Experiment to get more efficiently off the ground: reverse asparagus configuration

[Cirocco]

Hi guys,

I'm very new to KSP (got it last week), but have been doing extensive research on the forums and other sources on how to get the most bang for your buck out of your rockets. Obviously I very quickly encountered the asparagus configuration where you progressively siphon fuel from the outer engines to the inner engines and break off the empty ones symmetrically. I like this configuration a lot and in space or high orbit, it works like a charm. However as an initial stage, I find that while it is extremely fuel efficient, it lacks the raw power (I believe that's delta v?) to get you out of Kerbin atmosphere quickly. Mostly because it dumps two engines rather fast and the longest time out of the entire flight is spent on one engine. Granted, at that point you should have cleared atmosphere and be using that one engine in a vacuum, but still, the quick dumping of two engines greatly decreases the early thrust which is what you want for escaping the atmosphere.

So while being bored at work and thinking about my rocket designs today, I came up with the following: what if you used a configuration that focuses on keeping as many engines live and firing for as long as possible? So I came up with the following: a reverse asparagus design.

How does it work? well much like the asparagus design, it consists of a center tank with a number of side rockets. Each rocket is linked to the stage above it by a stack decoupler. The center tank feeds fuel to all the side rockets at the same time, and the side rockets pump fuel from one to another in the classic asparagus design. So when you lift off, the center tank will run out fairly rapidly. Because of this, it's best to have a big center tank with a big engine (or even several medium sized ones by using the tri- or quad-coupler). Your first step is to ditch the center tank. This is quite a hairy manoeuvre and requires enough clearing between the center tank and the side rockets (make sure to use TT70 decouplers rather than TT38 ones in the stage above this one) but once the center tank is gone, you are left with all side rockets firing and can simply peel them off in a symmetric fashion as you would with regular asparagus. The biggest upside to this configuration is that every stage after the initial one has 1 more engine burning than the classic asparagus configuration. And your final and longest step uses 2 engines the entire time instead of 1.

Note that this only works as a lower stage because it requires you to dump the center tank first and then peel off in the classic asparagus design way. You need those side rockets to still be attached to something at the top when the center tank falls away. Also, you'll need struts linking the outer rockets to each other to prevent them from swinging wildly out of control and colliding. The struts should break easily as you separate them off though.

I've done some initial launch testing, and so far results were very promising. The reverse asparagus reaches an apoapsis almost 50% higher than the regular asparagus configuration on identical vehicles (the only difference being the direction of fuel pumping on the fuel lines and a few extra lines on the reverse one) when shooting straight up from launch.

I'm a total noob though, so if the more experienced players could give me some idea as to whether or not I'm on to something, give some ideas for more thorough testing and tell me whether or not this has been done already, that would be awesome

Keep those rockets firing.

Cirocco

[Soda Popinski]

Pictures might help. Conceptually, it sounds similar to a slack tank asparagus system, where you loose mostly tankage first, with no (or in your case) few rocket engines.

[Tank Buddy]

Asparagus doesn't lack raw power, as a matter of fact, it probably has more Delta V than any other staging method.

(I don't think you have the right idea of Delta V, Delta V is the change in velocity your craft can achieve by burning fuel. You need to change velocity to get into transfer orbits, landing, plane changes etc. It is pretty much a measure of the range and capabilities of your craft, not power, as one engine (in orbit at least) provides more Delta V than multiple engines)

The only downside is you loose thrust every time you drop a stage, therefore you have to manage acceleration.

[RogueMason]

You know, I was actually pondering this sort of design just the other day. While I actually have no need for such a rocket, it would be an interesting engineering challenge that I may try out at some point, just because I can

[Cirocco]

I'll see if I can post some pictures up tomorrow. I'm still very much in the concept testing phase, but so far every single rocket with this design has turned out pretty awesome. Mind you I've only tested it with rather small fuel tanks, not with the giant Roccomax (or however it is spelled) ones.

One big problem is the inherent collision risk with stack-decoupling long stacks: they can easily smash into each other. Still, so far only one launch has gone catastrophically wrong this way and that was because I didn't wait until the rocket stabilized after the loss of thrust when engines died.

As I said, will try to put up pictures tomorrow, won't have time today

[draeath]

I find a hybrid to work well - a few engines that you keep until the last minute, but you're shedding "drop tanks" the whole time.

[FenrirWolf]

You actually don't want too much power (or TWR, more accurately) in the lower atmosphere since you'll exceed terminal velocity.

If losing outer engines as you stage is causing you to lose too much thrust then you probably need to alter your engine configuration.

Since the engines in the central stack of an asparagus lifter will be the ones you hold onto the longest, it can be beneficial to have those be stronger than the individual boosters along the sides. An example would be having a skipper as your central engine and LV-T30's for the side rockets.

I'll have to some digging around to uncover an old post by this guy who distributed a rather nice series of lifters called the zenith family. It explains what trying to get at much more elegantly.

