Space Travel

Part 3 - (Turbo) Pumping Gas

Ross, Going Postal

There are lots of technical problems in designing a liquid fuelled rocket; such as how do you light it, relight it, how do you stop combustion instability shaking the rocket to pieces, how do you manage to direct the thrust to give you guidance control.

Surprisingly one of the most significant obstacles is designing the fuel pump. If you have ever done any car maintenance you will know that modern car fuel pumps are small cheap and last a long time. Why should a rocket be different?

Think about all the times in your life that you have stood pumping fuel at the petrol station, if you are like me you are probably passing the time wondering why your spouse always leaves the car on less than a quarter full. Now think of all of those times put together for a 50 year motoring lifetime as one long pump into a fuel tank. Well here is a video of a tank on a rocket as it is launched being emptied of more fuel than you will pump in your life.

This tank is almost 2 metres wide and 16 metres long and holds 38,000L-ish; it is a Redstone rocket fuel tank (nicked from the parts bin to put in a Saturn I booster where it was one of nine fuel tanks used simultaneously on the first stage).

By launch vehicle terms this Redstone tank is a tiddler. The Saturn V first stage had 770,000 litres of RP1 and 1,250,000 litres of LOx (about 50 times the quantity held in this tank).

In fact the Saturn V first stage would pump your entire motoring life’s fuel volume in just over 2 seconds; it would empty 2 million litres of propellant in about a two and a half minute burn. The LOx alone required five 17 inch diameter pipes to feed the massive F1 engines.

The engineering problems of shifting this much fuel that quickly are significant, and it isn’t just volumes of fuel it is the hostile nature of the environment – high G, very low temp LOx, the fact that if the fuels come into even the minutest contact e.g. leaking through the pump seals, then the astronauts sitting on top of the stack can kiss their butts goodbye.

One additional problem that this is not like the fuel pump for the car, in a car engine the fuel does not have to be pumped into a cylinder against the pressure of combustion – at the time the fuel is injected/sucked in no combustion is happening; In a rocket engine combustion is happening in a very big way, so the pump pressure damn well better be higher than the combustion pressures inside the Nozzle otherwise the burning gas will come back up the feed pipes and into the pump.

In fact the fuel pumps are so powerful they often have their own dedicated rocket engine to drive them, crazy as that sounds. The fuel pump and how it is arranged is sufficiently big an issue that some in the space industry say, only half jokingly, that a rocket is basically a fuel pump with some associated hardware!

In this video a group of Danish shed-dwellers create and test their own hydrazine powered rocket fuel pump, very similar in design to that used on the V2 (or A4 as we shall call it following next week’s instalment). [if you are not a shed dweller, you may want to skip to the test firing at 3 min 35]

Like everything else on a rocket, the pumps also have to be incredibly light. This combination of high power and lightness makes these components some of the highest power to weight ratio machines ever designed. Each of the five F1 fuel pumps on a Saturn V was 55,000Hp and weighed only 2,500lb, giving a power to weight of around 48000:1, even more amazing is the smaller Merlin 1D fuel pump used in the Falcon 9 – this weighs only 150lb (about the same as Taylor Swift) and produces close to 10,000Hp of power (about the same two large railway diesel locomotives coupled together) this equates to a power to weight of around 145,000:1.

There are several low power ways of pumping fuel that upper stages sometimes use – both simple pressurisation of the fuel with another gas, or simple heating of the propellants allowing them to expand. Both these methods are simple but have the double downsides of low thrust and low exhaust velocity therefore lower specific impulse (roughly – miles per gallon).

In fact there is an equation that allows us to determine the exhaust velocity based on a number of other parameters in the combustion chamber (some terms not defined for simplicity)


Ross, Going Postal

We know we want the maximum possible exhaust velocity – so if OT were here for Monday maths, we’d know that this equation means we want high combustion chamber pressure and high temperature in the chamber to get the best Ve. These parameters are linked in that the higher the pressure you inject propellants, then the higher temperature in the chamber.

So the poor old fuel pump has not only to deliver a shed load of fuel it has to do it at maximum pressure.

As an aside: The performance of the engine relating to the fuel pressure means that during high G acceleration you can get real problems as the G forces fuel into the pumps – this can result in a problem known as “pogo-ing” an oscillation along the thrust axis with the fuel surging and this disturbing the combustion and fuel delivery in a feedback loop. Unmanned Apollo 6 (the second time the Saturn V flew) was almost a complete disaster as a result of pogo-ing with the oscillations being so violent that they seriously damaged the second and third stages (whole panels were thrown off and more importantly fuel lines were ruptured) – the rather bodgy fix to this was flown first on the manned Apollo 8 – which we will cover later in the series.

As we saw from our Danish friends in the video above one way to to power the fuel pump is to use another fuel in a“gas generator”, or more commonly a gas generator is run from the same fuel source and tanks as the main rocket –


Ross, Going Postal

This is the way most launch vehicles work today, but there are higher performance designs in particular “staged combustion”.

These are really quite odd and involve burning in the main flow – by adjusting the mixture to be really “off” i.e. adding a little of the other propellant and actually burning it in the flow of the main propellant. You can do this on one side only or both sides.

This has the advantage that all of the combustion contributes something to thrust – where in a gas generator design – the fuel spent in the pre-burner and all that horsepower doesn’t really contribute to thrust (it gets vented as exhaust at below mach 1 as opposed to mach 30ish out of the combustion chamber).

The RD180 that the Russians sell to the american United Launch Alliance is of this staged combustion design.

Then at the very top of the high performance designs you have a “full flow staged combustion cycle” design. These are sufficiently exotic to have never been flown, the RD-270 was tested by Energomash in the 1960s and currently in development and testing is the Raptor for SpaceX next generation rocket, which may fly a test hop next year.


Ross, Going Postal

In a “full flow” staged combustion both sides of the propellant circuit have pre-burners and effectively the full flow of all of the propellant is used to turn the turbines (thus the name). This gives the very highest pressure (and thus temperature) in the combustion chamber and if you can design a combustion chamber to withstand that, then you have a very high thrust and high efficiency engine, you also tend to get high thrust to weight (of engine).

The F1 used in the Saturn V had like most large rocket boosters used a Gas generator – with a chamber pressure of about 70 atmospheres, the Merlin 1D has a chamber pressure of about 100 atmospheres, the RD180 has a chamber pressure of 260 atmospheres and the Raptor will be expected to achieve 300 atmospheres eventually. This will result in better specific impulse, thrust and thrust to weight ratio.

So even though it feels like we reached the pinnacle of space design in the 1960s, we are gradually making further progress in improving performance using new materials and applying new brains to the problem.

Next Time – German Rocketry in the 1930s

© Ross 2018

Audio file


*Image credits:

By Duk – Own work, CC BY-SA 3.0,

CC BY-SA 3.0,