Sorry if this is the wrong place to ask this question, as its very much hypothetical.

In the computer game Kerbal Space Program , there are 3 main Staging Options that are used

  1. Conventional Staging where each stage separates, then reveals the engines above it, the same as used in the Saturn V Rocket.
  2. Onion Staging where an outer tank feeds into an inner tank and when empty is dropped, similar to the crossfeed between the Shuttle and Main Orange Tank
  3. Asparagus Staging where Engines and tanks surround the main rocket, which feed into the rocket beside it around the central core then to the next to the next and then to the central core which above it has the final stage, this setup in KSP allows lifting huge spacecraft into orbit.

Asparagus Staging Diagram

This means that the Fuel and Oxidizer Tanks in S4 actually feed all 7 Engines, until empty and dropped, then S3 feeds all 5 until empty, the dropped, Same with S2, and then S1 is now already very high in the atmosphere traveling very fast but is basically fully fueled and continues on into orbit

This is the meat of the question, Asparagus staging in KSP is easy, just done with magical fuel lines, but in reality, it would take huge pumps and fuel lines and structural supports. so all this added weight would reduce the effective delta-V you could gain from it. but I'm not sure how much

Now I know that without specific information in terms of Mass of the stages etc it's impossible to do the actual maths, but:

The Question

Is this even realistically possible in real life? or would the drawbacks outweigh the benefits

  • 2
    $\begingroup$ Not quite a duplicate, but very relevant: space.stackexchange.com/questions/18665/… $\endgroup$
    – Lex
    Nov 6, 2018 at 16:47
  • 2
    $\begingroup$ It was considered for Falcon Heavy byt eventually abandoned. en.wikipedia.org/wiki/Falcon_Heavy#Propellant_crossfeed $\endgroup$ Nov 6, 2018 at 16:57
  • $\begingroup$ @lex, Cheers, i looked through a couple of the how did the shuttle do it questions. Its genuinely interesting stuff but doesn't really look into if it is actually possible, people just seem to wave it off as not possible without explaining why. $\endgroup$ Nov 6, 2018 at 16:59
  • $\begingroup$ @SteveLinton, again, cheers, i see options to it often but never any reasons behind why it wasn't done. just it won't be done $\endgroup$ Nov 6, 2018 at 17:08
  • 1
    $\begingroup$ "Staging where an outer tank feeds into an inner tank and when empty is dropped, similar to the crossfeed between the Shuttle and Main Orange Tank" -- Is that true? I thought that the main engines weren't used after ET separation, and instead the OMS (with its separate fuel tanks) was used for orbital injection. $\endgroup$ Nov 6, 2018 at 19:18

2 Answers 2


It's possible, but not as easy in real life as it is in KSP.

To maintain proper tank pressurization, the crossfeeds have to be pump-driven; the fuel and oxidizer crossfeed lines have to be pretty large to move the required amount of propellant (i.e. on the order of the same power as the core section's engine turbopumps). All this increases weight and complexity.

In your proposed design, a minor issue is that the pumping direction of the outer tanks produces a rolling torque on the rocket which has to be countered (via gimbaled engines or other attitude-control mechanisms). No torque is produced by the S2->S1 feeds in your design, so a single-level N-to-1 crossfeed, like Falcon Heavy's proposed 2-to-1, doesn't have to worry about it.

It turns out to be possible to get some of the benefits of asparagus crossfeed by throttling the core engines down while the boosters are running -- the outer tanks thus empty first because they're consuming propellant faster. This is much easier to engineer -- no crossfeed plumbing, just throttlable engines. This is what Falcon Heavy actually does. Installing more or larger engines on the outer boosters would be generally equivalent, as well.

  • 2
    $\begingroup$ If you want to empty and drop off S4 first, its tubopumps have to have 4x as much power as the S1 turbopumps (S4 has to be emptied at 4x the rate of S1). $\endgroup$
    – Hobbes
    Nov 6, 2018 at 19:04
  • $\begingroup$ 3x for crossfeed, 1x for its own engines, but that's only if you need the core to stay 100% full -- crossfeed can still be useful at less than 100%. Your point is taken, though, and I've edited. $\endgroup$ Nov 6, 2018 at 19:13
  • $\begingroup$ @RussellBorogove "To maintain proper tank pressurization, the crossfeeds have to be pump-driven" Are you sure this is true? If all the tanks are connected at their bottoms (with large enough connections as to not impede flow), and only the ones you are trying to empty are being fed ullage pressurant, it should cross-feed right?. $\endgroup$
    – Lex
    Nov 6, 2018 at 20:40
  • 3
    $\begingroup$ Ahhh, okay, the pumps need to move as much propellant, but not achieve the same pressure, as engine turbopumps. $\endgroup$ Nov 9, 2018 at 12:58
  • 1
    $\begingroup$ Yes, pump power is proportional to massflow and pressure difference. You may not need the additional pump if the inducer of the main pump can handle the reduced inlet pressure. Several engines, such as SSME and several Russian engines have a boost pump spinning at a different speed. That allowes them to work down to almost a vacuum. $\endgroup$ Nov 12, 2018 at 1:38

Asparagus staging is possible, but not viable.

Mainly because pumping fuel around at the speeds and volumes needed for launch is a really big job, but also partially because the actual tank dry mass we achieve is so much better than KSP does, that the benefit is much reduced.
KSP tanks are about 10% dry mass(tanks only, no extras), thus ropping them provides very solid gains for your rocket. The Falcon 9 first stage with engines, avionics, gridfins and legs still manages dry mass of only 6.5%. For bare naked tanks the dry fraction will be much less, possibly even under 2%.

The proportional gain from dropping just a tank would be....minor. Not worth it, compared to the mass, intricacy, power needs and introduced failure modes that come with pumping several tons per second of cryogenic liquids around.

Nor, apparently, is even the somewhat more benign "onion" staging viable.
The closest we've ever had to this was the Space Shuttle's external tank, which was sort-of a single-tank onion system, except that the orbiter did not carry any fuel for the SSME's onboard, so it's more of just an external tank not even a drop tank.
SpaceX did take a serious look at implementing crossfeed on their Falcon9Heavy, this would have been a basic onion staging system, but have decided that the problem is either undoable or more likely just not worth doing.

What is worth doing is the sort of thing that the Delta IV Heavy, and the Falcon9Heavy do: Run the launch with full throttle on the side boosters, and reduced throttle on the core booster. This allows the rocket to stage off the side boosters on depletion, while still having some fuel available on the core booster. This is a compromise solution that provides some of the benefit of crossfeed, but completely sidesteps the added complexity and risk of the plumbing needed.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.