Edited January 7, 2014 by FenrirWolf

[nholzric]

SRB's are often the cure for low power on or during launch. I use a ring of SRB's at launch so that I can keep my asparagus stages at 70% throttle for the first minute or so. Speaking of SRB's, being able to limit their thrust is awesome - go tweakables!

If the rocket is underpowered after I've dropped all the outer stages, I fly with a delayed the gravity turn. This gives me more time above the atmosphere to reach orbital speed with the underpowered engines.

[6.forty]

I find a hybrid to work well - a few engines that you keep until the last minute, but you're shedding "drop tanks" the whole time.

I do this, except I take "shedding tanks the whole time" a bit further than I probably should.

[Soda Popinski]

Asparagus doesn't lack raw power, as a matter of fact, it probably has more Delta V than any other staging method.

(I don't think you have the right idea of Delta V, Delta V is the change in velocity your craft can achieve by burning fuel. You need to change velocity to get into transfer orbits, landing, plane changes etc. It is pretty much a measure of the range and capabilities of your craft, not power, as one engine (in orbit at least) provides more Delta V than multiple engines)

The only downside is you loose thrust every time you drop a stage, therefore you have to manage acceleration.

But power can effect delta-V while still getting into orbit. There's gravity drag, so if you have too low of a TWR, you're not climbing fast enough, and lose delta-V to gravity drag. If you have too high a TWR, you lose delta-V to atmospheric drag (going faster than terminal velocity). So, power (thrust) does effect delta-V, or at least delta-V usage.

[Cirocco]

If losing outer engines as you stage is causing you to lose too much thrust then you probably need to alter your engine configuration.

I don't lose thrust as such, but the lower stack of the rocket wobbles for 2-3 seconds after a stage runs out of fuel because of the sudden change in thrust due to the engines dying. Because you have no center stack anymore (as that is the first to go), the rocket is hollowed out and lacks a bit of stability. This means that in the reverse asparagus, you need to wait a few seconds after the engines burn out to give the rocket time to stabilize before you decouple. If not, collision of the stacks is almost assured. This is the biggest problem I can see right now, and I'm trying to resolve it by trying to add more stability to the outer rockets, especially the final 1-2 stages. But because everything else has shedded already, including the center sack, I'm having to resort to more unorthodox means than simply adding struts to the center mass.

Many explosions are sure to follow in the test stages, but hey, we're Kerbonauts dammit!

[G'th]

Couldn't you achieve basically the same thing (or perhaps drastically outperform it) by simply clustering the engines you'd have on the drop tanks onto the center tank? It'd save you a lot of potential headache when your reverse asparagus system decides to want to go out of control

[Alistone]

You have this idea to climb out of the atmosphere as quickly as possible. That would minimize gravity losses (the loss you have from using your thrust to fight gravity; the clearest example is when you are hovering and constantly burning fuel while getting no change in velocity)

But there is another factor which is air resistance. Air resistance force is proportional to your velocity squared. If you travel 3 times as fast you will encounter 9 (3^2) the air resistance. You will also escape the atmosphere 3 times faster. So you will have 1/3 of the gravity loss, but you will have 9 times the air resistance loss.

It turns out that the air resistance loss would significantly greather than your gravity loss and your ascent would actually be very inefficient.

It turns out that the best balance is to keep your gravity loss and your air resistance loss to be equal to eachother. To determine what that speed is we use "terminal velocity". It turns out that an unpowered falling object will have a force of gravity pulling it down with the same force as the air resistance is slowing it down. Because the forces balance eachother out at terminal velocity, the speed of the falling object remains the same.

You are correct that you want more thrust at the very beginning to get from 0 m/s at launch up to about 100m/s which is terminal velocity below 1000m altitude. But to avoid losing dV due to air resistance you maintain about 100m/s vertical speed and slowly accelerate to about 260m/s over the next 11,000 meters.

There is a chart of terminal velocity on the wiki here http://wiki.kerbalspaceprogram.com/wiki/Kerbin

Often people use extra SRB for the initial impulse to get up to 100m/s quickly. But after that a very low thrust to weight ratio is adequate untill you get out of the thick atmosphere. After that a higher thrust becomes more efficient again because air resistance declines very quickly and as you noted in your original post, the faster you can reach orbital speed the less energy you lose fighting gravity.

From 0 to 100m/s you want as much acceleration as possible (under 1,000 meters altitude)

5 m/s/s would reach optimal speed in 20 seconds at an altitude of 1000 meters. Good

4 m/s/s it would take you 25 seconds to reach speed at an altitude of 1250 meters. OK, but you're not quite optimal.

3 m/s/s it would take you 33 seconds to reach speed at an altitude of 1650 meters. Bad, you really need more thrust.

From 100 m/s to 260 m/s you are accelerating slowly through thick atmosphere between 1,000 and 12,000 meters in altitude

That's about 100 seconds to increase your speed by 160 m/s

You only need about 1.6 m/s/s acceleration to maintain terminal velocity.

