Why is 'Apollo style' so inefficient in KSP?

[eddiew]

Context:

A contract needs me to land at least two kerbals (a pilot, and a guest scientist, which I think is from the Field Research contract pack) on Mun. My tech level is quite low. I don't have the smallest engines, or anything bigger than a skipper.

Problem:

I was trying to put a mission together today in a classic Apollo stype; 2-man lander, 3-man command module with enough delta-v to get it and the lander to Mun, bring just itself back, and enough ascent stages to get the whole kerbang into LKO. It ended up weighing and measuring more than a tier 2 VAB and launchpad can handle - even after I grabbed the ALCOR lander capsule mod.

Which ok, I'll accept, Saturn V was massive and the crowning glory of rocketry; but we all know that a tier 2 setup is plenty to get 3 kerbals to Mun and back, with an all-in-one command module and lander and return vehicle, or potentially a single-stage spaceplane. I'll redesign in a more conventional way to complete the contract, but it made me wonder...

Question: why are micro-stages like Apollo used, less efficient than brute force, in KSP?

Surface gravity of Kerbin and Mun are roughly terrestrial and lunar equivalents, so my guess is it's the delta-v required to get from one to the other is far lower. Return from Mun is a sniff at 280m/s, while the descent and ascent are around 6-800 each way. Making a lander that weighs less than a small fuel can is... impossible without an ion engine and an external command chair (if then).

Any thoughts? Has anyone actually made a dual lander/return vehicle setup that was sensible? Is there an optimal flight plan? Are KSP's engines too heavy relative to the thrust/fuel use? Is it that Apollo's command module was a lot bigger than the lander and should be on 3.75m tech? Or is it really just a thing that you can do as a curiosity when you have enough tech and funds, but has now place in an efficient career?

I'm not really expecting a solution, but I am curious as to wherein the inefficiency lies. Thanks for answers

Edited July 6, 2015 by eddiew

[MarvinKitFox]

Actually, you have answered each of the objections yourself.

In KSP, the delta-v needed for a Mun trip does not justify doing a LOR moon mission. It almost doesn't justify using separate Mun lander/ascent stages!

KSP engine TWR are *abysmal* compared to Earth analogues.

KSP tankage mass ratios are even worse!

Using an Apollo-style lander will make a lot more sense when you need to go further, do more.

For example, a manned Tylo mission virtually demands that you use a separate orbiter and lander, and that your lander/return vehicle be multi-stage.

[GoSlash27]

Actually, you have answered each of the objections yourself.

In KSP, the delta-v needed for a Mun trip does not justify doing a LOR moon mission. It almost doesn't justify using separate Mun lander/ascent stages!

KSP engine TWR are *abysmal* compared to Earth analogues.

KSP tankage mass ratios are even worse!

Using an Apollo-style lander will make a lot more sense when you need to go further, do more.

For example, a manned Tylo mission virtually demands that you use a separate orbiter and lander, and that your lander/return vehicle be multi-stage.

All this, although an Apollo- style mission really shouldn't work out anywhere near 120 tonnes in KSP.

Best,

-Slashy

[Chaos_Klaus]

While it is not more efficient, it is still completely doable. My old Apollo 11 recreation launched on a 3.5m launcher.

You just have to be extremely careful about engine choice and how much fuel you bring. On the mun, you can use a single Spark and a few round-8 fuel tanks to launch the MK2 lander can back into orbit. The desent stage could use a Terrier with, say two T200 fuel tanks. Or you could do the whole lander in a single stage.

[Xannari Ferrows]

Well, brute force often entails high TWR, which means you spend less time fighting gravity. Smaller stages on rockets I typically use, contradictory to your point, actually end up working better [probably because I love going to Jool]. I try to stay away from the brute force solutions, and try to be as optimal and conservative as possible.

[rofltehcat]

I think in theory it can be boiled down to these options:

1. Have a transfer stage with a fuel-efficient engine, leave it in orbit, land with very light engines, come back up, dock, use transfer to go back or siphon the fuel then go back with the light lander engines.

vs.

2. Just have a lander with a heavier engine than you require, drop empty fuel tanks along the way and on the surface.

vs.

