Selecting A Landing Location When Aerobraking?

[The Flying Kerbal]

To keep it simple, let's just say you have a vessel returning home from Mun or Minmus and you're going to use aerobraking to slow the thing down in preparation for landing.  Is there any way of timing where and when you enter Kerbin's atmosphere so that you can be reasonably confident of landing somewhere safe as opposed to slamming into the side of a mountain?

I've had a few close shaves at times, and one or two impacts on mountain slopes which have proven fatal, and I've never been able to see how to avoid this.  So is it really just a question of dropping your Pe. into Kerbin's atmosphere and trusting in the gods, or can a descent like this be made a little less haphazard?

 

Thanks everyone.

[Vanamonde]

Setting your PE at 32kms or less will brake you to a vertical fall from just about any starting speed. So if you're still moving laterally when you're low enough to hit the ground, you are setting your PE WAY too low, exposing the ship to unnecessary deceleration Gs and possibly excessive heat. This is true even of planes. Bring it down belly-first and it should still slow enough to switch to controlled horizontal flight during which you can steer the thing away from obstacles. 

[The Flying Kerbal]

I'm gonna have to try that, I normally set it at 35 Kms.

Thanks for the answer. 

[katateochi]

When returning from Mun, I tend to put the Pe around 40-45km (depends on craft) which is usually enough to lower the Ap to a few 100 km, then you can trim it down a bit more on another pass (assuming you've got some thrust left to adjust Pe once at the Ap again).  If I'm not using mods to help predict landing site then my approach is aerobrake into an elliptical orbit, trim it down to a more circular orbit with another 1-2 passes and then I know where to perform a deorbit burn over Kerbin to get back to KSC.

If you're using Mechjeb you can use the landing autopilot to help predict landing site (without actually using the autopilot).  I'm not sure how accurate it is in the current versions, but I used to use it quite a lot to predict aerobrake and landing site. open up the landing autopilot tab and then select the predict landing toggle (can't remember exactly what it's called) and that will show a red line of predicted descent. 

[Geschosskopf]

19 hours ago, The Flying Kerbal said:

To keep it simple, let's just say you have a vessel returning home from Mun or Minmus and you're going to use aerobraking to slow the thing down in preparation for landing.  Is there any way of timing where and when you enter Kerbin's atmosphere so that you can be reasonably confident of landing somewhere safe as opposed to slamming into the side of a mountain?

The only way to be completely safe when landing is to return to LKO, wait until it's daylight at the place you want to land, then de-orbit.  Landing directly from Mun or elsewhere is a crapshoot.  Most times you'll be OK because the bulk of Kerbin's surface is safe.  However, sometimes you will come down in the mountains.  If it's daylight, you can maybe make some last-minute emergency burns to avoid them, but you have about a 50/50 chance of arriving in the dark so you just have to take your chances.

The Trajectories mod is quite useful for all this because it shows, in the map view, your predicted trajectory after atmospheric interaction.  This includes not only your landing spot if your Pe is low enough, but also your post-atmosphere Ap if you stay high enough to just aerobrake.  Thus, you can aerobrake into a reasonably low parking orbit without needing much fuel.  There you wait 1 or more orbits until the area you want to land in is in daylight.  Then you can set up your de-orbit burn to make sure you land in exactly the spot you want.

NOTE that Trajectories doesn't take Kerbin's rotation into account.  It just draws a predicted path from the point in space where you enter the atmosphere, but Kerbin rotates under this point.  Thus, if you set up a landing using Trajectories when you're still all the way back at Mun, Kerbin will rotate 1 or 2 times before you actually get there, and there could be a mountain at your destination by then.  IOW, even with trajectories, trying to pick a landing spot from all the way back at Mun is still a crapshoot.

[OhioBob]

20 hours ago, The Flying Kerbal said:

To keep it simple, let's just say you have a vessel returning home from Mun or Minmus and you're going to use aerobraking to slow the thing down in preparation for landing.  Is there any way of timing where and when you enter Kerbin's atmosphere so that you can be reasonably confident of landing somewhere safe as opposed to slamming into the side of a mountain?

Yep, that's fairly easy to do.  This is the method I use...

What I do for a return from Mun or Minmus is to set the periapsis at Kerbin at about 22-24 km.  At that height I find that I usually land at a point that is just about directly beneath the periapsis.  Of course it depends on the aerodynamic characteristics of the reentry vehicle, but if we're using anything like a Mk1-2 command pod, we'll be pretty close.

After setting up the maneuver node for the transfer back to Kerbin, I check the time to periapsis.  Let's say the time to periapsis is 1 day, 1 hour, 45 minutes.

