It occurs to me that "Asparagus Staging" has become a common term here...

[softweir]

...

I mean, seriously. Why drop the tank AND the engine?

You burn the fuel, you drop the tank - and suddenly the craft weighs less and needs less engine to keep accelerating. So why waste fuel carrying that engine even higher?

As you probably know, the Saturn 5 first stage had 5 engines and shut one down towards the end of its burn to keep acceleration below dangerous levels. NASA seriously considered dropping the middle tank at the time it shut down. This would have allowed them to shut it down even earlier, and meant the first stage would have achieved even greater altitude and velocity in the last few seconds. And all this in spite of the fact that they weren't dropping tanks, just burning huge masses of fuel.

EDIT: AN afterthought: The proposed Falcon 9 Heavy will have the following phases:

1. Launch: All three engines firing at full power
   
2. MaxQ: inboard engine throttles back to prevent breakup due to aerodynamic forces as it transitions to supersonic
   
3. Outboard tanks and engines jettison, inboard engine at full throttle
   
4. Inboard engine throttles back
   
5. 1st stage dropped, upper stage ignites

Note phase 4: By this time, the inboard tank and upper stage are now so light that the inboard engine is throttled back, partly to improve efficiency, partly to reduce excess acceleration. If the Falcon had carried those outboard engines all the way up here, the whole lot would now be too heavy to get into orbit!

Edited December 6, 2013 by softweir

[magnemoe]

That already happens with serial staging. Otherwise the Saturn V would bring the entire stack to Lunar orbit.

As to the issue KSP currently leading to unrealistic asparagi builds, it must be fixable, since FAM + Ferram's Kerbal Joint Reinforcement mod (specifically so FAM works better with realistic vertical builds) does a pretty good job.

Squad has stated they want to put in a more realistic drag model. If they do something similar to FAM and RL, where cross-section matters, we may see the end of asparagus over use.

Yes this will solve two issues, first an narrow rocket will work better it will also make it stable, today you want an wide rocket to keep it stable.

It does not solve the low TWR and heavy tanks who is the main reason why asparagus is so important.

And yes, we need the option to just throttle back the central stage while running the boosters at max power during accent.

[Deathsoul097]

@Thescientist

Yup. That's pretty much the gist of it.

[Ryu Gemini]

Indeed. If we assume (yes, this is overly simplistic and likely inaccurate to some degree) that an engine of 1 weight produces 1 thrust...

For serial staging: Lets say you need 5 thrust at launch. Later you need 3, and the 5 engine was already dropped with its tanks. The last stage needs 1. You need 8 weight worth of engines, because the ones at the higher stages cannot be fired until the previous stage is done.

For asparagus, You have 5 engines. You drop off 2 at a time. End result, you only need 5 weight worth of engines. All the engines can be fired from the get go. No need for extra engines that have to "wait their turn."

Naturally, this does NOT take into account air resistance, and the weight of the fuel pumps. So it becomes a question of whether having fewer overall engines and associated systems on the rocket outweighs the downside presented by the added air resistance and the logistics of fuel pumps. The Falcon Heavy's designers seem to think it viable for that craft. It may not be viable in all circumstances though.

On a side note, if air-breathing rockets like that SABRE start becoming viable and common, then I could see a bit more looking into asparagus staging, since said engines do not need to pump oxidyzer, only fuel. And because you have unavoidable drag to a degree since you NEED the extra cross-section to take in air for the engines (thus partly mitigating the drag issue of real-life asparagus design, since you would be forced to have some extra drag regardless). In fact, if you ignore that fuel pumping benefit I mentioned and design it to cross-feed the air too, then the sabre engine could potentially work to a higher altitude if you detach the empty tanks, but KEEP the top part of the side-stage (which would be the air intake and air cross-feed part), such that you would supposedly have 2 extra cross sections of air intake feeding the center rocket's SABRE, increasing the altitude it would be viable to.

That may require some thinking as to how to make that realistic though.

[magnemoe]

For an sabre its thee steps you could use to increase performance easy.

First is a launch trolley, two benefits you save weight on landing gears as they are only used then you land empty. it might have its own engines to help get the plane up to takeoff speed.

second is mid air refueling. replace fuel used during takeoff and let you launch with more fuel than you could take off with.

Last is then launching satellites to higher orbits than LEO you need an extra stage for this, now it you make that stage larger and only goes suborbital with the plane you could increase weight to orbit a lot.

[KASASpace]

I know a guy who calculated the D-v of a realistic asparagus rocket. It weighed less than the SLS NASA is working on, had more payload, and only used RP-1 instead of LH2. So, very advantageous to anyone who wins the race to asparagus

[KASASpace]

Well, actually, asparagus tech already exists. Space shuttle ET fuel lines into Orbiter. Worked for 30 years, so maybe on asparagus rockets? But hey, you wouldn't NEED aspargus if you could add a vertical track that accelerated the rockets up the first few 100 meters going an extra few hundred meters a second. That would be cheaper and more practical.

