🛰️ Launch Vehicles: Saturn V, Shuttle, and Falcon

What you’ll learn

1. What Is Expensive?Frame every launch vehicle as an argument about what is scarce, given that the physics is identical and settled.Tsiolkovsky's equation, the ~9.4 km/s to orbit, mandatory staging and chemistry's ~450 s Isp ceiling apply identically to every vehicle ever built and have not moved since 1903. Yet Saturn V, the Shuttle and Falcon 9 are radically different machines built by serious people who all knew the same physics. The reason is that physics says what is possible and never what is worth it — and 'worth it' is a question about money, schedule, politics and risk. Saturn V decided time was scarce; the Shuttle decided hardware was; Falcon 9 decided hardware is, but only if refurbishment and cadence are solved too.
2. Saturn V: Money Was Not the ConstraintAccount for Saturn V's scale, its expendability and its flight record as correct answers to a schedule constraint.Saturn V stood 111 m (363 ft) tall and 10 m (33 ft) across, massed 2,822,171–2,965,241 kg at liftoff, and put 140,000 kg into LEO and 43,500 kg toward the Moon — five times a Falcon 9's LEO payload, from a vehicle retired in 1973 and never matched since. Five F-1 engines gave 33,000 kN at sea level; the S-IC burned LOX/RP-1 and the S-II and S-IVB burned LOX/LH2, the density-impulse argument made physical. It threw away everything, because the constraint was a calendar rather than a budget: recovery hardware is mass you cannot use, and refurbishment is a process that takes years to invent. Thirteen flights from 9 November 1967 to 14 May 1973: twelve successes, one partial failure (Apollo 6), no crew lost on ascent — achieved with enormous margins and a launch escape tower carried every time and never used.
3. Why Nobody Kept the ToolingExplain why Saturn V was retired, and correct the 'we lost the plans' myth.Saturn V cost $185 million per launch in 1969–1971 — 'equivalent to $995 million in 2024' — but the fatal problem was not cost, it was that there was nothing else to do with it. 140 tonnes to LEO was a commercially useless number in 1972: no 140-tonne satellites, no customers, no market at any price. It was the instrument of a single argument with the Soviet Union, and once won, the instrument had no purpose and an enormous running cost. Apollo 18, 19 and 20 were cancelled and the tooling was scrapped or repurposed along with the workforce. 'We lost the plans' is a myth — the drawings survive. What was lost was the tooling, suppliers, process knowledge and the undocumented judgements that live in people's hands, which is far less recoverable.
4. The Shuttle's PromiseExplain the Shuttle's reuse rationale and how Air Force requirements — especially cross-range — determined the vehicle.The Shuttle's diagnosis of Saturn V's disease was correct: no other transport industry scraps its vehicle after one trip. Build one that comes back, amortise it over dozens of flights, and cost per kilogram collapses. But funding required Air Force backing, and the Air Force required lifting '29,000 kg (65,000 lb) to an eastward LEO or 18,000 kg (40,000 lb) into a polar orbit' with a '4.6 by 18 m (15 by 60 ft)' payload bay — and, decisively, cross-range: launch polar, orbit once, land at the launch site, which the rotating Earth turns into hundreds of miles of sideways flight during re-entry. NASA favoured straight wings; Air Force engineers argued they 'would not provide the required cross-range capability'. The resulting delta wing meant more structure, more tiles, and orbiter mass carried up and back on every flight — for a profile essentially never flown.
5. What the Shuttle Actually ReusedDistinguish reuse from cheap reuse, and identify cadence as the variable the Shuttle got wrong.The external tank 'was the only major component of the Space Shuttle system that was not reused'. The SRBs were recovered — from salt water — then disassembled, cleaned, inspected, requalified and re-cast, with a recovery fleet. The orbiter came home to months of processing: engines pulled and inspected, thousands of individually shaped tiles checked essentially one by one. The lesson: reuse saves the cost of building hardware, not the cost of trusting it — and trusting it required a standing army that was a fixed cost. So cost per flight depended on flight rate, and the rate was 135 missions from 12 April 1981 to 21 July 2011, about 4.5 a year against a promise of dozens. Marginal cost was US$450 million (2011); the US$211 billion (2012) programme over 135 flights is roughly $1.5 billion each — more per flight than expendable Saturn V, for a fifth of the payload.
