Efficient Rocket Landings

[Orian Gertdithe]

I have been playing KSP for over a year now and I honestly have no idea how to land efficiently on a moon (to be more specific the Mun and Minmus). I've researched how to do this but I can't find what I need or I do find what I need but I can't comprehend it.

Please give me some advice.

-Orian

[FullMetalMachinist]

9 hours ago, Orian Gertdithe said:

I have been playing KSP for over a year now and I honestly have no idea how to land efficiently on a moon (to be more specific the Mun and Minmus). I've researched how to do this but I can't find what I need or I do find what I need but I can't comprehend it.

Please give me some advice.

-Orian

Welcome to the forums! 

@GoSlash27 has a great tutorial on performing a reverse gravity turn.

Not only does he give a great procedure for that, but also does a quick overview of the other two most common landing techniques. I suggest you read that, and then ask if you have any further questions. 

Edited March 31, 2016 by FullMetalMachinist

[p1t1o]

On airless moons, there is no reason to use a "reverse gravity turn".

The shorter your burns, the less dV you waste fighting gravity.

On airless moons, it is entirely legitimate to simply zero your horizontal speed on orbit above your chosen landing area. Do this with as hard a burn as you can, a tool to view your horizontal and vertical velocity components is very useful here.

Now you should be falling vertically towards the spot on the moon where you will be landing.

Zeroing your vertical velocity in one sharp burn at an appropriate time - sometimes called a "suicide burn".

These short, sharp burns are quite efficient, and though the seperate treatment of your horizontal and vertical speeds might seem quite clunky and counter intuitive, it can approach maximum theoretical efficiency by minimising gravity losses. The vertical and horizontal components of your velocity must be zeroed in order to land, there is no curve or thrust regime which will get around that, doing each component separately does not add any extra dV.

On bodies with an atmosphere, things are of course quite different

[Starhawk]

1 hour ago, p1t1o said:

On airless moons, there is no reason to use a "reverse gravity turn".

I'm afraid that you are mistaken.

The method you describe (sometimes called 'stop and drop') is not the most efficient way to land on an airless body.  So there definitely is a reason to use a different method.

The most efficient way to land is to lower your periapsis until it is just barely above the surface near your landing site.  Then execute a suicide burn very close to the periapsis which is very close to the surface.  In this way you zero both your horizontal and vertical velocity in one burn making maximum use of Pythagoras.

Separating the horizontal and vertical components is less efficient in exactly the same way that you have to travel farther walking around two sides of a field rather than walking diagonally from corner to corner.

As to Slashy's 'Reverse Gravity Turn' method.  That is designed specifically for (and highly recommended for) doing pinpoint landings on an airless body.  For example, when you want to land right at your Munbase.

Happy landings!

[Foxster]

...or you could use Mechjeb's landing assist. You could either just use it as a learning aid or, if you like how it takes the tedium out of multiple landings, then carry on using it. 

Edited March 31, 2016 by Foxster

[Wemb]

1 hour ago, Foxster said:

...or you could use Mechjeb's landing assist. You could either just use it as a learning aid or, if you like how it takes the tedium out of multiple landings, then carry on using it. 

And if you want something a but less automated, take a look at KER - it has what is effectively, a landing computer which won't do the landing for you, but will calculate times to impact, suicide burn times, etc. as well as displaying really, -really- essential data, such as the radar altimeter rather than ASL.

Wemb

[N_Danger]

 MecJeb also has a Surface Info screen which gives true altitude above the ground plus a read out of current horizontal and vertical velocities that is necessary in manual landings.

[Wemb]

And, if you want to go entirely stock, one of the most useful things you can do is to slap a couple of downward facing spot lights on your ship - if you've every seen the movie The Dambusters you'll know why.this is useful.

Also, if you want practise - do it on it on Minmus, It's much easier and forgiving to land on and required a lot less dV meaning you can practise more.

