Climb inside the rocket

1 / 5

RP-1 tank (kerosene)

The lower tank, full of a very pure kind of paraffin — rocket-grade kerosene.

It sits at the bottom, nearest the engines. But the oxygen it needs to burn is in the tank ABOVE it — which is a problem the rocket has to solve with a pipe.

For engineers: Skin-and-stringer carries axial compression through discrete stiffeners, which is the right answer for the tank taking thrust loads at the base. The LOX tank above is pressure-stabilised monocoque instead. Same stage, same barrel diameter, opposite structural philosophies — because the load cases genuinely differ.

Drawn to order, not to scale — SpaceX publishes no internal tank dimensions.

2 / 5

Double-wall transfer tube (the downcomer)

A pipe carrying freezing liquid oxygen straight down through the middle of the kerosene tank.

The oxygen is on top and the engines are at the bottom, so the oxygen has to get past the fuel somehow. It goes right through the middle of it — inside a pipe with a gap in its wall, so the two never touch and the kerosene never freezes solid around it.

For engineers: It's a lovely illustration of coupled design: the common dome buys mass, and immediately spends some of it back on a double-walled duct and its thermal management. Note also what SpaceX does NOT say — it states the tube is double-walled but never states why, and it never mentions a transfer tube in the second stage at all. Do not assume one exists there.

Drawn to order, not to scale — SpaceX publishes no internal tank dimensions.

3 / 5

Common dome

One shared wall between the two tanks — the roof of the kerosene tank IS the floor of the oxygen tank.

Two separate tanks would need two walls with a gap between them. Sharing one wall deletes a whole chunk of rocket. Every metre you delete is a metre you don't have to lift.

For engineers: The mass argument is straightforward; the interesting part is the coupling it creates. LOX at ~90 K against RP-1 that freezes near 225 K means the interface is now a design problem, and the transfer tube through the fuel tank is double-walled for the same reason. Note Saturn V made the opposite call one stage down: the S-IC used a full intertank, because kerosene-against-LOX yields far less thermal payoff than the S-II's hydrogen-against-oxygen did.

Drawn to order, not to scale — SpaceX publishes no internal tank dimensions.

4 / 5

LOX tank (liquid oxygen)

The upper tank, full of liquid oxygen so cold it would freeze anything it touched.

Fire needs oxygen, and there is none in space — so the rocket carries its own. There's far more oxygen aboard than fuel, by weight: burning needs more oxygen than you'd ever guess.

For engineers: Sub-cooling the LOX densifies it, packing more oxidiser into the same volume: free Δv for the price of a colder ground system. Monocoque is viable here because a pressurised cylinder in tension is the cheapest possible structure — but it means an unpressurised tank is a structurally compromised one.

Drawn to order, not to scale — SpaceX publishes no internal tank dimensions.

5 / 5

Helium COPVs — inside the oxygen tank

Bottles of helium that sit INSIDE the liquid oxygen tank, in the cold.

As a tank empties, the space left behind has to be filled or the tank crumples — so helium is pushed in behind the propellant. The bottles live inside the cold oxygen because cold helium squeezes up smaller, so more of it fits in the same bottle.

For engineers: Cold-packed pressurant storage is sound and long-established. SpaceX's investigation traced the AMOS-6 failure to buckles in the COPV liner that allowed solid oxygen to accumulate and ignite under the carbon overwrap. The lesson is the one engineering keeps re-teaching: the concept was right and a manufacturing detail was fatal.

Drawn to order, not to scale — SpaceX publishes no internal tank dimensions.

1 / 5

RP-1 tank — 216,000 gallons

The kerosene tank — 216,000 gallons of it, enough to fill a swimming pool several times over.

Glued into the bottom of it is a block of foam whose entire job is to take up space, so less kerosene can hide in the corners where the pumps can't reach it.

For engineers: The wall thins as it rises because it only carries what's above it. The S-IC LOX tank does the same in eight steps, .254 → .190 in. Structure follows the load, and the load shrinks with every metre.

2 / 5

Five LOX suction tunnels

Five pipes of liquid oxygen running straight down through the kerosene tank to reach the engines.

Falcon 9 does exactly the same thing today — with one pipe instead of five. When the oxygen is on top and the engines are at the bottom, there's only one place for the pipe to go: through.

For engineers: The counts don't mirror, which is the giveaway about locality: five LOX lines cross the tank from above, but fuel feed is local — 'two 12-inch ducts to each F-1 engine', ten in total. Oxidiser travels; fuel is already there.

3 / 5

Intertank — the road not taken

A whole section of rocket between the two tanks, just holding them apart.

The stage above this one shares a single wall between its tanks to save weight. This stage doesn't — it spends the weight instead. The same rocket made opposite choices one stage apart.

For engineers: The common-bulkhead trade is dominated by the ΔT across it and the mass of the section it replaces. For LH2/LOX (a 70 °C gap and an enormous hydrogen tank) it wins decisively. For RP-1/LOX it largely evaporates, and you take the simpler, inspectable, penetrable structure instead.

4 / 5

LOX tank — 345,000 gallons

345,000 gallons of liquid oxygen — and hidden inside it, four big bottles of helium sitting in the cold.

The helium bottles are filled twice: partly before the oxygen goes in, then topped right up once the oxygen has made them freezing cold. Cold gas squeezes up smaller, so you fit more in.

For engineers: Note that last detail: the S-IC LOX tank is AUTOGENOUSLY pressurised — it makes its own pressurant from its own propellant. So is the S-IVB's hydrogen tank, from the J-2's fuel manifold. 'Autogenous pressurisation is new, Starship pioneered it' is flatly false; it flew in 1967. Falcon 9 is the one that's pure helium.

5 / 5

The common bulkhead

One wall with liquid hydrogen above it and liquid oxygen below. On one side it's −253 °C; on the other, −183 °C. In places that wall is barely two millimetres thick.

It replaces two whole tank ends AND the section that would have separated them. One wall doing the work of three.

For engineers: MSFC-MAN-504 gives LOX at −297 °F and LH2 at less than −423 °F: a 126 °F gap, which is precisely 70 °C, across a bonded sandwich. The core's taper is the thermal design — thick where the gradient has room, thin where structure dominates at the rim.