I love building asparagus rockets in KSP but don't see them much in real life.

[Kasuha]

The main unnatural thing in KSP is IMO the fuel pipe. A connection which allows to pump insane amounts of fuel between fuel tanks and closes safely when the tank is disconnected is a major safety hazard in any real situation.

Yes I know it was used in space shuttle. But there was just one, not the numbers used by players in KSP.

Without fuel pipe there's no asparagus as you lose all its advantages.

[Melcom]

I think, money is the main factor IRL. Every design needs to be financed. If a Delta IV Heavy works with normal staging, why to develop much more complex systems for the same job. If more starts go to private firms like SpaceX, money will even more be the factor for design.

Didn't said Von Braun something like "Keep it simple and test it."?

[Urban_K]

It's true that the Falcon Heavy that has been mentioned in this thread will use some basic variant of asparagus staging (two side stages that pump their fuel into the center one).

Only if someone pays for it. Normally it'll work much like the Delta IV Heavy by throttling down the centre: http://www.spacex.com/falcon-heavy

A Falcon Heavy launch with fuel crossfeed will cost a lot more, 75%: http://www.spacex.com/about/capabilities

[Jarnis]

Only if someone pays for it. Normally it'll work much like the Delta IV Heavy by throttling down the centre: http://www.spacex.com/falcon-heavy

A Falcon Heavy launch with fuel crossfeed will cost a lot more, 75%: http://www.spacex.com/about/capabilities

It doesn't cost more because it costs more to do this. It costs more because the payload is larger. In rockets you do not pay based on how much it costs do do something. You pay by the payload mass delivered to orbit. Crossfeed one delivers a lot more.

Also the higher price for bigger payload on FH also covers the fact that bigger payload will go with expendable FH that will not return boosters. You pay for the fact that at least two out of three cores just turned into scrap rather than landed back.

[Nao]

Adding to the many good comments here. The reason for asparagus staging is so good in KSP is that both engines and fuel tanks themselves are much heavier in KSP. Just by making simple 3 stack one stage configuration into 2-1 two stage asparagus increases the Dv by around 24%. Comparably, real world engines are ~6x lighter and fuel tanks are ~4x lighter, that Dv increase would be less than 5%. And thats not counting additional pumps and fuel line mass (that would probably bring the Dv increase to a whooping ~4%). The additional costs for the complicated fuel pumping system isn't just worth the saving.

Also from what i've read "asparagus" staging on falcon V heavy is only going to be used to lift the most heavy payload. The asparagus system is just a way of maximizing payload at a considerable cost that is still lower than making competently new design.

Edit: @Jarnis do you have any sources on the comparison of different versions of Falcon V? I'm somewhat out of the loop on this.

Edited October 22, 2013 by Nao

[Paul Kingtiger]

Another factor in real life (although minor compared to others mentioned here) is where all those stages are going to land. Kennedy Space Center is where it is so that staged rockets (or failed launches) land in the ocean. But even with that you can't afford to drop stages everywhere.

Asparagus is great as an engineering solution in KSP, but in the real world there are different challenges that require a different optimal solution.

[Jarnis]

Lots of good discussion on Falcon Heavy

http://forum.nasaspaceflight.com/index.php?topic=32528.15

Probably more than you can possibly have time to read

[LameLefty]

Lots of good discussion on Falcon Heavy

http://forum.nasaspaceflight.com/index.php?topic=32528.15

Probably more than you can possibly have time to read

Join the L2 portion of the site. There's more info there (including indications that cross-feed isn't even being actively worked on at this time, other than preliminary designs and possibly scarring of the cores for future plumbing/components). SpaceX has a LOT on their plate right now - F9H with cross feed is pretty far down the list.

[lazarus1024]

I hope this is in the game at release. It would put more of a emphesis on escape system planning, which, I always setup (Engines cut and detach, parachutes on crew compartments deploy). I mean hell, I spent time building, launching, and docking an escape pod for my Space Station, just in case.

No, SQUAD has said they'll never do random part failures.

[lazarus1024]

Another factor in real life (although minor compared to others mentioned here) is where all those stages are going to land. Kennedy Space Center is where it is so that staged rockets (or failed launches) land in the ocean. But even with that you can't afford to drop stages everywhere.

Asparagus is great as an engineering solution in KSP, but in the real world there are different challenges that require a different optimal solution.