Wemb

Edited March 31, 2016 by Wemb

[FullMetalMachinist]

6 minutes ago, Wemb said:

And, if you want to go entirely stock, one of the most useful things you can do is to slap a couple of downward facing spot lights on your ship - if you've every seen the movie The Dambusters you'll know why.this is useful.

Wemb

Another all-stock option is to place a maneuver node right at the spot that your orbit line meets the surface. Then pull the retrograde handle until the orbit is gone. The "est. node burn time" is how long your suicide burn will take. Then hit the gas when the "time until node" equals the "node burn time".

(Actually start slightly after that, because the stock node burn time doesn't account for your TWR increasing as you burn fuel.) 

[Vegetal]

After roughly a thousand hours playing ksp, executing nodes and landing on moons got old, so I installed mechjeb, which does a fine job on that.

Before this, I used a pretty standard landing method, trying to follow what I think it's the theoretical ideal. I might be wrong, but the ideal is to make a single, full power retro burn that zeroes your velocity right at the ground level. That would be extremely hard to nail in practice.

What I do, is a retro burn in orbit to bring my trajectory impacting somewhat after my target LZ. After that I keep locked retrograde, and adjust my landing trajectory just using the throttle. That way I don't have any cosine losses, and can get pretty close to the theoretical ideal. If you have a strict target, like a surface base, you might have to burn sideways to retrograde to adjust your trajectory.

Remember, the shallower your approach is, the more efficient it gets, but also more dangerous.

[Madscientist16180]

11 hours ago, Starhawk said:

I'm afraid that you are mistaken.

The method you describe (sometimes called 'stop and drop') is not the most efficient way to land on an airless body.  So there definitely is a reason to use a different method.

The most efficient way to land is to lower your periapsis until it is just barely above the surface near your landing site.  Then execute a suicide burn very close to the periapsis which is very close to the surface.  In this way you zero both your horizontal and vertical velocity in one burn making maximum use of Pythagoras.

That sounds pretty fun and dangerous! Gotta try it, i always "stop and drop" as it is easy and allow for precise enough landing for me. 

Any Idea by what kind of margin the way you describe is more efficient? I mean if we are talking a few percent D-V, the ease of "stop and drop" could win the debate for what i'll use in the futur

[Starhawk]

2 minutes ago, Madscientist16180 said:

That sounds pretty fun and dangerous! Gotta try it, i always "stop and drop" as it is easy and allow for precise enough landing for me. 

Any Idea by what kind of margin the way you describe is more efficient? I mean if we are talking a few percent D-V, the ease of "stop and drop" could win the debate for what i'll use in the futur

I don't have any hard numbers.  The higher the gravity, the more it will matter.

It probably doesn't help much at all at Minmus.  At the Mun on the other hand I'm sure the difference is quite noticeable.

Think of it like this.  Imagine you're orbiting the Mun at 20 km. If you lower your periapsis to just above the surface then, when you get there, you will be moving somewhat faster when you reach periapsis than you were in a 20 km circular orbit.  If you burn and zero your horizontal velocity at that point, you are only a small distance above the surface and only require a small burn to touch down.  The increased speed at your periapsis is much less than what you will spend fighting the velocity imparted by the 20 km fall.

If you zero your horizontal at 20 km and drop from there think about how much fuel you burn to zero out the velocity that gravity imparted on the way down.

The difference can be calculated without too much difficulty, but I'm somewhat lazy about math. 

Happy landings!

[Plusck]

37 minutes ago, Madscientist16180 said:

Any Idea by what kind of margin the way you describe is more efficient? I mean if we are talking a few percent D-V, the ease of "stop and drop" could win the debate for what i'll use in the futur

The simplest way to check is to first note your horizontal velocity (call this "b") when doing a "stop and drop", then just let yourself fall to the ground and note what speed you're going when you impact (call this "c"). Those two figures taken together give the abolute minimum expenditure on a "stop and drop" maneuvre.