You suddenly can get by with a LOT less thrust. So jettisoning engines really doesn't hurt.

If we wanted to be really clever we would adjust our staging by adding more fuel to some of the stages and less to others to really balance out the staging. But the truth is our T/W ratio varies significantly within a stage as it usually puts the same thrust for the duration of the stage, but often with significantly less weight near the end due to all the expended fuel.

Just try to stay close to the terminal velocity and you'll do pretty good. I suspect that your tests may have results skewed because the velocity profile during launch was not optimal (did not follow the terminal velocity at given elevation chart).

[MockKnizzle]

The main benefit of traditional asparagus is that you drop unneeded engine mass during ascent. With proper asparagus design, you have a roughly constant TWR profile all the way up.

[Andersenman]

Simple and clear explanation, Alistone! I especially like this part:

From 0 to 100m/s you want as much acceleration as possible (under 1,000 meters altitude)

5 m/s/s would reach optimal speed in 20 seconds at an altitude of 1000 meters. Good

[...]

From 100 m/s to 260 m/s you are accelerating slowly through thick atmosphere between 1,000 and 12,000 meters in altitude

That's about 100 seconds to increase your speed by 160 m/s

You only need about 1.6 m/s/s acceleration to maintain terminal velocity.

Rules-of-thumb, scientifically sound enough and easy to grasp at the same time, making KSP enjoyable to play!

[tntristan12]

one engine (in orbit at least) provides more Delta V than multiple engines)

False! Multiple engines with the same ISP will have the same delta-v. You burn through fuel faster, but you get higher thrust. Delta-v is driven by ISP and mass difference. The more mass of fuel you take, the higher your delta-v. Power is represented by TWR, which needs to be sufficiently high when burning normal to a surface to minimize delta-v losses.

[maltesh]

False! Multiple engines with the same ISP will have the same delta-v. You burn through fuel faster, but you get higher thrust. Delta-v is driven by ISP and mass difference. The more mass of fuel you take, the higher your delta-v. Power is represented by TWR, which needs to be sufficiently high when burning normal to a surface to minimize delta-v losses.

You're forgetting one thing: More engines give you more dry mass, which lowers your mass ratio, and thus your delta-V. All other things the same, a spacecraft with five engines of a particular type will have lower delta-V capability than a spacecraft with one engine of that same type. How much lower is going to depend on what fraction of spacecraft mass the engine is.

[boolybooly]

Not false. If you are carrying extra engines they slow down the final velocity as they are extra mass, what they do is convert more fuel to thrust per second but the total acceleration for the crafts fuel supply is reduced by the engine mass. Some designs I dump lifter engines once in orbit but leave the tanks.

[mediobogdum]

You have this idea to climb out of the atmosphere as quickly as possible.....

Thank you very much for this post.

[tntristan12]

You're forgetting one thing: More engines give you more dry mass, which lowers your mass ratio, and thus your delta-V.

...how did I forget that? You're right. However, if you carry enough fuel the increase in dry mass becomes less important than the extra thrust you gain.

[Hyomoto]

This is something I've been pretty interested in too, and I used to be a bit more into elaborate asparagus-style rockets. But now I look at the problem a little differently.

Agreeing with above, the thing about having engines is depending on how you build your rocket, those engines might just be dead weight sucking up your fuel. A nice strong engine with descent ISP that can lift your craft is all you need, minus your payload, everything else can be fuel. So obviously you can play with multiple engine configurations but if you want to talk efficiency and cost, dropping engines is the most expensive, least efficient method. What I'll usually do now is reduce the size of my lifting stage, and place that fuel into drop tanks. The returns are diminishing, but TWR > 1.07 seems to work well for me.

Typically I'll do two sets of three, string them together and drop them as I rise. I may need some small engines on the third set of tanks but I've found intelligent rocket design (I know, KSP isn't exactly founded on that, but FAR and DE support it) is easier to handle and generally just does a better job. I couldn't find a really good screenshot of something I've already built, but this illustrates the most basic version. This ship has quite a lot of fuel on it, but the final tank is quite small in comparison, about 1/5 the total.

Edited January 8, 2014 by Hyomoto

[Kasuha]

I used this idea on the landing stage of my Tylo lander - it had one large central tank which it dropped when it was empty, keeping small fuel tanks and engines around it. But the point was - it needed all those engines because only when they were all burning and the remaining fuel tanks were half empty it crossed TWR 1 on Tylo.

For liftoff in atmosphere, too much TWR is as useless as too little TWR. To terminal velocity (which you can usually achieve quite fast) you can use any TWR your ship can hold, but after that, TWR 2 is optimum all the way to the upper atmosphere. Keeping your engines while dropping empty fuel tanks ultimately raises your TWR, meaning you either launched with too little TWR, or you have unnecessarily too much TWR and are burning fuel to warm up the atmosphere. Or both.