Also new designs are starting to focus more on reusability of stages. With asparagus staging that adds a lot more complication to try to recover the stages. For example, SpaceX (IIRC) actual re-lights the engine on the first stage (or maybe second?) before it splashes down as well as pops a chute. Now figure the complexity of doing that with multiple, multiple tanks/stages and the odds that you are going to lose stages/tanks goes up.

I could see how a design that maybe uses drop tanks that are used up very soon after launch...maybe...but the complexity might just not be worth it.

Fuel isn't actually all that expensive in the end for a rocket. It is the rocket parts themselves that are the most expensive part of any launch.

Not sure on current price of LOX, but Kerosene only runs around $4-5 a gallon for the kind of stuff you put in a rocket (since that is what most use, though not all of course). If you rocket burns 100t of fuel plus oxidizer that is only around $150,000 worth of rocket fuel. If you gain 10% efficiency by changes in staging, that only saves $15,000 worth of fuel.

NOW, there is an argument for possibly getting 10% more payload in to space and cutting down on the number of launches you need to do...as the rockets and launches themselves (not the fuel) are darned expensive. However, you can potentially make up for some of that by moar rocket.

There comes a point though where it is difficult to make moar rocket, and then you need to do something else, like boosters or asparagus staging. In the past moar boosters was the way to go and it remains to be seen which one is most cost effective as technology improves.

Also until recently prototyping and testing was STONKING expensive and time consuming. Test even a modest sized scale model of a rocket with different staging and stuff might cost millions of dollars and take years. Then scaling it up and testing real parts and rockets would take many times more money and lots and lots of time. NASA and others are just starting to get good at rapid prototyping and making cool machines that'll allow you to 3D print complex metal objects, so they can do things like make one-off turbo pumps, valves, rocket engines, etc so they can much more cheaply test things like asparagus staging.

It doesn't save much money if you spend a billion dollars on a failed asparagus staging design to allow you 10% more payload in the same rough size, mass and construction/operating cost of a rocket. If it works and is reliable great, but if it doesn't work out, you wasted massive R&D on a failed design. Or worse you chase it for years and years spending billions to try to perfect a design to save you 1 launch in 10...or maybe save a couple of hundred million dollars a year in launchs...but spent billions.

[Scarecrow88]

Is it possible to turn on random engine failures?

I hope not. Imagine doing a 2 year mission to Eve, one of the most difficult planets to return a manned landing from, and you get an engine failure on your Eve lander just as you take off to return to Kerbin. What a complete pain in the proverbial that would be, not to mention trying to build craft that have contingency abort systems on every stage that will be used during the flight.

[HeadHunter67]

Is it possible to turn on random engine failures?

I don't understand why people are so eager to advocate un-fun "features" in games. The usual reason is to make things more "challenging", but I think it's possible to make a game challenging without randomly punishing players even when they do everything right. Would sports be more "challenging" if the refs tossed a coin to see if each goal was allowed?

For those who want random failures, feel free to go into your tracking screen at some randomly determined interval and terminate your active craft. But I am thankful that Squad doesn't want this either.

[Nao]

Thanks for the links Jarnis! After skimming some data, I've found that FH boosters will have propellant mass fraction of 0.966+ Which is crazy if you think about it. It's like a big fuel truck that can transport 40000L (30,8ton) of fuel while weighting as much as a small city car... with a power of a train... that can go to space!

This would make singe stage flight really efficient, although I forgot that Merlins have quite low ISP (compared to numbers on SSME or KSP) so I guess thats one of the reasons to use cross-feed on Falcons (they run closer to Dv limit so the Dv increase from staging is more significant).

[LameLefty]

Also new designs are starting to focus more on reusability of stages. With asparagus staging that adds a lot more complication to try to recover the stages. For example, SpaceX (IIRC) actual re-lights the engine on the first stage (or maybe second?) before it splashes down as well as pops a chute.

After staging, 3 of 9 engines are ignited and run for a short period to kill most of the velocity and allow a controlled re-entry of the first stage. One engine is re-ignited a second time very shortly before touchdown to perform a "suicide burn" and landing. SpaceX attempted this on their recent mission, shooting for a test "water landing" downrange, during the first launch of the F9 v1.1. Apparently the stage developed a roll during final approach, prior to igniting the center engine the second time. The roll exceeded the attitude control system's ability to control it. The center engine was able to be restarted but flamed out due to centrifuging of the fuel in the tanks. Musk thinks they'll be able to overcome this with some tweaks and once landing legs are installed on the stage (which will essentially act as fins to help control roll).