Then you apply Pythagoras (since your horizontal and vertical vectors are at right angles, so a = sq.rt.(b²+c²)) to work out what the maximum-efficiency suicide burn would have been by burning along the hypotenuse.

However, if you are slowing yourself down on the way, you also need to add the time difference between going down slowly and just dropping straight to the surface. Every second longer that it takes, you must multiply by the gravity of that moon or planet (so on the Mun that would be 1.6 m/s per second). Let's call this "d".

By using the stop-and-drop method, you are wasting slightly less than (b+c+d)-a.

Of course, no suicide burn can quite equal that efficiency - but your savings by making a minimal burn to just graze the surface then doing a suicide burn will be close to this amount if you have a high TWR.

Admittedly, a further approximation is that this doesn't take account of the fact that the suicide burn can't be perfectly aligned along the hypotenuse of these vectors... but that approximation will only be a few percent  

 

So for example, if you are at 20km above the surface of the Mun, and you orbital velocity is 600 m/s (tbh I can't remember what it is at this altitude), you will take time t = sq.rt.(20,000*2/1.63) = 157 seconds to hit the ground, and will hit at 157*1.63 = 255 m/s. Minimum possible burns for a "stop and drop" will be 855 m/s. Minimum possible (approximated) suicide burn would be 652 m/s. So that's a 23% saving (or 31% loss, depending on how you want to look at it).

Edited March 31, 2016 by Plusck

[Orian Gertdithe]

Thanks for the advice guys! This has helped me a lot.

[Superfluous J]

6 hours ago, Madscientist16180 said:

Any Idea by what kind of margin the way you describe is more efficient? I mean if we are talking a few percent D-V, the ease of "stop and drop" could win the debate for what i'll use in the futur

Roughly speaking, when you come to near-stop 15km over the Mun's surface, note how much fuel and/or dV you have left. Then, when you're landed, note it again. Whatever you used falling down, that's what you wasted.

It's actually a *LITTLE* bit less than that, but not much, and for most moons and worlds it's significant.

[p1t1o]

On 31/03/2016 at 0:43 PM, Starhawk said:

I'm afraid that you are mistaken.

The method you describe (sometimes called 'stop and drop') is not the most efficient way to land on an airless body.  So there definitely is a reason to use a different method.

The most efficient way to land is to lower your periapsis until it is just barely above the surface near your landing site.  Then execute a suicide burn very close to the periapsis which is very close to the surface.  In this way you zero both your horizontal and vertical velocity in one burn making maximum use of Pythagoras.

Separating the horizontal and vertical components is less efficient in exactly the same way that you have to travel farther walking around two sides of a field rather than walking diagonally from corner to corner.

As to Slashy's 'Reverse Gravity Turn' method.  That is designed specifically for (and highly recommended for) doing pinpoint landings on an airless body.  For example, when you want to land right at your Munbase.

Happy landings!

Didn't mean to imply that it was always the most efficient route, I only said it was "perfectly legitimate", though I probably went a little overboard when I said "...it can approach maximum theoretical efficiency " - if I was going into more detail I would have said something about how its efficiency increases the lower your altitude, as "stop-and-drop" more closely approaches a single burn.

Thus, the most efficient pathway which you describe is actually following the same general idea, as it will most likely still involve a separate vertical burn of some size or other.

Sorry again if I was just muddying the waters

[Snark]

23 hours ago, FullMetalMachinist said:

Another all-stock option is to place a maneuver node right at the spot that your orbit line meets the surface. Then pull the retrograde handle until the orbit is gone. The "est. node burn time" is how long your suicide burn will take. Then hit the gas when the "time until node" equals the "node burn time".

+1 to this technique, it works pretty well!

23 hours ago, FullMetalMachinist said:

(Actually start slightly after that, because the stock node burn time doesn't account for your TWR increasing as you burn fuel.) 

It's true that you need to start slightly after that... and that increasing TWR can be one contributing factor... but the biggest reason is a mathematical one that has nothing to do with TWR.  Even if you have the "infinite fuel" cheat turned on so that your TWR stays perfectly constant, you'll still want to start somewhat after the point where "time to burn" equals "time to impact".