6. Side-MountExplain side-mount as a safety architecture, and state the two accidents' causes soberly as the boards found them.Side-mount is structurally defensible — it lets the tank be a tank and the orbiter be an aeroplane — but it puts the orbiter in the tank's debris path and removes any escape option; the Shuttle 'was designed without launch escape systems', with early ejection seats 'disabled and later removed', because 'expected high reliability would preclude the need for one'. Challenger: the Rogers Commission found hot gas bypassed O-rings in a field joint — a joint existing because the SRBs were railed in four sections from Utah — and blamed 'a faulty design of the field joint that was unacceptably sensitive to changes in temperature'; O-ring temperature was calculated at 26 °F against a coldest previous launch of 53 °F, and Thiokol engineers had warned they 'did not have enough data'. Columbia: foam from the left bipod ramp 'struck the reinforced carbon–carbon (RCC) panels on Columbia's left wing at relative velocity of 625 to 840 feet per second'; the CAIB found 'NASA management did not consider the potential risk to the astronauts as a safety-of-flight issue' despite a history of strikes.
7. Falcon 9: Reuse, Argued DifferentlyExplain Falcon 9's narrower reuse thesis and the mechanics of propulsive landing.Falcon 9's insight is a refusal: recover only the first stage, because it is both the most valuable (nine Merlins, most of the structure; Block 5 first-stage thrust 7,600 kN / 1,700,000 lbf) and the easiest — it separates well below orbital velocity, so no tiles are needed. The second stage would need to return from Mach 25, exactly the problem the Shuttle solved with tiles and a standing army, so it is thrown away. Landing is hard for two reasons: a rocket on its exhaust is an inverted pendulum with no stable equilibrium, held up only by software gimballing many times a second; and the nearly-empty stage cannot hover, since even one Merlin at minimum throttle out-thrusts its weight — hence a suicide burn that must reach zero velocity exactly at the deck. First flight 4 June 2010; first successful landing 21 December 2015; 623 landings from 634 attempts; one booster has flown 36 times.
8. Does Reuse Pay?Quantify the payload cost of reuse and establish cadence as the mechanism, with company-stated figures flagged as such.Reuse costs payload: Falcon 9 lifts 22,800 kg (50,300 lb) to LEO expendable but 17,500 kg (38,600 lb) landing on a drone ship — about a 23% penalty, and returning to the launch site costs more still. That mass is grid fins, legs, landing-rated structure and above all the propellant for the entry and landing burns. It only pays if the booster is worth more than the payload forgone, which is a question about cadence and refurbishment cost. The mechanism is fixed costs: a workforce, factories, pads and ships exist whether you fly four times a year or four hundred, so flight rate is an input to a cheap rocket rather than an outcome. Wikipedia's Falcon 9 article records 666 launches and US$74 million per launch (2026); a company's price is not the same as its cost, and should be read as such. The Shuttle never entered the loop — and being crewed, winged and side-mounted on every flight meant fixed costs it could never divide by a big enough number.
9. The LedgerClose the ledger across the three vehicles and generalise the 'what is expensive?' question.Saturn V gave up all its hardware every flight and got simplicity, scale and schedule — 140,000 kg to LEO, 13 flights, no crew lost on ascent — then lost its tooling for want of a second customer. The Shuttle gave up payload, abort options and flight rate to bring hardware home, and ended at roughly $1.5 billion per flight for a fifth of Saturn V's payload, with two accidents in which the side-mount configuration is implicated. Falcon 9 gave up its second stage and 23% of its payload, and got nine engines back on a boat. Nothing in the physics moved across sixty years: the answer to 'what is expensive?' did. So when part of a launch vehicle looks like a mistake, the useful question is not why they missed the obvious — it is what they thought was expensive, because the shape of the machine encodes the economics of the decade that built it.

