Reverse gravity turn landing technique for airless bodies

[GoSlash27]

I've been doing a lot of testing with landers this past week, and would like to share a landing technique that I've figured out in the process. Not sure if it's already commonly used or has a different name. If so, please let me know.

The two most common landing techniques are the "suicide burn" and "zero descent rate" approaches.

The suicide burn is a reverse vertical ascent-then-circularization burn. On the plus side, it can drop you in the neighborhood of where you want to be, but relies on perfect timing, high thrust to weight, and is hugely wasteful of fuel.

The zero descent rate approach is the most efficient approach and doesn't require a high thrust to weight. It's the reverse of an immediate prograde burn to orbit.

But on the downside, it's highly inaccurate AFA final touchdown point, and may smack you into terrain if you're not careful.

I call this the "reverse gravity turn approach" and it has advantages over both of these techniques.

It has medium efficiency, works with low t/w ratios, is very easy to execute, and will drop you right where you want to be with no fuss.

I will demonstrate what you need to use this technique, starting with the vehicle:

We are going for Tylo in this example, and this is our lander.

DV is twice the "perfect" DV for Tylo according to WAC's DV chart

In the case of Tylo, it says 2270 m/sec, so we'll double that to 4,540.

T/w should be at least 1.2 at the body you wish to land on. No harm in going a little higher to start with, but keeping a low t/w will make your touchdown less twitchy because you have finer throttle control.

Edited December 24, 2014 by GoSlash27

[GoSlash27]

Step 1: Setting up

So here we are in orbit around Tylo. First order of business is to make sure that we're ready to shoot an approach. Engines should be armed and throttled down. Nose is pointed retrograde and we are rolled upside- down according to the nav ball.

Next, we identify our landing site.

Now we create a maneuver node. It will be set up for a retroburn of roughly 1/2 our orbital speed. Then you drag it into position so that it drops into the landing area.

This maneuver node is your guide for when to begin your approach. Once you are over the front intersection with the surface, it is time to begin your retro burn. But before we do that...

We want to make sure our view is where we want it (looking forward at the top of the ship with the surface on the left) and look everything over one last time.

When we are over the leading edge of our maneuver node curve, it's time to begin our approach.

[GoSlash27]

Step 2: Deorbit burn

We are at the leading edge of the maneuver node, so it's time to begin.

Go to full throttle retrograde and burn until your flight path intersects the landing area, then throttle down to cutoff.

If you have a reference on the surface, you no longer need the maneuver node. It's okay to get rid of it at this point. You can keep the maneuver node up if you want it to provide you a target for landing.

We will free- fall until we are halfway between our apoapsis and the surface.

At this point we will begin our powered descent.

[GoSlash27]

Step 3: Powered descent

We are mid-way between apoapsis and touchdown, and it's time to begin our powered descent.

We will keep the nose a little above the retrograde marker and increase throttle so that we remain halfway between apoapsis and touchdown.

If we are low (closer to the surface than apoapsis), we increase throttle.

If we are high (closer to apoapsis than the surface), we decrease throttle.

If we are long (overshooting the landing site), we pitch toward the retrograde marker.

If we are short (falling short of the landing site), we pitch more vertical.

Once we can't clearly see where we're at, it's time to transition to the final visual approach.

[GoSlash27]

Step 4: Final approach

This is pretty much the same as all visual approaches. You just fly to correct your drift while aiming for your desired touchdown spot, and gradually reduce your sink rate as you approach the surface.

This particular approach placed me within 1.5 Km of my target point. I didn't overshoot or undershoot, but ended up right.

Normally I can parallel- park this approach precisely on the spot, but snapping pics like a tourist kinda ruined my finesse.

Still... a klik- and- a- half ain't nuthin' to sneeze at.

That's all there is to it. Time to go and science!

There's really nothing demanding about this technique, which is why I use it. I don't have to practice it or figure out any timing. So long as the ship is properly built for it, I don't have to worry about missing my target or hitting terrain.