3. Pretty much exactly like 2 but have a lander with light engines, spend a little more fuel on the transfer but a bit less during landing/ascent.

So I think the answer you are looking for is in engine fuel efficiency and engine weight. Basically option 1 is only better if the transfer dV is sizeable amount of the mission or if your mission is massive (engine weight negligible, need higher TWR anyways).

A heavy but efficient engine would be Terrier/Poodle/NRV and a light but slightly less efficient engine would be Ant/Spark. So the fuel efficiency difference would be ~10-15%. if your lander's weight and fuel requirement fits for a Terrier or Poodle there would be even less of a reason to do the whole docking thing. In general I think most people probably just use these engines for vacuum landers (=option 2), thus foregoing all the hassle and their efficiency offsets a lot the engine weight.

Are there fuel efficiency stats for the different stages used on the moon missions? I'd expect the difference to be much larger than 10-15%.

[cantab]

I think the low delta-V requirements are the main cause. The Mun departure burn from Munar orbit is so tiny it's not much extra fuel to take down the the surface, and the small size of the Mun means taking it down isn't hard. The mass of a separate orbiter - duplicate Kerbal accommodation and duplicate engines in particular - will end up worse overall. The three-Kerbal pod being a complete boat anchor doesn't help things either.

If on the other hand you're going to Dres, with similar landing delta-V but requiring a much bigger return departure burns, "Destination Orbit Rendezvous" as I call it becomes much more appealing.

But even then, better still might be to leave a basic drop tank in the destination orbit, and pick it up again after the landing.

[Coga19000]

Well, what the others before me said is completely correct. But there are also many other reasons I did not see being mentioned , which are more or less these :

1) Engineering constraints- literally, the engines were a mess to work with. Most engines were unable to be throttled at almost any degree, and could only be fired once (The S-IV actually was designed to fire twice, but even that feat was stretching it) : both problems are still a major problem for the space industry. Therefore,each part of both the Saturn V and its payload was designed to meet certain Propulsion needs at specific times.

The Decent Propulsion System (the LM's engine) managed to circumvent both problems at a certain degree, but with ways too experimental to be fully trusted, and the solution posed its own problems; the engine was pressure fed with helium gas, whose pressure would build up through the trip, until the system vented it into space- saving the engine from blowing up, but rendering it inoperable. This usually happened a while after the lander was safely on the Moon, which would not be a concern only if the Engine was not expected to fire again after that point.

2. Celestial body diameter. Both the Moon and Earth are HUGE compared to the Mun and Kerbin respectively. While they have the same respective gravitational forces, the Delta -V requirements to circularisation are give or take four times higher. In those ranges, and along with the distance between Earth and Moon, that's when the "every gram counts" rule really kicks in.

3) Redundancy. Oh yeah, that's the big one. See, manned spaceflight is nowhere as simple and reliable as stock KSP has lead us to believe -especially since NASA didn't have other space companies to ask which way they did it. Systems may break up, engines may fail, tanks may leak, guidance may go haywire, connection may be lost, accidents may happen -and let's face i: I don't care how bonded you are to your little green big headed (literally) suicidal maniacs, a humans life is worth much more than the life of a digital character, and much harder to preserve in space. And when something as grand as a round trip to the Moon is done for the first time in the history of forever, the risks are huge. Therefore, you have to provide your pioneers with only one way to complete the mission, but dozens of ways to walk on the Earth's crust again.

Here are some major examples of necessary redundancy on the Lunar Orbital Rendevous Profile, as demonstrated throughout the Apollo Program itself :

i) As demonstrated on Apollo 10, the distinction between the descent and ascent stage provided the astronauts with a way to Abort a Moon landing while it's done; if the descent engine failed for whatever reason, the ascent stage could be separated and ignited while still on a suborbital trajectory (which would obviously not be enough after a certain amount of Velocity was shed) and circularise again, with the useless descent stage fully embracing the regolith in all its glory.

ii) Obviously, Apollo 13 is the greatest example of why redundancy was necessary. The Lunar Module served both as a lifeboat, providing communication and life support, and as a Propulsion System to head the whole Assembly back at home.

iii) During the Apollo 15 landing, the DPS collided with the Lunar soil, which obviously is not as easy to get away with IRL as it is in KSP. The engine bell was designed to crumple harmlessly in such a turn of events, but an engine without a nozzle is extremely inefficient and totally gimbal-free, which would be a problem if the Engine had to fire again.