Although we expect to land directly beneath the current periapsis, the terrain beneath the periapsis at the time of landing is not what it is now.  We have to take into account Kerbin's rotation.  In one day Kerbin will return to its exact current position, so we can ignore the number days.  What is important is the hours and minutes, because that represents how much Kerbin will change from current.  In this example we have 1:45, or 105 minutes.  Kerbin rotates one degree per minute, so in 105 minutes it will rotate 105 degrees.  Therefore the landing site is going to be 105 degrees west of the spot that is currently beneath the periapsis.

You can estimate the 105 degrees by eyeballing it, or you can use a map like that below.  The white vertical lines are spaced 15 minutes (15 degrees) apart, with the longer lines being 1 hour (60 degrees) apart.  Just find the current spot under the periapsis and move the appropriate distance to the left to find the predicted landing site.

The time to periapsis that you read with the maneuver node may not be the same as you get after completing the burn (there's always going to be some error).  Therefore it pays to double check the landing site after you've completed the burn to make sure it hasn't change by a lot.

If you don't like the looks of the predicted landing site, then just delay performing the transfer burn for a later orbit.  Let's say you're orbiting Mun once every 45 minutes.  That means waiting one orbit to perform the burn will shift the landing site about 45 degrees to the west.  This way you can plan ahead and schedule the burn for the time that will place the landing site right where you want it.

[FancyMouse]

51 minutes ago, Geschosskopf said:

NOTE that Trajectories doesn't take Kerbin's rotation into account. 

It does. It marks a red cross on the surface for your landing spot, which is not necessary at the end of your trajectory path. The path doesn't take rotation into account, but the red cross does. I don't have KSP right in front of me so I don't remember for sure if this one is an optional setting or not, but the red cross is definitely there for quite a long while and I've been using it a lot.

Edited December 18, 2017 by FancyMouse

[Snark]

My own solution to this problem,

21 hours ago, The Flying Kerbal said:

Is there any way of timing where and when you enter Kerbin's atmosphere so that you can be reasonably confident of landing somewhere safe as opposed to slamming into the side of a mountain?

...is to just aim to land in the middle of a huge mountainless expanse, either in the middle of the ocean or on broad plains.  That way, it doesn't matter if my actual landing spot is a few dozen kilometers  away from the "ideal perfect" location. 

Another technique is to just design the reentry vehicle to be low and squat, so that it doesn't matter if it lands on a slope.  This is what really makes a difference.  As long as the vehicle isn't tall and "tippy", it can descend on parachutes and land basically anywhere on the planet and it'll be fine.  Even in the very rare occasion when it lands on a so-steep-it's-practically-a-cliff mountainside (which has happened to me once or twice) ... doesn't matter.  It touches down, then slides down the steep slope with the parachutes still keeping it manageably slow, until it comes to rest at the bottom.

One thing that the above two suggestions don't accomplish is the ability to land in a specific spot.  Both of these suggestions are about solving the problem by making it so that it doesn't matter where you come down, rather than precisely targeting the landing location.  I like this because, 1. in my experience, it's a whole lot easier to do, and 2. in my own gameplay, "where exactly do I land" almost never matters.

However, every once in a while, I might have a situation where it does matter exactly where I come down.  For example, if I've landed one ship on an atmospheric body, and then I want to land another ship right next to it.  (Doesn't come up very often, given the way I play, but it does sometimes occur.)  In such cases, the solution is to give the descending ship a bit of aerodynamic control, so that I can pilot it to the desired touchdown site.

It doesn't have to be a lot of aerodynamic control-- for example, I don't have to build a fully-functioning, fully-controllable airplane with wings and tail and such.  It just needs to be enough to fly like a "controlled brick".  For example, if the descending ship is mainly a cylinder, just stick a couple of AV-R8 winglets on the back.  That's enough to give pretty good attitude control, and I don't need wings because I can use the cylinder's "body lift" to fly the thing like a glider with a really crappy glide ratio.

Or, if the ship's pretty small and has a fair amount of reaction torque relative to its size, you may not even need any aero control surfaces at all.  For example, consider a ship consisting of Mk1 command pod, sitting on top of a 2-ton LFO tank, on top of a Terrier.  That's a fairly common reentry vehicle I use for returning from LKO (it's small and light enough that it doesn't need a heat shield).  Just the reaction torque from the command pod lets me "fly" it, engine-first, pretty well.  By pitching the ship up or down, I can make a big difference in how fast it descends.  It's controllable enough, in fact, that often when it gets down to about 10 km, I can even cancel out the descent entirely and actually ascend for a kilometer or two before it slows down enough to drop once more.  It's remarkably controllable.

