Why is it so hard to build rockets with dV > 10000 m/s

[Mat2ch]

Hi,

I'm trying to build a lifter for Eve, which shall be my next mission. I know that I need around 12000 m/s dV to get back from the surface, worst case. I want to go in style and bring a small party to the planet, so 3 lander cans is the minimum.

Now I'm trying and trying different designs and they all end up being huge monsters. It's actually pretty easy to build something with 6000-8000 m/s dV, but from then one it's hell. Lightest build so far had 400t. I have no idea how to get that into LKO or land it on Eve...

So, tell me, what did you build to get down to Eve and back up? I've found one thread, but the rocket there used a heavily clipped design and that's what I want to avoid. Traveling light, but in style I say

Edited April 14, 2015 by Mat2ch
I think I'm done here. Thank you

[Starhawk]

You need to look at the Eve Rocks Challenge thread.

Eve return is one of the most difficult challenges in the game. You definitely won't be traveling light.

Happy landings!

Edited April 14, 2015 by Starhawk

[Red Iron Crown]

The thread Starhawk linked is the most current compendium of Eve ascent designs. You can also check out Lightest Eve Lander for some lightweight, high altitude ascent vehicles, as well as Eve One for general design discussion.

[Mister Dilsby]

In direct answer to the question "why is it so hard", it's because the fuel requirement increases exponentially, not linearly, as you ad dV.

Cars are easy, their fuel requirement increases roughly linearly. If your efficiency is 6 liters/100km (~40mpg) and you want to go 200km, you carry 12 liters of fuel. If you want to go 500km, you need 30 liters. If you want to go 10,000 km--well, that's when it starts to go non-linear, as carrying 600 liters around is going to hurt your fuel economy. Up to that point it really doesn't matter, because over the ranges involved in surface transport it does not take many grams of fuel to move around.

But fighting gravity, ESPECIALLY Eve's gravity, costs loads of fuel. And it costs fuel to lift fuel. It's easy to build a ship that weighs about 10tons, has enough thrust to lift on Eve, and gets 1000dV. But then you have to add a stage to lift 10t of dead weight for the next 1000dV, and maybe that stage weighs 15 tons. Now you have 25t of dead weight to lift for the next 1000dV, and maybe that next stage weighs 30tons. Etc, etc, etc.

You probably know all this, but getting familiar with the math really helps. If you can do logarithms, you can solve dV equations by hand rather than relying on mods like KER. (see http://wiki.kerbalspaceprogram.com/wiki/Tutorial:Advanced_Rocket_Design)

As Starhawk and RIC mentioned, the Eve Rocks Challenge Thread is a great place to learn how others made an exponential series converge to a reasonable and finite number. Don't be afraid to steal someone's design ideas (and leave them +rep for it!), it's the ultimate compliment.

[Hannu]

Nature of the rocket propulsion is nasty. Mass of propellant is exponential function of dv. Every km/s multiplies the mass by certain factor. If we assume ISP 350 and typical masses of KSP parts, every km/s multiplies initial mass by a factor of about 1.4. It means that if you want to increase dv from 8 km/s to 12 km/s you need approximately four times as large craft.

[KerikBalm]

3 lander cans? send 3 ascent vehicles, not one with a capabity of 3

That will make it much easier already - practice your precision landings.....

Next: send the lander up to LKO unfueled/partialy fueled. Send a seperate launch up to just fuel it up.

Then send up the transfer stage to push the vessel to Eve.

Trying to do it in a single launch is a terrible idea.

4,500 m/s to LKO, 1,100 to Eve capture -> 5,600 m/s

12,000 m/s for eve ascent -> 17,600 m/s

That's quite the requirement!

Its even worse if you don't do orbital rendevous at eve, and your eve ascent craft needs another 1,100 or so to get back to kerbin.

18,700 m/s in such a case.

18,700/4,500 = 4.155

If you look at the launcher-ception thread, you basically need to be able to do 4 of those.

A launcher ot get 3 to orbit -> a launcher to get that to orbit -> a launcher to get that* to orbit -> a launcher to get that** to orbit....

What is your mass fraction for your best rocket launcher? 20%

so each launcher underneath is 5x bigger? 5^3 = 125x the size of your simple 3 man to LKO launcher.

So yea... its tough.

[Mat2ch]

Oh, wow. To return with three Kerbals from Eve it really needs a huge rocket... I always thought I'm doing something wrong, but apparently I only have to fiddle a bit more around to create something that will work.

Thank you for your answers!

- - - Updated - - -

3 lander cans? send 3 ascent vehicles, not one with a capabity of 3

That will make it much easier already - practice your precision landings.....

Next: send the lander up to LKO unfueled/partialy fueled. Send a seperate launch up to just fuel it up.