[pincushionman]

Two words: Dry. Mass. Leaving aside the engine engineering problems mentioned by Coga19000 above, all hardware in KSP is ludicrously heavy compared to the nearest real-world counterpart. In Apollo, the fuel mass that needed for a direct mission was prohibitive; in KSP the hardware mass needed for a rondezvous mission is.

[eddiew]

Some great answers here guys, thank you

Suddenly I realise why NASA don't scour these forums for good ideas - our ruleset is just too different vs reality to produce anything viable ^^;

[FancyMouse]

Are these relevant discussions?

http://forum.kerbalspaceprogram.com/threads/79775-multi-stage-vs-single-stage

http://forum.kerbalspaceprogram.com/threads/105236-When-to-and-when-not-to-use-two-stage-landers

[eddiew]

All this, although an Apollo- style mission really shouldn't work out anywhere near 120 tonnes in KSP.

I think my problem was a 2-man lander requirement. I'm sure I could have cut it back if only one of the crew went down. The absence of mainsails might also be a factor, since I was having to load five skippers onto the base of it, which pushed up the fuel mass, which... etc.

But even then, better still might be to leave a basic drop tank in the destination orbit, and pick it up again after the landing.

I actually did exactly that with Duna landings in 0.90 with TAC-LS. Something like 10 tons of food and fuel left in orbit shrank the lander an awful lot

Well, what the others before me said is completely correct. But there are also many other reasons I did not see being mentioned , which are more or less these :

1) Engineering constraints- literally, the engines were a mess to work with.

All very good points, but the engines in particular strikes a chord. Inability to re-light, and invariable thrust means pretty much everything is going to fly like an SRB, and means that over-provision of fuel (so easy in KSP) is actively harmful.

- - - Updated - - -

Are these relevant discussions?

http://forum.kerbalspaceprogram.com/threads/79775-multi-stage-vs-single-stage

http://forum.kerbalspaceprogram.com/threads/105236-When-to-and-when-not-to-use-two-stage-landers

Yes, actually. And if I'm taking a 1.25m can to Mun, I'll use droptanks tanks which also hold the landing legs and materials bays, and get ejected half way through the ascent Usually doesn't seem necessary for larger landers however. It's just easier to make an all-in-one and recover a bit more of the costs upon return. One of the problems here is that KSP fuel cans are in such discrete sizes, and as pincushionman said above, the dry mass of the tank is very big compared to its fuel. Fine tuning your munar descent stage to, say, 650 delta-v, is easy, but the relative change is in mass when you drain a bit out isn't as big as it should be.

Edited July 6, 2015 by eddiew

[Wemb]

Some great answers here guys, thank you

Suddenly I realise why NASA don't scour these forums for good ideas - our ruleset is just too different vs reality to produce anything viable ^^;

It is, in fact, rocket science.

Wemb

[eddiew]

It is, in fact, rocket science.

Wemb

Well, KSP is not, NASA is

[davidpsummers]

I did my moon missions this way. The only difference is that I put my lander and orbiter in Kerbal orbit with two different launches. I also left the lander in orbit and reused it on subsequent missions.

Not sure why NASA didn't do it this way. Why did they build the huge Saturn V when the already showed they could rendezvous in orbit?

[Sampa]

I did my moon missions this way. The only difference is that I put my lander and orbiter in Kerbal orbit with two different launches. I also left the lander in orbit and reused it on subsequent missions.

Not sure why NASA didn't do it this way. Why did they build the huge Saturn V when the already showed they could rendezvous in orbit?

The reason for this was mostly because we had no way to efficiently get fuel to the lander to refuel the said lander. Also, by the shear design of the lander...efficient for its time, as far as weight is concerned.