Reason:  The "time to impact" (or the maneuver node, if you will) is how long to get there if you don't do any burning beforehand.  It's essentially assuming that your burn will be instantaneous at the moment of the node.  Since you're starting your burn before that, you're slowing down, which means you will take longer to reach the node (or impact) than predicted.

Exactly how much longer will be a function of your local TWR and your angle of descent.  Higher TWR = shorter time.  Steeper angle = shorter time.  In the extreme optimum case (infinite TWR, going straight down), you'd start your burn when it's exactly half of the time remaining.  In practice, you'll need more than that-- I find that 65%-70% is a happy medium that works most of the time.  However, if you have a very low local TWR, you may actually need significantly longer than the estimated time-to-impact to come to a halt, so be warned.

Incidentally, while everyone's discussing various mod-related solutions to the problem (KER, or MechJeb, for example), may I also suggest BetterBurnTime:

It's a bit less intrusive than the MechJeb or KER experience, so somewhat closer to the stock experience.  What it does for vac landings is subtle, but useful:  when you're falling towards the surface of a vacuum world, it displays a "time until impact" display in the spot that would normally show "time to maneuver", and an "estimated burn time to kill velocity" display in the spot that would normally show "estimated burn time for maneuver".

So essentially it gives you an easy automatic equivalent of the maneuver-node trick that @FullMetalMachinist points out.  Also, it does take changing-TWR-as-you-burn-fuel into account. 

[Starhawk]

4 minutes ago, Snark said:

What it does for vac landings is subtle, but useful:  when you're falling towards the surface of a vacuum world, it displays a "time until impact" display in the spot that would normally show "time to maneuver", and an "estimated burn time to kill velocity" display in the spot that would normally show "estimated burn time for maneuver".

So essentially it gives you an easy automatic equivalent of the maneuver-node trick that @FullMetalMachinist points out.  Also, it does take changing-TWR-as-you-burn-fuel into account. 

I had no idea BBT had this feature.

Going to download now.

Happy landings!

[FullMetalMachinist]

12 minutes ago, Snark said:

Reason:  The "time to impact" (or the maneuver node, if you will) is how long to get there if you don't do any burning beforehand. 

*facepalm* 

Totally forgot about that part, which is dumb of me because, yes, that is the biggest contributer to starting your burn after the "time to node" time. 

[soundnfury]

If I might take the opportunity to plug one of my projects: konrad is very effective for targeted landings on airless bodies.  Like KER, it gives you useful data rather than flying the landing for you; unlike KER, that data is derived from extrapolating your motion based on your current attitude and throttle setting.  It does need a bit more setup (you have to tell it things like your dry mass, Isp and thrust, and give it a terrain map from SCANsat, because it's an external program running off telemetry rather than a mod inside KSP), but with its help it's possible to make a landing with pin-point precision, while using less delta-V than a reverse gravity turn (you perform the majority of the burn at a constant attitude, rather than gradually pivoting upright).

[Loren Pechtel]

On 3/31/2016 at 2:46 AM, p1t1o said:

On airless moons, there is no reason to use a "reverse gravity turn".

The shorter your burns, the less dV you waste fighting gravity.

On airless moons, it is entirely legitimate to simply zero your horizontal speed on orbit above your chosen landing area. Do this with as hard a burn as you can, a tool to view your horizontal and vertical velocity components is very useful here.

Now you should be falling vertically towards the spot on the moon where you will be landing.

Zeroing your vertical velocity in one sharp burn at an appropriate time - sometimes called a "suicide burn".

These short, sharp burns are quite efficient, and though the seperate treatment of your horizontal and vertical speeds might seem quite clunky and counter intuitive, it can approach maximum theoretical efficiency by minimising gravity losses. The vertical and horizontal components of your velocity must be zeroed in order to land, there is no curve or thrust regime which will get around that, doing each component separately does not add any extra dV.