This technique works the same way on all airless bodies.

Best,

-Slashy

Edited December 24, 2014 by GoSlash27

[Mystique]

Good guide, I was using kinda same approach after all those try-and-fails and found it to be most efficient (first thing to do was to break horizontal speed to 0 in orbit and then just control vertical speed, then I tried that "zero descent rate" and found it to be totally real-time inefficient, boring and not fitting into Kerbal space exploration conception so eventually I came to this).

The only thing I was not doing was this:

"If we are low (closer to the surface than apoapsis), we increase throttle.

If we are high (closer to apoapsis than the surface), we decrease throttle."

Didn't pay attention to map view, just guided ship visually to avoid hitting bigger rocks (I kinda love shallow landings, feels like a wingsuit basejumping ) while controlling speeds using MJ/KER info screens; I was checking map view only to see if I haven't deviated from landing coordinates too much.

And I must admit that it's quite precise - I managed to land a rescue fuel tank on Mun for rover-lander mission that ran out of fuel in just about 50 m. That was a great relief, because first one dropped in 5 km and I've chosen to launch 2nd rescue mission instead of spending bunch of time driving rover across those craters

[SmashBrown]

Yeah I use a similar method myself, although I do occasionally use suicide burns as well.

[RAINCRAFTER]

I used this method for my first successful Minimus landing a few minutes ago.

[Mystique]

Yeah I use a similar method myself, although I do occasionally use suicide burns as well.

For a long time I thought that suicide burn is thing like one described in OP, but only with retrograde burn timed precisely to get both vertical and horizontal speed to 0 right before touching ground. So I used MJ's suicide burn timer to do it, and, well, after some FUBAR landings I corrected method as written above

[Kesa]

For a long time I thought that suicide burn is thing like one described in OP, but only with retrograde burn timed precisely to get both vertical and horizontal speed to 0 right before touching ground. So I used MJ's suicide burn timer to do it, and, well, after some FUBAR landings I corrected method as written above

I was thinking the same until now

IRL, this is also called a gravity turn. An ideal gravity turn is made at full throttle and speed is killed exactly at touch down. But precise timing and high TWR/ full thrust matter less than for a suicide burn.

My thumb up rule is burn when impact time < speed / (2 acceleration)

(time to impact is given by KER)

It's essentially a rough generalisation of suicide burn (it is a suicide burn if you fall straight and substract gravity to acceleration).

It's pretty fuel efficient, but not that accurate.

I'll try your method Slashy

[rkman]

The suicide burn is a reverse vertical ascent-then-circularization burn. On the plus side, it can drop you in the neighborhood of where you want to be

I'm not sure that is a proper definition of "suicide burn". Suicide burn generally means waiting until the last moment to decrease speed to a safe landing speed. It does not need to be vertical.

Starting from orbit with the intention to 'wait until the last moment' for the final burn, that means a valid option is to spend only a small amount of dV on maneuvers other than the suicide burn. It might not be the only way to do a suicide burn, but it isn't not a suicide burn and in many ways it is very different than "reverse vertical ascent-then-circularization".

If only a small amount of dV is spend on the initial 'reverse circularization' (deorbit burn), the angle to the ground is shallow so it's near impossible to precisely land where you want. But it does make for a rather efficient suicide burn landing.

but relies on perfect timing, high thrust to weight, and is hugely wasteful of fuel.

Landing on Tylo with a vessel twr of 2.6 (@Tylo), and 77m/s initial deorbit burn (from 60km orbit), the suicide burn lasts about a minute. Total dV spend: 2606m/s.

If i see it correctly, during the landing technique you described throttle is at max for most of the final burn, reduced only during the last few hundred meters. That is in effect a suicide burn, it takes a long time because vessel twr is low.

I think that aside from the initial deorbit burn, 'not doing a suicide burn' involves several burns along the way to keep speed within some perceived safety margin. All other techniques involve some sort of last-minute burn the length of which depends on vessel twr.