[eddiew]

The reason for this was mostly because we had no way to efficiently get fuel to the lander to refuel the said lander. Also, by the shear design of the lander...efficient for its time, as far as weight is concerned.

Tbh, I'm not entirely sure I have fuel pumping ability yet myself ^^; Other than this, sending a lander on a fuel tank (to top it up with again) does sound like an interesting approach. Assuming you have a scientist to reset your experiments

[Sampa]

Tbh, I'm not entirely sure I have fuel pumping ability yet myself ^^; Other than this, sending a lander on a fuel tank (to top it up with again) does sound like an interesting approach. Assuming you have a scientist to reset your experiments

Yeah, although the CSM and the LEM were equipped with an umbilical, it was only intended for use for transfer of electricity from the Command Module to the LEM during the LEM full power up procedures and, I'm guessing, just after the LEM ascent module returned from the surface and they began transferring surface samples, moon rocks, equipment, and whatever the astronauts decided they may want to keep as a souvenir from their venture to the surface. This umbilical WAS used in reverse during the famous Apollo 13 mission to provide the extra power necessary to top of the CM's batteries prior to LEM release and subsequent reentry. However, there was NO way to repurpose the umbilical for fuel due to the fact that the LEM fuel tanks contained a different fuel type than the CSM did.

[ajburges]

I did my moon missions this way. The only difference is that I put my lander and orbiter in Kerbal orbit with two different launches. I also left the lander in orbit and reused it on subsequent missions.

Not sure why NASA didn't do it this way. Why did they build the huge Saturn V when the already showed they could rendezvous in orbit?

Saturn V was already in development. You should see the concept for the direct ascent mission plan!

Earth orbit rendezvous was considered to add too much risk vs the cost of a single launch. Less overhead for a single launch as well. Remember, we were still trying to prove orbital rendezvous could be done when they were selecting the Apollo mission profile.

[NathanKell]

As for why we didn't leave the LEM in orbit around the moon....we did. Thing is, though, the moon is very harsh on orbits (look up 'masscons'), and it gained a negative periselene quite quickly. To stay in low lunar orbit you need quite an insane quantity of stationkeeping delta V.

[davidpsummers]

Well, the LEM left a stage on the surface, so I don't see how they could have reused it.

The main questions was about the Saturn V vs. launching the two modules separately and meeting up in orbit. My cynical side wondered if there wasn't lobbying involved. But then I'm picturing a smaller rocket as simpler and easier and if they already had done enough development on the Saturn to have more confidence in it, that makes sense.

[ringerc]

One of the issues I run into in KSP when doing Apollo-style missions is the sheer mass of the lander modules. The size-1 module is 80% the mass of a full command module. The size-2 module, 2-man, is over half the mass of a robust, re-entry capable 3-man command module. There's no life-support to worry about so the short livable duration of the lander module vs the command module doesn't come into play.

IMO that's one of the reasons it's often cheaper to just land the lot. Especially since you can then save on RCS mass, etc.

Then again, I'm a semi-routine user of ablative lithobraking as a landing technique, so I might just be crazy. Burn out just above the surface and use the empty stage as a crumple-zone.

Edited July 8, 2015 by ringerc

[FlowerChild]

The Mun is also a lot smaller/less massive than the moon, so you don't have as steep a gravity well to climb out of, and the savings you get out of a CM/LM style design are thus less pronounced.

[tater]

It's because the kerbol system is tiny. Rockets are less powerful, but not 11x less powerful, so the mun is trivial. It's actually bad gameplay, because there is not really a choice to consider. Realism is clearly better gameplay in this case because choices---that all could work---are better than it being trivial.

[kurja]

In career mode I find it extremely useful to use separate descent and ascent stages, on the Mun and elsewhere - descent stage contains all the science bits, solar panels / rtg and an antenna plus a probe core, so in the future when there are "return or transmit scientific data from x" I can just transmit from the abandoned descent stage, that's basically free money.

edit: and ofc putting a probe core etc on the ascent stage as well makes for an (almost) free satellite in case of future "scientific data from orbit of x" contracts, useful if there was no satellite in place there already.

Edited July 8, 2015 by kurja