Questions this course answers

Grounded in trusted sources

• Wikipedia — Saturn V (height 111 m / 363 ft; diameter 10 m / 33 ft; liftoff mass 2,822,171–2,965,241 kg; payload 140,000 kg / 310,000 lb to LEO and 43,500 kg / 95,900 lb to TLI; 5 × F-1 at 33,000 kN / 7,500,000 lbf sea level; S-IC LOX/RP-1, S-II and S-IVB LOX/LH2; 13 launches with 12 successes and 1 partial failure (Apollo 6); first flight 9 November 1967 (AS-501, Apollo 4), last 14 May 1973 (AS-513, Skylab); cost per launch '$185 million (equivalent to $995 million in 2024)'): https://en.wikipedia.org/wiki/Saturn_V
• Wikipedia — Space Shuttle (135 missions; first flight 12 April 1981 (STS-1), last 21 July 2011 (STS-135); payload to LEO 27,500 kg / 60,600 lb; cost per launch US$450 million (2011); total programme US$211 billion (2012); Air Force requirements of '29,000 kg (65,000 lb) to an eastward LEO or 18,000 kg (40,000 lb) into a polar orbit' and a '4.6 by 18 m (15 by 60 ft)' payload bay; Air Force engineers argued straight-wing designs 'would not provide the required cross-range capability'; 'The ET was the only major component of the Space Shuttle system that was not reused'): https://en.wikipedia.org/wiki/Space_Shuttle
• Wikipedia — Falcon 9 (height 69.8 m / 229 ft, Full Thrust; mass 549,000 kg / 1,210,000 lb; payload to LEO 22,800 kg / 50,300 lb expendable and 17,500 kg / 38,600 lb landing on a drone ship; Block 5 first-stage thrust 7,600 kN / 1,700,000 lbf from nine Merlins; first flight 4 June 2010; first successful booster landing 21 December 2015; 666 launches and 623/634 successful landings as of the article's July 2026 content; booster reflight record 36 flights; cost per launch US$74 million (2026)): https://en.wikipedia.org/wiki/Falcon_9
• Wikipedia — Space Shuttle Challenger disaster (Rogers Commission findings: hot gas bypassing O-ring seals in the right SRB field joint with 'no other potential causes for the disaster'; 'a faulty design of the field joint that was unacceptably sensitive to changes in temperature, dynamic loading, and the character of its materials'; 36 °F air temperature and 26 °F calculated O-ring temperature against a 53 °F coldest previous launch; Morton Thiokol engineers warned they 'did not have enough data to determine whether the O-rings would seal at temperatures colder than 53 °F'; SRBs built in four sections in Utah creating three field joints; Feynman's ice-water demonstration and appendix; no launch escape system — ejection seats 'disabled and later removed for the operational flights' because 'expected high reliability would preclude the need for one'): https://en.wikipedia.org/wiki/Space_Shuttle_Challenger_disaster
• Wikipedia — Space Shuttle Columbia disaster (CAIB: 'At T+81.9 seconds, the foam struck the reinforced carbon–carbon (RCC) panels on Columbia's left wing at relative velocity of 625 to 840 feet per second'; foam's low ballistic density making the collision essentially inevitable given the geometry; 'Despite a history of foam strike events, NASA management did not consider the potential risk to the astronauts as a safety-of-flight issue'; NASA lacked 'appropriate communication and integration channels to allow problems to be discussed and effectively routed and addressed'): https://en.wikipedia.org/wiki/Space_Shuttle_Columbia_disaster
• Report of the Presidential Commission on the Space Shuttle Challenger Accident (the Rogers Commission), 1986
• Columbia Accident Investigation Board Report, Volume 1, NASA, August 2003
• Wikipedia — Rocketdyne F-1; RS-25; Space Shuttle Solid Rocket Booster; Space Shuttle thermal protection system; Space Shuttle orbiter; Space Shuttle Discovery

Every Wunder lesson is built from real, reputable sources — never invented.

Related Science courses