Optimal Descent (to the Mun)

[Tarmenius]

Inspired by the Optimal Ascent challenge (http://kerbalspaceprogram.com/forum/index.php?topic=14571.0), the objective here is to discover and demonstrate the most fuel-efficient method of descending to the Mun\'s surface as determined by amount of fuel remaining. I know there has been plenty of discussion on the subject in the past (and feel free to have more of it here), but I couldn\'t find a previous challenge that singled out this phase of a mission to allow more direct comparison.

So, I built a very simple lander and strapped it onto a rocket that would get it into Munar orbit. Once there, I jettisoned and destroyed the TMI/MC (Trans-Munar Injection/Munar Capture) stage then saved my persistance file which I have included below. Simply move your existing persistance file (found in the 'KSP/saves/default' folder) to another location for easy recovery, download mine and place it where yours was.

The lander is all-stock (0.15.2) and the only craft in the file. Its orbit is 100km over the equator and has just entered the day side of the Mun, so you won\'t have to wait to find a good landing spot. Also note that you will have to activate the lander stage with spacebar before you can fire the engines. I did this intentionally as a way to demonstrate that you would be starting with a 100% full tank of fuel.

My method was to make a fairly steep approach and do most of my burning over the last 2500m or so.

I freaked out a bit at around 3500m becasue the ground was coming up way too fast for my comfort level. Thankfully the lander is pretty light so I was able to reduce my descent rate in time. In the end I touched down at ~3.5 m/s with 113.3 Gubs of fuel remaining and everyone\'s just happy to be on the ground.

I think I have enough fuel to at least get back into Munar orbit so a rendezvous can be made to get the intrepid crew home, but if someone can land this and get back to orbit around Kerbin I will be impressed (but not surprised given the skills of many forum members). I say 'orbit' because the lander has no parachute or decoupler to provide a safe return.

[NEW] In the interest of satisfying curiosity, I have included a version of the lander with MechJeb. It is the second file listed.

[NEWER] As JellyCubes pointed out: 'the \'kg\' value seen when right-clicking a fuel tank has absolutely no real meaning.' Knowing this, I have unofficially renamed fuel units for this challenge. They shall now be referred to as 'Gubs,' unless something more amusing is proposed

[NEWEST] As this challenge progressed, the optimal descent method became clear. I have added a description of that method to the bottom of this post. Of course, anyone is still free to claim their spots on the Leaderboards as variables like landing location and pilot skill are still a factor.

Now it\'s your turn. Honor, glory and bragging rights are at stake here! Claim them if you can!

Exemplary Service Medal

Awarded to those pilots who succesfully returned the crew to Kerbin. Parades and talk-show appearances are assured.

Zephram Kerman

BlazingAngel665 (MechJeb assisted)

pushingrobot

mager42 (MechJeb assisted)

Leaderboard (No MechJeb):

1. JellyCubes: 154.5 Gubs

2. Apotheosist: 153.8 Gubs

3. pushingrobot: 150.5 Gubs

4. Kosmo-not: 148.6 Gubs

5. Zephram Kerman: 148.5 Gubs

6. PakledHostage: 147.6 Gubs

7. closette: 144.9 Gubs

8. tjoreilly: 144.2 Gubs

9. Tarmenius: 143.9 Gubs

10. Bluejayek: 141.8 Gubs

Leaderboard (with MechJeb):

1. mager42: 147.7 Gubs

2. BlazingAngel665: 144.4 Gubs

The Method

First a disclaimer: The following demonstrates the most efficient Munar descent method when starting from an already-established orbit. By planning your Munar Insertion properly, it is possible to use even less fuel. But those details will be for another time.

Once a rough landing site has been chosen, wait until your orbit has taken you to the opposite side of the Mun from it. Burn retrograde to bring your Periapsis (Pe) as low as you feel comfortable while still above the terrain (this will likely be a guess so don\'t worry too much about precision here). Flip back to prograde then wait until Pe, watching the terrain ahead of your craft to avoid an unscheduled (and destructive) landing and to locate a desirable landing site. Nearing Pe, you should be almost pointed retrograde. Find your landing site and be ready to finish turning retrograde (if you haven\'t already) and burn off your horizontal velocity. Since each lander design will require different burn times to eliminate horizontal velocity, you\'re going to have to make a best guess about when to begin. Toward the end of the burn, when most of your horizontal velocity has been reduced, the retrograde marker on the NavBall will begin to climb toward +90^o vertical. At this point you may want to reduce throttle and pitch over to keep up with the retrograde marker, making sure to reduce thrust enough that you are still losing altitude. Ideally, this 'reverse gravity turn' will have you pointed vertical by the time you\'re just about to land. From here, simply control your descent rate to suit the limits of your lander and you should be on the ground with plenty of fuel left (asuming you brought enough in the first place )

Where Do We Go From Here?

