When to and when not to use two-stage landers

[Captain Sierra]

I'm a bit of a KSP veteran. I've played since 0.20.2 and have 400+ hours, and even I don't know the answer here.

Where is it more efficient/economical/generally better to use a two-stage lander with a descent stage left on the surface vs. a single stage 'down-and-back' lander?

Before answering, please assume the following:

• There is no intention to re-use the lander
  
• We are not on Eve, Kerbin, or Tylo
  
• The lander will not be coming back with us (emptied of excess useful fuel and dumped before burning for return)
  
• This is in stock solar system, not Jumbo 32, 64K, RSS, or the like
  
• This is in sandbox mode, but funds can still be taken into account
  
• There is zero recovery on the lander itself
  

Edited December 29, 2014 by Captain Sierra
armagheddonsgw made me see some additional necessary clarifications.

[Anquietas314]

Interesting question If we're talking about career mode, and you're not going to try to reuse the lander, you should still at least try to bring it back to Kerbin to get a (partial) refund, in which case a single-stage lander would be best (of course, if you're going to leave it behind, whichever design is cheapest!). Personally I haven't had much reason to use a 2-stage lander even on Duna, but I hear with RSS apollo-style landers are pretty good

[Captain Sierra]

Interesting question If we're talking about career mode, and you're not going to try to reuse the lander, you should still at least try to bring it back to Kerbin to get a (partial) refund, in which case a single-stage lander would be best (of course, if you're going to leave it behind, whichever design is cheapest!). Personally I haven't had much reason to use a 2-stage lander even on Duna, but I hear with RSS apollo-style landers are pretty good

Right, that brings me to some clarifications I need to make in the OP. Thanks for making me see that.

Making them now.

[foamyesque]

On Laythe, a two-stage ascent vehicle makes sense, though it's not anywhere near required. Duna is closer to an edge case, I think. On the Mun, a two-stage lander can also make some sense, particularly if you're not doing an OR and are headed straight back to Kerbin, but you've got to be running tiny, tiny engines on the ascent stage; the Rockomax 48s or smaller. I think it'd be worth the trouble on Eelo, Vall and Moho, as they're both airless and larger and hence have much larger land&return deltaV requirements than Duna or the Mun.

[Anquietas314]

Right, that brings me to some clarifications I need to make in the OP. Thanks for making me see that.

Making them now.

Hmm. I think that makes it much harder to give you a solid answer. I suppose Moho or Vall/Laythe might be worth using a 2-stage lander due to the gravity and so on, but personally I'd just use a 1-stage anyway - I hate to leave junk lying around unnecessarily

Edited December 29, 2014 by armagheddonsgw

[GoSlash27]

This is actually part of the question I'm researching right now.

Neglecting cost, it comes down to mass.

I don't have any definitive answers just yet, but I should within the next couple weeks.

Best,

-Slashy

[Voidryder]

Agree with GoSlash (or do you prefer 77?), it's a question of mass and dv. (Ok, I added the dV.) Is the mass of an ascent stage, engine, fuel, etc and the dV required to get it off the surface less than the dV necessary to haul a nearly empty descent or non-reusable single stage back into orbit? If you aren't going to reuse the lander again, shed the weight in my opinion.

Plus the original landing site can become a tourist attraction if the descent stage is still there.

[GoSlash27]

Agree with GoSlash (or do you prefer 77?), it's a question of mass and dv. (Ok, I added the dV.) Is the mass of an ascent stage, engine, fuel, etc and the dV required to get it off the surface less than the dV necessary to haul a nearly empty descent or non-reusable single stage back into orbit? If you aren't going to reuse the lander again, shed the weight in my opinion.

Plus the original landing site can become a tourist attraction if the descent stage is still there.

My suspicion is that in all cases a 2 stager is preferable, but I haven't confirmed that yet.

It has a lot to do with how you approach the descent stage.

If you're using a zero descent rate profile, then clearly you're interested in absolute minimum mass for all phases of the mission. There may be some cases where ditching the empty tanks is cheaper than a decoupler, but I don't know yet.

If, OTOH, you're running with a suicide burn or reverse gravity turn, then you're willing to expend more mass in the descent stage. And in that case, why would you *not* dump that extra weight for the trip home?

This may not be a question that can be answered mathematically; may just come down to player preference...

Best,

-Slashy

[Greep]

This could seriously depend on what you're trying to do... especially now that there's base construction contracts, dropping a base and returning would pretty much always necessitate more than one stage.

Otherwise, it also does depend on funds vs mass a lot... e.g., if it were decided on mass alone, for a probe you might want to have a lower rockomax stage, with an upper monopropellant/ 0-10 stage. But that makes no sense whatsoever if you cared about funds...