On bodies with an atmosphere, things are of course quite different

Huh??  I don't see how you conclude that doing the burns separately doesn't cost you anything.

1)  Oberth.  You want to do as much of your burn as low as possible.

2)  Time.  Every second you're not in orbit you're expending fuel to counter gravity.

I have found I use the least fuel when I set my periapsis as low as I dare, then kill most of my orbital velocity at periapsis, aim retrograde and burn when the suicide timer (Mechjeb) is down to a few seconds.  If your rocket is not very maneuverable you need to be a bit higher up--I have a self-refueling science lab rocket that would certainly crash if I tried this approach with it.

[Blaarkies]

On 4/1/2016 at 2:08 PM, p1t1o said:

Thus, the most efficient pathway which you describe is actually following the same general idea, as it will most likely still involve a separate vertical burn of some size or other.

No it's not following the same idea and it should not involve a separate vertical burn, at all. A "Reverse Gravity Turn" is exactly like launching into orbit as efficiently as possible. The differences:
> At launch your TWR grows as your tanks empty.
< At landing, your TWR grows as altitude decreases(if you burn up fuel)

> At launch, you aim prograde and launch with a pitch as low as possible so that your AP is at minimum orbital altitude (gain horizontal velocity as quick as possible...without high-fiving mountains)
< At landing, you aim retrograde and de-orbit from as low an altitude as possible (lose horizontal velocity as quick as possible...without low-fiving mountains(?))

No stock engines on any craft will give you 600m/s dv in 0seconds, so you will burn vertical based upon how long it took to lose all the horizontal velocity...this is inevitable, but it works the same way in landing

In short, imagine you launch a perfect gravity turn where you one-burn the engines to orbit. Remember where you turned them off? That is where you start the suicide burn, ending with a landing where you come to a stand still 0m from the ground. If Minmus was only flats, you could get into a 10m high orbit, start burning retrograde and then just start pitching up to keep your variometer pointed at zero and this would be the most efficient landing for the craft's TWR(without 0.1m risk)

Notice that nowhere should you ever split burns into horizontal and vertical separate burns.
- 1.41 m/s dv in a 45'degree pitch gives
-> x = 1.41 * cos(45') = 1 m/s (horizontal)
-> y = 1.41 * sin(45') = 1 m/s (vertical)

To achieve that with two burns, you need 2 m/s dv total. Doesn't seem like a lot, but it can be multiplied to keep this ratio...1414m/s is considerably cheaper than 2000m/s. But don't fret, this is worst case scenario. If you aim within 25'deg of your intended aim, you lose less than 10% dv

Edited April 5, 2016 by Blaarkies

[p1t1o]

5 hours ago, Loren Pechtel said:

<snip>

 

2 minutes ago, Blaarkies said:

<snip>

Bit late to the party guys.

On 4/1/2016 at 1:08 PM, p1t1o said:

Didn't mean to imply that it was always the most efficient route, I only said it was "perfectly legitimate", though I probably went a little overboard when I said "...it can approach maximum theoretical efficiency " - if I was going into more detail I would have said something about how its efficiency increases the lower your altitude, as "stop-and-drop" more closely approaches a single burn.

Thus, the most efficient pathway which you describe is actually following the same general idea, as it will most likely still involve a separate vertical burn of some size or other.

Sorry again if I was just muddying the waters

 

[Wemb]

On 01/04/2016 at 4:37 PM, FullMetalMachinist said:

*facepalm* 

Totally forgot about that part, which is dumb of me because, yes, that is the biggest contributer to starting your burn after the "time to node" time. 

OTOH, regarding when the best time to start a suicide burn it's also important to remember that the consequence of starting your burn too early is comming to a halt some distance above the surface and some wasted fuel, while the consequence of starting your burn too late is becoming a new crater. People new to this might want to err on the side of caution. :-)

 

Wemb

Edited April 5, 2016 by Wemb