[GoSlash27]

I'm not sure that is a proper definition of "suicide burn". Suicide burn generally means waiting until the last moment to decrease speed to a safe landing speed. It does not need to be vertical.

Starting from orbit with the intention to 'wait until the last moment' for the final burn, that means a valid option is to spend only a small amount of dV on maneuvers other than the suicide burn. It might not be the only way to do a suicide burn, but it isn't not a suicide burn and in many ways it is very different than "reverse vertical ascent-then-circularization".

If only a small amount of dV is spend on the initial 'reverse circularization' (deorbit burn), the angle to the ground is shallow so it's near impossible to precisely land where you want. But it does make for a rather efficient suicide burn landing.

I differentiate this technique as the "zero descent rate" approach. My use of the term "suicide burn" is in the generally- used context.

If i see it correctly, during the landing technique you described throttle is at max for most of the final burn, reduced only during the last few hundred meters. That is in effect a suicide burn, it takes a long time because vessel twr is low.

I think that aside from the initial deorbit burn, 'not doing a suicide burn' involves several burns along the way to keep speed within some perceived safety margin. All other techniques involve some sort of last-minute burn the length of which depends on vessel twr.

The only reason it appears that way is because I was busy taking screen caps while flying.

What differentiates this approach from the others is that the throttle setting isn't necessarily "full" at any part of the descent. The throttle setting is whatever keeps the ship midway between apoapsis and the surface at the moment.

It is not a "series of burns", but rather one continuous burn at various throttle settings.

Likewise, the pitch angle isn't a fixed value, but constantly changing to correct overshoot/ undershoot of the landing area.

What sets this technique apart from the others is that the ship is under active guidance to the desired landing area throughout the approach, which yields very high safety and accuracy.

Best,

-Slashy

Edited January 2, 2015 by GoSlash27

[rkman]

I differentiate this technique as the "zero descent rate" approach. My use of the term "suicide burn" is in the generally- used context...

Thanks for the clarification.

Though i think it is confusing to label a certain sequence of landing maneuvers (deorbit + subsequent burn(s)) with the name of one of those maneuvers ("suicide burn").

What you call a suicide burn is the sequence of a fairly large deorbit burn followed by a suicide burn:

The suicide burn is a reverse vertical ascent-then-circularization burn.

I think a suicide burn is a particular kind of burn, not a particular sequence of burns.

About the efficiency of the landing sequence that you call a suicide burn: I think the large deorbit burn contributes more to inefficiency (steep decent, much dV spend on fighting gravity), than the actual suicide burn does. A minimal deorbit burn followed by a high twr (relatively short) suicide burn would be more efficient than that (though not as efficient as a CAL).

[cerberusti]

I usually drop my pe to the surface where I want to land (or at the tallest point I need to clear.) I then make a maneuver node at pe and zero out velocity on that, then execute the burn before the node (using it as a rough guide for timing and burn duration.)

The burn is almost entirely horizontal at the last moment I am comfortable with, there is very little vertical velocity at the start of the burn as I still have orbital velocity. I thought that was a suicide burn, and most of the point of doing one was to maximize efficiency by keeping orbital velocity as long as possible?

It is fairly precise without an atmosphere, and you can always burn a little higher and earlier to glide over your landing site at a more reasonable speed if you want a precision landing next to another vehicle.

edit:

I had also been playing with realism overhaul for a while before this patch, so a single engine ignition and max throttle as the only setting were sometimes requirements. I would think a burn at the last moment would be ideal, much as pushing for orbit immediately on airless worlds is ideal (clearing terrain when necessary.)

Edited January 3, 2015 by cerberusti

[Hund]

Thanks so much for this clear concise simple method for landing. I was able to land a moon base on the first try. I usually try to land by the seat of my pants and burn alot of fuel in the process. So thanks for the tutorial.

[StrandedonEarth]

Hey thanks, can't wait to try it.