[Jason Patterson]

As others have said, this depends heavily on your intent and the location, but I have to think that for most situations shedding mass will be superior. That may not mean full stages, perhaps only drop tanks... In lots of cases this is probably going to be a marginal savings in either direction - there are a bunch of bodies that are trivial in terms of delta-v to land and return to orbit.

[The Yellow Dart]

To me, in KSP, the second stage of a lander is essentially a backup plan. When sucking science out of the Mun and Minmus, I generally try to design landers that can land, hop around 1 to 3 times, and return to orbit to be refueled and reused. Sometimes I get too ambitious and make one hop too many; that's when I use a second stage, to ensure I can always return to orbit even if the first stage is out of fuel or damaged. Then I just dispose of the old lander and launch a new one.

[Laie]

Depends on the payload. After all the bodies excluded in the OP, Moho has the next-highest gravity. If all you want to take down to the surface is a lander can + three legs, the minimal lander is so small that I doubt whether it can reasonably be split into two stages. In terms of raw cost (only the lander itself), I'm pretty sure it can't be worthwhile. Careful optimization using the more expensive radial and oscar-b tanks may save on the order of 300kg... I'm not convinced that a two-stage lander will be cheaper even if you continue the calculation all the way back to the launchpad.

However, if you also want to land some extra weight, like a science package, two stages should be obviously superior for all bodies that require chemical propulsion. If ions are feasible anywhere, staging will be unnecessary / pointless.

[Foxster]

I've just completed a challenge to do the full science sweep at each biome on Eeloo in a single mission. My lander ended up as a multi-stage ship, not just two stages. This ensured it had plenty of dV left for the return journey.

While exploring it lost 8 used goo and science juniors, then a large empty fuel tank, then a pair of nuke engines and tanks, then a final pair of tanks and landing struts were discarded for the return leg. The rest of the engines were discarded before final Kerbin re-entry. Making it at least a five stage lander.

On Eeloo arrival:

On return to Kerbin:

and by splashdown:

[OhioBob]

On any planet with an atmosphere, then a one-stage lander is better. You'd be landing with parachutes, so very little of the lander's mass is dedicated strictly to the landing. The only mass that is really disposable is the landing gear, parachutes, and perhaps some science experiments. Since most of what you're landing is needed for the ascent, then two-stages is rather pointless. You might consider using decouplers to jettison the landing gear, etc., but you'd have to do the math to see if the gain is worth carrying the mass of the decouplers.

It is on the airless bodies that two-stages might make some sense. Although I've landed on most planets and moons, the only ones I've taken off from are Mun, Minmus and Duna, therefore I don't have a lot of experience. Of course Duna has an atmosphere, so I've already commented on that. For Mun and Minmus, I don't think two-stages is worth the trouble. Those bodies don't require all that much delta-v, so both landing and ascent can be accomplished with a pretty low mass ratio. Low mass ratios can easily be obtained with one stage, so why add the complexity of two? Even at Moho you can probably land and take off again with a mass ratio of about 2, which is easily obtainable with a single stage. If you can find a way to build two stages, each with a mass ratio of about 1.4, for less money and less mass than a single stage with a mass ratio of 2, then use two stages.

Edited December 29, 2014 by OhioBob

[MarvinKitFox]

Here is the final, definitive, absolute answer to your question.

Do you consider it fun to do so?

If your criteria is mass, mass, mass... then a 2-stage lander is a good idea. or 3 stage, 7-stage, 45-stage, whatever works out lightest in the end.

If your criteria is pure cost, after recovery from a KSP landing, then single-stage everything-returns-to-hangar might work out best.

If your criteria is effort, then likely a 1-stage lander moar boosters is the way to go.

For me, the criteria is all of the above, subject to maximizing the fun quotient.

So with me, the answer to the question "1 or 2 stage lander" has to be... "green with purple polka dots"

[Red Iron Crown]

I single stage everywhere but Kerbin, Tylo, and Eve. Though I generally don't discard landers after reaching orbit, I also never recover them. If I were trying to be as mass efficient as possible with a single-use lander staging might make more sense.

Pro-tip: Make your landers single stage, reusable and dockable (where possible), and leave them in orbit around the body you've just landed on. Next mission to that body instead of pushing another lander to the destination you just need to bring fuel for the lander.

Edited December 29, 2014 by Red Iron Crown

[Janos1986]

If there's an atmosphere to aerobrake and slow your descent with chutes only, and/or you feel confident about your TWR for going back up in orbit, use a single piece lander.

If there is no atmosphere and you feel the need to have a braking engine, use a two-staged lander to shed the unnecessary mass. It's useful even for the Mun, especially if you have tons of scientific equipment. You can visit different biomes until you run out of juice in your first stage, take all the science from the science modules and ditch the empty tank and modules to get back safely to Kerbin.

Or you can settle halfway and build a two-staged lander powerful enough to get to orbit in one piece, and having the second stage as an emergency bail for getting into orbit anyway should something go awry.