Edited October 22, 2013 by LameLefty

[sgt_flyer]

there was one planned russian concept which used an early form of asparagus - the UR-700,which was N-1 rival

[Kerbart]

Asparagus rocketry is basically why the Russians failed to beat the US to the moon.

Pogo oscillation is a potentially dangerous type of self-excited combustion oscillation in liquid fuel rocket engines. This oscillation results in variations of thrust from the engines, causing variations of acceleration on the rocket's structure, giving variations in fuel pressure and flow rate. That's why the N1 rocket wasn't successful. It had over 30 engines

The number of engines has nothing to do with pogo oscillation. Apollo 6 had it, and Apollo 13 had it pretty seriously but nobody ever talks about that because Apollo 13 had other issues during the flight that got more attention from the press, and pogo oscillation has been a known problem on earlier (one engine) rockets as well.

There was only one case of an N-1 where pogo oscillation was a problem and ironically that was the most successful N-1 flight (in all fairness it might have occurred on the other flights if they didn't blow up first). The other flights ended even earlier, because it is simply impossible to get reliability out of something that uses 30 engines, and because there were some serious engineering problems with the N-1.

Edit: as you mentioned, having 30 engines doomed the N-1. But there's little relation between 30 engines and pogo oscillation

Edited October 22, 2013 by Kerbart

[LameLefty]

The other flights ended even earlier, because it is simply impossible to get reliability out of something that uses 30 engines, and because there were some serious engineering problems with the N-1.

I bolded the part that's relevant. 30 engines isn't the key - it's 30 engines with the notoriously-poor Soviet quality control system, combined with the lack of an efficient engine control system to handle all 30 at once. I have no doubt SpaceX is going to do just fine with the F9H and its 27 engines (plus 1 on the second stage). But the number of engines, all by itself, isn't the issue.

[Jarnis]

Join the L2 portion of the site. There's more info there (including indications that cross-feed isn't even being actively worked on at this time, other than preliminary designs and possibly scarring of the cores for future plumbing/components). SpaceX has a LOT on their plate right now - F9H with cross feed is pretty far down the list.

I am in L2. I just can't link to those threads... and yes, it is still just something that exists "on paper". Still, I think they will eventually try it.

[Jarnis]

I bolded the part that's relevant. 30 engines isn't the key - it's 30 engines with the notoriously-poor Soviet quality control system, combined with the lack of an efficient engine control system to handle all 30 at once. I have no doubt SpaceX is going to do just fine with the F9H and its 27 engines (plus 1 on the second stage). But the number of engines, all by itself, isn't the issue.

Read "Rockets and People" volume 4 (available as free ebook from NASA website) and you learn a LOT more about N1. http://www.nasa.gov/connect/ebooks/rockets_people_vol4_detail.html

N1 was killed by a large number of errors...

- It had completely unrealistic schedule. In USSR nothing was done without a decree and a schedule. If one was set, it was done - even when it was completely impossible. So a lot of effort was wasted on completely unrealistic deadlines.

- It was being built without suitable test rigs. The only way to test the first stage was to launch the thing. The engines were not designed to be reused - they couldn't be "hot-fired" (or full-duration test fired) on a test stand and then flown - the first time the engines were used was the real flight. Oh, sure, they did test individual engines in test stands and hoped to know how the mass-produced engines behaved, but nothing beats actually test-firing every engine before flight (like SpaceX does it) or at least testing whole stages on test stand to see that they work as advertised. Soviets just took 30 mass produced engines, never fired them, built a whole stage complete with analog KORD control system to ensure that any failing engines would be shut down on the fly (it had multiple-engine-out capability), and hoped they all worked in a full-blown launch. The answer to their hopes was "nope".

- "No suitable test rig" was due to Soviet development style of "we'll just build a bunch, keep launching them and fixing things as they crop up" used in ICBMs. Except that they couldn't exactly afford to build a dozen N1s just to get one to work. Incredibly they built almost a dozen and flew four of them, full stack, with every single one going all Kerbal before end of 1st stage flight. One of them happened to take out the launch facilities, causing substantial delays as it was rebuilt. Last try got pretty close - about 7 seconds short of staging. Each one failed due to a different issue but at least three of those could have been debugged out of the system in test stand firings. So much for the plan of testing in flight..

N1 1L and 2L were test models

N1 3L - first launch attempt, engine fire, exploded at 12 km

N1 4L - never launched, parts used for other launchers

N1 5L - partially painted gray; early launch failure destroyed pad

N1 6L - launched from the second pad 110, deficient roll control, destroyed at 1 km

N1 7L - all white, last launch attempt; engine cutoff at 40 kilometres (22 nmi)

Rockets 8-10 were scrapped when the program ended.