Then send up the transfer stage to push the vessel to Eve.

Sending three ascent vehicles instead of one is a great idea. I'll try to build one and see how it behaves. As it will be a lot lighter it will be easier to land anyway...

And I already thought about a tug ship. In fact I had one with 12000 m/s dV in orbit... don't know how much this will be after I connect three ascent vehicles...

[Streetwind]

There is a limit to how much dV a single stage can produce. This limit derives directly from the fuel mass fraction of the fuel tanks used in that stage.

Without going into detail here (though I could if you wanted to see math), that limit for LF/O in Kerbal Space Program is 21.5 times the stage's Isp. That's the theoretical maximum, for carrying the entire observable universe's mass in fuel while using only massless parts for everything else, including a massless engine. In practical application, you'll find 20 times impossible to achieve and 15 times still a major chore. Most rocket stages in KSP have around 10 times their Isp in dV, which is a reasonably efficient middle ground between the diminishing returns of staging too often and the diminishing returns of trying to cram too much fuel into one stage. For really high performance cases like Eve return missions, though, you probably want to shift the balance towards a higher dV number per stage simply because you'd like for things to remain somewhat compact.

[OhioBob]

I just successfully completed my first manned launch from Eve. It took what essentially amounted to eight stages, giving me a total ÃŽâ€v of about 12.8 km/s. In simulations I determined it could achieve Eve orbit from an elevation as low as 500 m, though I launched it from about 900 m. From that elevation I got into a 120 km orbit with 200 m/s to spare. All I launched was a Kerbal in a command seat and my total liftoff mass was 83 tonnes (the landed mass was 91 tonnes). The total delivered to orbit was about 320 kg. I also did an unmanned practice launch using a smaller design (also capable of 12.8 km/s). The unmanned version delivered a little over 200 kg to orbit with a launch mass of 51 tonnes.

In both cases I'm getting a payload fraction of only about 0.004. There are probably others who have done better than that, but the point is, it takes a huge vehicle to launch a small mass from Eve. Each additional m/s of ÃŽâ€v is very, very hard to attain.

(edited to add)

The 12.8 km/s ÃŽâ€v is based on vacuum Isp. It's more like 12.1 km/s in actual practice.

Edited April 14, 2015 by OhioBob

[EtherDragon]

Since Street got me going on it... it's math time with SimGamerTV!

Dv = Isp*G*Ln((M+P)/M)

The thing about Dv is that it's completely dependent on the Full to Empty Mass Ratio, i.e. (M+P)/M

As Streetwind eluded to, there is a theoretical max for this ratio - it's ~9.33:1. That is, the best you could ever hope for (in KSP), assuming no other parts have mass, is:

Example A.

Ln(9.3/1) ~ 2.23

2.23*Isp*G = ~17,483 Dv for the best engine in the game (The LV-N).

But to get close to that, you have to have only one engine, and so many fuel tanks that the cockpit and rocket motor have a negligible fraction of the weight. (hint, it would be MASSIVE, and have a thrust to mass ratio so low it would be practically useless)

If you used a rocket motor with more umph, you get more thrust, but less ISP. The theoretical best (assuming a 9.33:1 full/empty ratio) looks like this:

Example B.

Ln(9.3)*370*9.8 = 8,099 Dv. A far cry less than 10,000.

Now, you can get over 10,000 by adding staging, but doing so requires you to add the entire upper stage as part of the empty mass of the stage below.

But if we turn the math around for the bottom stage (let's assume the upper stage looks like the one in example we need only 2000 more Dv. So we get something like this:

2000 Dv = Ln((M+P)/M)*9.8*370

0.55 = Ln((M+P)/M)

1.74 = (M+P)/M

Which means, for every ton of full mass, you need 0.74 tons of fuel for the lower stage.

TL/DR?

The reason it's so hard is due to the diminishing returns found in the Ln((M+P)/M) function, coupled with the theoretical maximum Full/Empty Mass Ratio.

Edited April 14, 2015 by EtherDragon

[Red Iron Crown]

EtherDragon, a small correction to your otherwise great post: The best mass ratio for fuel tanks in KSP is 9.333:1, for some of the Mk3 tanks.

[EtherDragon]

EtherDragon, a small correction to your otherwise great post: The best mass ratio for fuel tanks in KSP is 9.333:1, for some of the Mk3 tanks.

I reject your reality and substitute my own!

(Or I could go and edit my post, and update the math...)

(Edit: Edit done.)