Edited December 29, 2014 by Janos1986

[Jouni]

Assuming that you want to maximize the payload fraction of the lander, there are three things that affect the optimal lander design: the kind of staging you want to use, the wet/dry mass ratio of each stage, and engine Isp.

If you use vertical staging (each stage has its own engines, and only one stage is active at a time), the optimal delta-v from from a stage using conventional rocket engines is something like 2000-3000 m/s. It's 2000-2500 m/s for Kerbin ascents, but lander stages usually have lower TWRs (and hence better mass ratios) and higher specific impulses than launch vehicles, so you may want to get more delta-v from each stage.

[Anquietas314]

If you use vertical staging (each stage has its own engines, and only one stage is active at a time), the optimal delta-v from from a stage using conventional rocket engines is something like 2000-3000 m/s.

In that case you should probably always use a single-stage lander; I want to say Duna's the biggest body of those considered that you would build a lander for (disregarding Kerbin and Laythe since both have oxygen atmospheres and hence you can/should use a spaceplane), and its escape velocity is 1372m/s. You're going to lose some deltaV due to the atmosphere, but not much.

[Kegereneku]

Landing on Mun and Duna I would typically use droptank attached to the landing gear, using the same engine(s). Any fuel surplus on the Mun allow me to reach more biomes and the return stage have its own surplus.

So I'm should be always winning as long as I use all fuel in the droptank.

[numerobis]

Whether to use a 2-stage lander depends on three things:

1. Delta-V and Isp from home to your landing site.

2. Mass of your decoupler, plus mass of your return engine if you aren't using the same one.

3. Delta-V and Isp from landing site to home.

Let's discuss a rocket that achieves 390s Isp and reaches its destination after a 6 km/s burn (roughly, KSC to Mun). If you use the smallest decoupler (15kg), then the fuel used to get the decoupler to the Mun is 57kg, straight from the ideal rocket equation. It's actually worse, because you need tanks to carry the fuel (and fuel to carry those tanks, ad infinitum), but how much worse depends on your staging setup, so let's just leave it at the best case.

What you need is that the return burn saves 57kg in fuel requirement. Assuming 900 m/s at 390 Isp (roughly, Mun to KSC), you need to leave 200kg beyond the decoupler itself.

That means that if you leave behind four landing legs and you're flying on an LV-909, you've broken even; leave a spent FL-T200 behind as well and you win. Do the math with the 48-7S and you need to leave behind a bit more.

But, do the math with an LV-909 on the way there and a 48-7S on the way back, and it's more demanding: you've got to carry that extra engine, which means 115kg in decoupling mass. So you need to dump about 1.5t on the lunar surface to make that worthwhile.

I just added a relevant function to the KSP-scripts, which lets you write:

import engine
# burn 6km/s at 390s Isp (aerospike), carry 115kg (48-7S + 15kg decoupler), how much fuel will we burn (ignoring tanks)
engine.propellantMass(6e3, 390, 0.115)
0.4359221842787079

# Return on the 48-7S, how much payload could we carry if we still had that fuel?
# That's the amount of payload we need to leave behind on Mun.
engine.payloadAllowance(900, 350, 0.436)
1.4565388115813005

# Subtract 15kg from that since we leave the decoupler behind, we need to find another
# this many tonnes to dump.
engine.payloadAllowance(900, 350, 0.436) - 0.015
1.4415388115813006

[Red Iron Crown]

^^ Also depends on whether you're optimizing for mass or cost. Optimizing for cost is a whole different calculation.

[numerobis]

Oh, yeah. Decouplers are expensive, fuel is cheap, plus if you go home with the lander bits, you get to recover it. So that's going to generally tilt pretty heavily in favor of single-stage.

Another thing I forgot though: I wrote the mass to go home, but left out the mass to land safely. If you're deciding between leaving a tonne of equipment behind or bringing it with you, you've got to take account of how many parachutes you'll need upon arrival -- and how much fuel it takes to carry that parachute.

[James M]

From my experience, two stage landers give you them best delta V return in lower gravity so you can make a bunch of hops from biome to biome. Honestly, I don't see why you wouldn't use a two stage lander when given the option.

[coredumpster]

An additional consideration, is that a two stage lander left behind with a probe core, solar, thermometer, and an antenna allows you to easily fulfill the 'Science from the surface of ....' contracts.

For me, in early career I do one-stage landers that return to Kerbin ( mun / minmus ).  Then I do two-stagers on bodies with higher gravity and to leave behind a science-probe.  After I advance further in tech tree and available funds, I start building stations in orbit and then leave my re-usable landers docked and use dedicated kerbin return ships.

With 1.2 now, how CommNet will factor into that strategy remains to be seen.  Might start doing two stagers, where lift-off from surface with both, and leave first stage in orbit with relays, and then smaller Kerbin return craft carrying the crew and science.