[BlackHat]

Very nice tutorial

[Nich]

Very interesting.  I guess I am confused as how the zero decent rate is the most efficient?  I would have thought an actual gravity turn would have been more efficient or are cosine losses minimal?

[Plusck]

On 26/02/2016 at 2:50 AM, Nich said:

Very interesting.  I guess I am confused as how the zero decent rate is the most efficient?  I would have thought an actual gravity turn would have been more efficient or are cosine losses minimal?

In all cases, the most efficient technique is where you spend the least possible time fighting gravity. That's a given.

The zero-descent rate / constant altitude burn is especially efficient for low-TWR craft because it gives time to kill horizontal velocity, only gradually adding a vertical vector until the very end, when TWR will be high enough to finish with what is in effect a suicide burn. In that sense it is very much like a reverse gravity turn.

Ideally, you want to lower Pe to just above the surface, and only start to burn when you get there. As with any orbit, your speed at Pe will be fastest and you benefit most from the Oberth effect. The start of your burn keeps you in orbit but lowers Ap until you deorbit, and during all of that time you are shedding only horizontal speed and gravity hasn't really affected you since you're following the curvature of the body.

From then on, the speed at which your impact point approaches will depend on TWR, and therefore automatically corrreponds to your ability to slow down in time to avoid impact. It is therefore a self-correcting system in a sense (i.e. your TWR defines how long this stage will last, which with a high TWR will be hardly any time at all), and by maintaining altitude you will ensure that the vertical vector to your thrust is the minimum necessary to avoid crashing. This phase ends when horizontal speed is low enough to turn the maneuvre into a low-speed suicide burn.

The disadvantage is the difficulty in aiming for a specific point on the surface, and the impossibility of doing this efficiently if there are mountains in the way at any point from Pe to your final landing zone. By extension, if you're wanting to avoid having aliens randomly land in your backyard, your best place to settle is just to the east of some big hills.

 

If your TWR is high enough, you can simply lower Pe a touch more so that it is below the surface, and this becomes a simple suicide burn. The risk of hitting mountains is lower (since you're coming in at an angle) but you'll never reach peak Oberth-effect efficiency. A "pure" suicide / retrograde burn will cause the point of impact to approach you quite quickly, and has to be planned for (or you die).

If you can't plan for the change in trajectory, then you need to turn this into what might be called a "constant-descent" burn which is very much like what GoSlash27 is describing here: you add a vertical component to the burn to keep your point of impact close to its initial position.

 

In any event, all these methods seem to converge to much the same thing at around the point that the initial Pe skims the surface. The thing I like most about GoSlashy's method is the increased accuracy of the approach while minimising the deviation from the maximum-efficiency retrograde-then-suicide burn which basically coverts a non-impacting orbit into touchdown in one single, pure-retrograde burn...

[Warzouz]

On 26/2/2016 at 2:50 AM, Nich said:

Very interesting.  I guess I am confused as how the zero decent rate is the most efficient?  I would have thought an actual gravity turn would have been more efficient or are cosine losses minimal?

The OP is confusing. What @GoSlash27 calls a "zero decent rate" is usually what most of us call a suicide burn. A light deorbit and a long final burn until few seconds before touch down. It usually focused on efficiency, not target landing precision.

And what @Plusk describes as "zero-descent rate / constant altitude burn" is indeed a variation using low TWR engines. In that condition you can't just follow retrograde because your vertical speed will rapidly go out of control. You have to burn partially radial to keep your vertical speed manageable while you slow down your horizontal speed. That costs a lot of fuel, but may be the only way to land with low TWR. This technique is not efficient and not precise because your main objective is focused on controlling you vertical speed, not your landing point. I used this technique to land on Tylo with a 1.00/2.15 TWR 14T SSTO ship (from 30km). I used 6000m/s(3150 to land and 2850 to return) from and to 30km orbit.

The OP method is a good in-between which has also the advantage of being quite precise at targeting the landing spot. Most useful.

Edited April 1, 2016 by Warzouz

[Esoteric106]

Thank you for this It helped me a ton