Edited April 14, 2015 by EtherDragon

[Laie]

Generic advice re: Eve landers:

What makes large rockets prohibitive usually isn't the sheer size but the part count. FLT-800 + Aerospike is the basic building block for most Eve lifters, but if you take that to it's logical extreme, you first stages will run for like 5 seconds each. You may want to try Skippers with Rockomax-32 tanks at least for the outer stages. Performs well and has no overheating troubles. If you need even bigger stages, the Kerbodyne Quad Cluster isn't half bad.

Stay away from Mk3 parts, or get the Joint Reinforcement mod. In the stock game, vessels built of Mk3 parts tend to fall apart at the slightest provocation; trying to land them on Eve is frustrating to the highest degree.

Most Eve landers can SSTO to LKO if you increase their fuel capacity by ~20%. Someone who's more math-savy than I am could probably tell you why this is so; I only know from experience that adding a small number of large tanks will take you to orbit in one piece. Refueling them in orbit usually isn't much of a problem; the big big tanker may be 150 parts on the pad, but by the time the vessels meet in orbit the tanker will be like 7 large tanks plus perhaps 20 bits an pieces (rcs, probe core, etc). The real problem with oversized vessels is that refueling them requires you to do three or five runs.

Eve is relatively close. If you're constrained by part count, you may find that just slapping on a few more tanks (like, re-using those from the launch) and doing the transfer on the lifter engines is totally affordable, while adding Nukes for the transfer will either lead to a very low TWR, or drive up the part count, or both.

The most lightweight way to carry a kerbal is the external seat; next comes the one-man lander can. Three lander cans weigh much less than a three-man pod, and accordingly, the lifter can be much smaller. Try it, the difference is huge.

You don't need many parachutes if you're willing to do a powered landing, which is not as difficult as it sounds. 150m/s worth of fuel is plenty. Carrying a few chutes ensures that you're pointed upwards and don't have to worry about steering, which helps a lot.

The best landing gear, ever, is the Liquid Fuel Booster. There's other engines that have a high impact tolerance, but chances are that they will break off under the landers' weight. The LFBs don't suffer from the same problem, presumably because of their length. You don't even need to add any more struts than usual. In my last designs, I ended up using the LFBs as a) extra tanks landing gear c) mount point for chutes and reaction wheels, but d) never ignited their engines.

[Mister Dilsby]

The best landing gear, ever, is the Liquid Fuel Booster. There's other engines that have a high impact tolerance, but chances are that they will break off under the landers' weight. The LFBs don't suffer from the same problem, presumably because of their length. You don't even need to add any more struts than usual. In my last designs, I ended up using the LFBs as a) extra tanks landing gear c) mount point for chutes and reaction wheels, but d) never ignited their engines.

That bit surprised me the most, I never would have thought of that. ^But listen to this guy, he has seen more Eve ascents than anyone else on the Forum.

I did find that the aircraft gear bay was quite a good landing gear as well; I had less than a meter clearance to my engine bells and landed on an 8 degree slope with no damage. Others have done great things with a stack of cubic octagonal struts. Probably the worst landing gear you can pick are the actual so-called "landing legs".

[Laie]

I've not only seen Eve ascents, but done plenty enough myself. Lately, I've been experimenting with low part count landers. Seems as if going above 6-8 asparagus stages yields only small gains in terms of efficency.

(LFBs as landing gear)

That bit surprised me the most, I never would have thought of that.

20m/s impact tolerance. You can slam down the vessel really hard, using only very few chutes and no thrust at all. It looks scary but is totally safe, if have enough ground clearance for your other engines. If you have the rest of the vessel stacked on top of the LFBs, it's advisable to slip in reaction wheels. Those seem to work as suspension, much like the cartilage discs in your spine.

The main benefit of LFBs as landing gear is part count. If you optimize for weight, cubic struts will serve you better. LFBs also aren't bad for a first stage. Yes, their ISP is abysmal, but for a first stage that hardly matters. Most of the first stage fuel is burnt in later stages' engines, after all.

[rockbloodystar]

http://forum.kerbalspaceprogram.com/threads/114250-Eve-misson-tips

paying it forward.

[Kelderek]

Generic advice re: Eve landers:

The most lightweight way to carry a kerbal is the external seat; next comes the one-man lander can. Three lander cans weigh much less than a three-man pod, and accordingly, the lifter can be much smaller. Try it, the difference is huge.

If you want to take 3 kerbals to Eve with a single container, then use the hitchhiker can (holds 4 kerbals), it only weighs 2.5t. That is what I used for my Eve Rocks submission a while back (see my signature).

I just ran another Eve mission with the three orange-suit guys using 3 of the small landing cans. This was not for the challenge, I just did it for fun with a sandbox game. I'm thinking this design may be worse than my first one, but it also got the job done with 3 kerbals. This is mostly stock with MechJeb because I am lazy and go AFK a lot.

Javascript is disabled. View full album