So I can put down a ship like that remarkably close to where I want it.  Use pitching-up-and-down for aerodynamic control to fly it to where I want to be, such that my trajectory would take me directly over (and overshoot) my intended landing position.  Then, right when it gets over the desired spot, pop the chutes.  They very quickly slow the ship to an effective halt, then it drops straight down.  It won't give you the to-the-nearest-meter control that you'd get with a spaceplane, but it's good enough to land within a few hundred meters of where you want to be, without too much trouble.  Occasionally I'll fly the Mk1pod-plus-Terrier ship described above to land within the bounds of KSC, just for fun target practice, and it works pretty well.

[Corona688]

It almost never matters for me.  I don't aim at all whatsoever, and hit an ocean, plain, or desert.  The odds of hitting anything else are really slim.

[OhioBob]

10 minutes ago, Corona688 said:

The odds of hitting anything else are really slim.

But not zero.  It was my very first Mun mission in which I killed Jeb by having him land on and tumble down the side of a mountain.  Although the odds are slim, up to that point it had happened to me 100% of the time.  (Of course that percentage has dropped dramatically since then.)  I'm now always a little paranoid of mountains and try to avoid them whenever possible.

[OrbitalBuzzsaw]

The thing to do is to do a 2-pass aerobrake - get your AP to something reasonable, and then boost back into orbit. Then land like a normal SSTO.

[Geschosskopf]

5 hours ago, FancyMouse said:

It does. It marks a red cross on the surface for your landing spot, which is not necessary at the end of your trajectory path. The path doesn't take rotation into account, but the red cross does. I don't have KSP right in front of me so I don't remember for sure if this one is an optional setting or not, but the red cross is definitely there for quite a long while and I've been using it a lot.

Well now, that's very nice to know.  Thanks!

[The Flying Kerbal]

Terrific answers as usual guys.  I'm currently cobbling together an experimental ship to test some of your suggestions out.  Val volunteered, or rather her buddies all took a step back when I asked for volunteers to step forward, and will shortly be sent hurtling skywards to put your ideas through their paces.

[Hotel26]

Great ideas.  I think I would do the aerobraking down to an altitude that will reduce my speed to a point I can get into a "nominal" LKO.  Then "square up" that orbit [circularize at, say, 70x70km]

From there, when I pass over a point in the daylight I like [KSC!!!!], I boost the apoapsis with a short prograde fire that raises my apoapsis just enough so that it "marks" the other side of my orbit.  Then schedule a burn at that apoapsis to reduce periapsis to some number, 32-35km (I am reading from this post), from which I know that the trajectory will degrade to a landing at that point.  [The downside of this is that raising the apoapsis slightly increases your penetration speed when you return to the periapsis for landing.]

You could easily just wait half of your orbital period and then lower your periapsis to whatever your particular vee-hickle likes for landing...    From the mechanics of von Hohmann [as long as you are in a circular orbit] you know you are operating on the other [original] side of the orbit...

It's kinda nice to eyeball your landing field going over it, 70 km above, and think, "I'll seeya in one more loop..."

EDIT: I've course.  I didn't allow for the rotation of Kerbin in 31m21.44s, did I?  But then that's factored into whatever periaptic altitude you know works for the V-Hickle...    Easy.

Edited December 20, 2017 by Hotel26

[Hotel26]

On 2017/12/19 at 12:59 AM, Snark said:

It doesn't have to be a lot of aerodynamic control-- for example, I don't have to build a fully-functioning, fully-controllable airplane with wings and tail and such.  It just needs to be enough to fly like a "controlled brick".  For example, if the descending ship is mainly a cylinder, just stick a couple of AV-R8 winglets on the back.  That's enough to give pretty good attitude control, and I don't need wings because I can use the cylinder's "body lift" to fly the thing like a glider with a really crappy glide ratio.

You might be right about that "brick thing".

Aquila Torpedo (a.k.a. "Bullseye; a.k.a. "Buzzbomb") steers in for touch-down at K.S.C.  It overshot and had to reverse course to do it, too.

Yeehaaah! (Just needs an EVA seat and a KERBAL strapped in real tight.)

"Let's park on the northern side, today..."

"Under chutes. Watch out for the engine shrapnel, BELOW!"

Happiness

TWO control towers; ONE airport. There's gonna be a fight...

Rather amazingly, the Torpedo can actually be coaxed into a controlled ascent.  It was a wild bronco until I added a reaction wheel.

Edited December 21, 2017 by Hotel26