x_2^2 - x_1^2). Writing 12k(x2−x1)2\tfrac12 k(x_2-x_1)^2 undercharges — it forgets the spring was already fighting. Big triangle minus small triangle, every time.

Moving the zero mid-problem. Choose the reference level once, declare it, and keep it. The classic self-sabotage: measuring the initial UU from the ground and the final UU from the table. All balances, one zero.

Signs into ΔU=−Wgrav\Delta U = -W_{\rm grav}. Gravity does positive work on the way down and the account drains. If your rising ball shows decreasing UU or your falling ball shows increasing UU, a sign died. Cross-check with common sense: higher = more banked, always.

"Energy stored in the mass." For conceptual/assertion-reason questions: gravitational PE belongs to the ball–Earth system; spring PE lives in the spring. Shorthand is fine in calculations; the concept question wants the system answer.

mgh beyond its passport. U=mghU = mgh assumes uniform gg — fine for buildings, wells, mountains; wrong for satellites and escape problems, where the gravitation chapter's −GMm/r-GMm/r takes over. If the height rivals Earth's radius, mghmgh has expired.

Which gg. CBSE/ISC: 9.89.8 m/s² unless told; JEE-Main/NEET: almost always 1010. State your choice.

Origins · 1 minWilliam Rankine and how this was discovered
WR
William Rankine
1820–1872 · University of Glasgow

A Scottish engineer-physicist who moved between railway cuttings and thermodynamic theory without changing gait, Rankine coined the term "potential energy" in 1853 — naming the silent half of the energy ledger while helping build the science of engines that spends it. His textbooks on applied mechanics trained a generation of engineers, his thermodynamic cycle still describes most of the world's steam power stations, and his insistence that "actual" (kinetic) and "potential" energy together form a conserved whole put the vocabulary of this entire unit into circulation. The word choice was deliberate: energy in potentia — capability on deposit, waiting.

History
1673
— Huygens' pendulum studies show a bob returns to its release height, hinting that something about position is banked and repaid
1717
— Johann Bernoulli treats virtual work, the germ of energy stored in configuration
1777
— Lagrange builds the potential function into analytical mechanics
1853
— Rankine names "potential energy," partnering Young's "energy" and completing the ledger's vocabulary
1882
— First pumped-storage hydro plants turn mghmgh into a grid-scale battery

Going further

Texts. NCERT Physics Class 11, Work, Energy and Power chapter (potential energy of a body and of a spring) · HC Verma, Concepts of Physics Vol. 1, Ch. 8 · Halliday–Resnick–Walker Ch. 8 (the potential-energy treatment this module downshifts).

Provenance. All results are standard. A verifier should recompute: the beam's 30 kJ ↔ 17 m/s round trip; the ramp check (60 J both routes); the spring ladder 0.04 / 0.12 / 0.16 J; the water tank (−105-10^5, +5×104+5\times10^4, 1.5×1051.5\times10^5 J); the rooftop double audit (ΔU=−100\Delta U = -100 J for both zeros); the spring pair (2.02.0 vs 0.50.5 J swapping winners between fixed-FF and fixed-xx).

Prerequisites. Work Done by a Force · Work by a Variable Force & F–x Graphs · Spring Force & Hooke's Law. Leads to. Conservative & Non-Conservative Forces · Conservation of Mechanical Energy & Efficiency · ARCHETYPE: Block Meets Spring · Conservative Forces & Potential Energy (UG deep dive).

Where is the potential energy actually stored — in the ball, the Earth, or where?

In the configuration of the pair — operationally, in the gravitational field of the ball–Earth system. Neither partner owns it: bring the same ball to the same height above the Moon and the account reads a sixth as much, because the account was never the ball's alone. For springs the address is more concrete: the energy sits in the stretched atomic bonds of the coil. The habit "the ball's PE" is harmless shorthand for calculations, but system-language answers the concept questions correctly and prepares you for fields, where the storage idea becomes central.

Why is potential energy DEFINED as minus the force's work? The minus sign feels arbitrary.

It's the sign that makes the total budget add up. When gravity does positive work, the body speeds up — kinetic energy grows — and we want the stored account to shrink by exactly that much, so the sum KE+UKE + U holds still. Define ΔU=−W\Delta U = -W and the bookkeeping is automatic: every joule the force pays out of storage lands in motion, and vice versa. With a plus sign, both accounts would rise together and "conservation" would count every transfer twice. The minus sign is the price of a balanced budget — the next module cashes it in.

Does a ball sitting on the floor have zero potential energy?

Only if you put the zero there — and you don't have to. Dig a well beside it and the floor-level ball has positive energy it can still spend (it can fall into the well). There is no absolute empty account in uniform gravity; there are only balances relative to a declared level. Exam questions dodge the ambiguity by declaring the reference ("taking the ground as zero…"), and you should too, in the first line of any solution where heights appear.

Why does the spring's energy not depend on whether it's compressed or stretched?

Because the deformation enters squared, and the physics justifies the algebra: pushing coils together or pulling them apart both distort the same bonds away from their preferred spacing, and the restoring force ramps up linearly in both directions. Work the triangle either way and you bank the same area. Watch the language in problems, though — "the spring is compressed by 4 cm from its natural length" and "the spring's length is reduced to 4 cm" are different statements, and only the first hands you xx directly.

Can I always convert the stored mgh fully back into kinetic energy?

In the idealized problems of this module, yes — that is what "refundable" means, and the next module makes it a law. In reality, air drag and friction skim the refund: a ball dropped from 5 m arrives with slightly less than mg×5mg \times 5 of kinetic energy, the difference having become heat in the air. The forces that refund honestly (gravity, springs) versus those that skim (friction, drag) are exactly the conservative/non-conservative split of the following module — this FAQ is its trailer.

A pendulum at its lowest point has zero height — has its potential energy vanished into speed everywhere along the swing?

Track the exchange continuously: at the extremes of the swing the bob is momentarily still (all account, no cash); at the bottom it is fastest (all cash, no account, if the bottom is your zero); in between, a running mixture. Nothing vanishes anywhere — the total holds while the ratio slides. This continuous slosh between mghmgh and 12mv2\tfrac12 mv^2 is the pendulum's whole personality, and computing the bottom speed from the top height (v=2ghv = \sqrt{2gh}) is the single most common energy exam question ever set.

Why does mgh stop working for rockets and satellites?

Because mghmgh was derived assuming gravity is the same force at every height — true to within a percent up to about 30 km, but badly false at satellite altitudes, where gg has visibly weakened. The honest account for large ranges uses the universal gravitation chapter's U=−GMm/rU = -GMm/r, which reduces to mghmgh (plus a constant — reference freedom again) when h≪REarthh \ll R_{\rm Earth}. The exam boundary is simple: buildings, wells, cliffs, mountains → mghmgh; orbits, escape speeds, "far from Earth" → the general formula.

Before this

Where this leads