[NovaSilisko]

There was a good post on NASASpaceFlight about why asparagus is very impractical. Mostly because piping fuel around a really, really, really non-trivial matter IRL, and Falcon Heavy's using of it is still only a concept. The shuttle was much different, that was just the same as if the tank were above the engines - the piping just went through a detachable truss, whereas FH requires a split in the lines and has to make sure that the right amount of fuel gets to all 27 engines at the right time... fuel starvation, or even just a bit of a slack in the amount of fuel coming in to the engine is a bad thing to have happen.

Also, I just want to say... fuel flowing from one thing to another is NOT the same as asparagus. Asparagus staging is when you have at least two pairs of boosters, with the second pair feeding into the first, and the first feeding into the core. So neither FH or the shuttle really qualify IMO

Sorry, I kind of have a peeve about all this. In some ways I feel like KSP has given a lot of people the impression that real-world spaceflight is much easier than it really is... as an example, I seem to remember seeing someone say we should "just" send up a Soyuz to go fix Kepler, for instance. Even if Kepler was in earth orbit (it's not), soyuz would have no way of docking to it, or grappling it, or conducting an EVA, or storing the required replacement gyros, etc.

[Catullus]

People forget that the main development which made space rockets possible was the turbopump. If you don't have a way of pushing fuel around fast enough, you're not going anywhere. Anything that gets in the way of that causes problems.

[agentKmurph]

I know a guy who calculated the D-v of a realistic asparagus rocket. It weighed less than the SLS NASA is working on, had more payload, and only used RP-1 instead of LH2. So, very advantageous to anyone who wins the race to asparagus

I wouldn't deny any of these claims on paper. However, as multiple people have stated, the logistics of building such a rocket is no trivial matter. The "realistic rocket" you speak of faces a daunting challenge of being able to actually pump all that fuel around and do it without upsetting the balance or throttling of the rocket. Whatever gains you could realistically get would be offset by a hell of an engineering conundrum.

I would think that given the challenges of building a true asparagus staged design, it might be best to find a compromise. A parallel staged rocket might be much more viable, where all radial boosters feed centrally rather than into each other and eventually to the central tank. This would distribute the central-feeding fuel flow across two or more parallel stages meaning that the flow rate at any given valve / junction between radial stage and central stage would be a fraction of feeding a whole booster. In this scheme, you would never have dead engines and the fuel flow at the junctions could potentially be small enough in magnitude to be acceleration-fed. You also wouldn't have to face the challenge of building a rocket that stages 3+ times like you would in true asparagus scheme.

[rdfox]

For the record, the biggest advantage in parallel-staging designs like the Atlas V Heavy (which doesn't use fuel transfer) and the Falcon Heavy (which does) isn't in drag, TWR, specific impulse, or anything directly related to the rocket's performance.

The big advantage is that, rather than building a special large first stage for heavy-lift payloads, you can use a cluster of medium-lift core stages to lift the same mass. This allows economies of scale to enter into effect, as you don't have to spend money designing, testing, and certifying the new heavy-lift rocket, nor do you have to set up a production facility to build it or any of the other infrastructure involved. Instead, you can just build a lot more of your existing, paid-for medium-lift booster core, thereby reducing the cost of the cores by increasing the number of units being built. So basically, the benefit is that you can get a given amount of payload into low orbit for a lower price tag than building a single big core would cost.

Note that even NASA considered this back in "evolved Saturn" studies in the late 60s; there was a "pie in the sky" proposal that would have launched a manned Mars mission in a single throw by using four(!) Saturn V first and second stages as a clustered-core booster, plus possible strap-on SRBs, to launch the required transfer stages and Mars spacecraft all at once. Total payload to LEO was calculated as being somewhere in the neighborhood of 500 tons(!) or more...

[Klopchuck]

The term originates in Tom Logsdon's "Orbital Mechanics: Theory and Applications" (1998); he calls it "Ed Keith's Innovative Asparagus-Stalk Booster"

[mrfox]

The UR-700 used such a concept

The RD-270 engine was used in all nine modules of the first and second stages, operating at 103% thrust at lift-off. At lift-off all nine RD-270 engines would fire; the engines of the second stage would feed from propellant tanks in the forward section of the first stage modules. Therefore at separation of the six first stage modules, the propellant tanks of the three second stage modules would still be full.

http://www.astronautix.com/lvs/ur700.htm

[Spartan125]

I don't know if this has been suggested before but instead of the pumps feeding into the other tanks they would feed directly into the other engines. When they are staged all the engines shut down then the side boosters are separated then once the boosters are far enough away the engines on the next stages reignite and the open places on the engines (where the pipes were pumping into) are sealed off. This may reduce the issues on fuel pumping but sealing off the original pump line may be extremely dangerous and unreliable and to reignite the engine may make the heavier. Aerodynamic drag shouldn't be much more of an issue than that of rockets that use side-mounted boosters and the boosters would be tightly packed.