Just as they were getting better engines (NK-33s, later used in Antares rocket) the program was shut down and efforts directed towards Energia-Buran - because Soviets had to have their own space shuttle and all that. Did an amazing job with that too, except that by the time it was ready, USSR was done for and the budget dried up.

Edited October 22, 2013 by Jarnis

[LameLefty]

Read "Rockets and People" volume 4 (available as free ebook from NASA website) and you learn a LOT more about N1. http://www.nasa.gov/connect/ebooks/rockets_people_vol4_detail.html

N1 was killed by a large number of errors...

(SNIPPED)

I know. I studied all of this years ago, when detailed info first began trickling out. For that matter, I was studying aerospace engineering when "Kremlinology" was still a relevant term (seemingly replaced these days by "SpaceX-ology," lol! ). Also, as you noted above and I stated out in passing, the lack of engine control technology really hurt the N-1 program. It should be realized that the NK-33 is not simply a 40+ year old engine shipped over to the U.S. and integrated into Orbital's rocket. Rather, they are pretty much disassembled, inspected and rebuilt, and equipped with a thoroughly modern digital engine controller and health monitoring system. This is what the N-1 needed but 40+ years ago wasn't available.

[Jarnis]

Hey, SpaceX-ology is the natural reaction to the secretive (understandable for a private company) cutting-edge space-related developments going on there. Lots of inquisitive and smart space nerds want to know more! And unlike what US govt. had back during the cold war, we don't have CIA and spy sats to get us info. Have to resort to very dated sat shots of test stands off Google Earth and "spy photos" snapped by other nerds who happen to visit SpaceX facilities

And speculation. Lots and lots of speculation. Mixed with over-analyzed statements by Elon Musk and carefully freeze-framed YouTube videos of launches and tests, plus some PR photos that SpaceX and/or Musk sometimes pass out. Image file metadata getting analyzed and videos getting digitally enhanced to fuel more speculation? You bet!

Edited October 22, 2013 by Jarnis

[Eric S]

A proper way to penalize asparagus staging in KSP, though one that would probably be somewhat complicated to model, would be to conserve the momentum of the fuel as it drains.

I have to disagree. I understand the conservation of momentum aspect. The thing is, yes, you get some angular momentum as the fuel enters the fuel line, but you'll get a corresponding opposite angular momentum when the fuel reaches the end of that fuel line. It's probably something that would still have to be taken into consideration because these are very large fuel flows we're talking about here, but given that the fuel lines would be full before the launch clamps release, the worst part is behind you by the time you leave the pad.

[Jarnis]

I have to disagree. I understand the conservation of momentum aspect. The thing is, yes, you get some angular momentum as the fuel enters the fuel line, but you'll get a corresponding opposite angular momentum when the fuel reaches the end of that fuel line. It's probably something that would still have to be taken into consideration because these are very large fuel flows we're talking about here, but given that the fuel lines would be full before the launch clamps release, the worst part is behind you by the time you leave the pad.

Yep. Realistic modeling would require...

- Making empty tanks a lot lighter. KSP stages have really poor mass fraction

- Making engines heavier

- Making fuel lines heavier and limiting the fuel flow per fuel line (or have multiple sizes of fuel lines with different specs)

[numerobis]

I thought the angular momentum issue was that if you start rolling at all due to exterior forces (which you will), the fuel on the outside gets angular momentum. When it is brought towards the core, the roll gets faster.

[EatVacuum]

Is it possible to turn on random engine failures?

Don't need it personally. My rate of design failures and staging screwups more than makes up for it.

But seriously, I was hoping that R&D would be about not just getting new parts but about improving existing. Do you go with the old reliable rocket engine you've had for years or put in a new, more efficient but untried one would be an interesting design decision. Random failures would make it worthwhile to have escape systems that aren't needed now. Other than the useful but unrealisticl revert to launch and F9 options of course.

On the main discussion point though, implementing a drag model where the cross section of what was pushing through the air and the aerodynamic design of the parts, especially those at the front and back of the rocket would go a long way to eliminating asparagus staging.

In KSP a really long thin rocket is harder to control than a short stubby thick one. Not sure that is so in the real world. It is easier to vertically balance a long thin pool cue or broom on your finger than it is a coke can or a ball.

But as it is the game mechanics actually seem to favour it. Not to mention the height of the VAB being a problem if you go too tall.

Edited October 22, 2013 by EatVacuum