If you\'ve been reading the posts in this challenge, you should already be aware that the discussion has shifted toward the Insertion phase of Mun missions. That discussion has spawned a new Challenge, which can be found here: http://kerbalspaceprogram.com/forum/index.php?topic=15844.0. I want to stress again that this is not necessarily the end of Optimal Descent and I will still be monitoring for new entries as well as occasionally attempting to improve my own. A huge 'Thank You' to all who participated (and preemptively to those who may in the future); each contribution is a valuable part in finding the answer to this challenge\'s question.

[Bluejayek]

Inspired by the Optimal Ascent challenge (http://kerbalspaceprogram.com/forum/index.php?topic=14571.0), the objective here is to discover and demonstrate the most fuel-efficient method of descending to the Mun\'s surface as determined by amount of fuel remaining. I know there has been plenty of discussion on the subject in the past (and feel free to have more of it here), but I couldn\'t find a previous challenge that singled out this phase of a mission to allow more direct comparison.

So, I built a very simple lander and strapped it onto a rocket that would get it into Munar orbit. Once there, I jettisoned and destroyed the TMI/MC (Trans-Munar Injection/Munar Capture) stage then saved my persistance file which I have included below. Simply move your existing persistance file (found in the 'KSP/saves/default' folder) to another location for easy recovery, download mine and place it where yours was.

The lander is all-stock (0.15.2) and the only craft in the file. Its orbit is 100km over the equator and has just entered the day side of the Mun, so you won\'t have to wait to find a good landing spot. Also note that you will have to activate the lander stage with spacebar before you can fire the engines. I did this intentionally as a way to demonstrate that you would be starting with a 100% full tank of fuel.

My method was to make a fairly steep approach and do most of my burning over the last 2500m or so.

I freaked out a bit at around 3500m becasue the ground was coming up way too fast for my comfort level. Thankfully the lander is pretty light so I was able to reduce my descent rate in time. In the end I touched down at ~3.5 m/s with 113.3 kg fuel remaining and everyone\'s just happy to be on the ground.

I think I have enough fuel to at least get back into Munar orbit so a rendezvous can be made to get the intrepid crew home, but if someone can land this and get back to orbit around Kerbin I will be impressed (but not surprised given the skills of many forum members). I say 'orbit' because the lander has no parachute or decoupler to provide a safe return.

Now it\'s your turn. Honor, glory and bragging rights are at stake here! Claim them if you can!

Leaderboard:

1. Tarmenius: 113.3kg

Also note: This is my first attempt at submitting a challenge, so helpful suggestions are always welcome

You say it is not possible to make a safe return. I would argue with that. The lander stage looks to be easily light enough to slow down significantly under kerbin gravity. You might be surprised at how little fuel it takes to drop from 100m/s (terminal velocity) to a landable 10m/s if you time it right.

Also, an ASAS on a lander stage is a massive waste of mass!

This challenge looks like fun though, I\'ll be sure to try it later.

[Tarmenius]

I didn\'t say it was impossible, just 'unsafe.' Though honestly, 'impossible' was pretty much what I was thinking so you got me there. But I love to be proved wrong, so by all means go for it. Just make sure to record the screenshot showing how much fuel you have left before taking off again.

The ASAS is there mainly as an aid to my lack of fine keyboard skills, and I wanted to keep this accessible to anyone. The first lander I was going to use even had RCS to help kill that pesky horizontal velocity while only meters from touchdown. Then it occured to me that without a way to measure how much RCS fuel was left, the results would be less than accurate.

I\'m looking forward to seeing what you come up with!

[Kosmo-not]

I\'ll do this challenge when I finally get back to my KSP capable computer in a couple days.

[Zephram Kerman]

Here\'s my entry: 134.5 kg fuel remaining.

My trick was to start with a short retro burn to descend; basically a modified transfer orbit. The remainder of the retro burn was done at 5 km altitude. After that, it was a short fall with a few horizontal bursts to trim the remaining lateral velocity, followed by a white-knuckle burn just before impact touchdown. We estimated the final burn by comparing altitude to speed; the target was Vertical Speed = Altitude / 20.

Upon impact touchdown even Bob smiled, after a short bout of screaming. However, Jeb was concerned about being double-parked back at KSC.

So, just to see if I could do it, I decided to send them home. Umm. Bob wants to go buy a lotto ticket, as soon as he changes into a clean spacesuit.

[Kosmo-not]

Now that\'s more towards what I\'m planning to do, except I will put my periapsis at around 1km.

[tjoreilly]

Looks like a great challenge! Thanks, Tarmenius.

Circularized orbit at about 12km, retro-burned down to about 300 m/s in order to get rid of some of the horizontal velocity, and started the landing burn at about 3km. Popped a leg off the lander, but other than that, a success.

140.3kg remaining.

edit: 2nd attempt

Immediate creation of decent path (no circularization). Reduced horizontal velocity at 3km, quickly followed by landing burn. Landed at 1003m altitude. Popped two landing legs off this time.

144.2kg remaining.

[Zephram Kerman]

Now that\'s more towards what I\'m planning to do, except I will put my periapsis at around 1km.

In hindsight, that would have been much less terrifying for my little kerbals in the can.

144.2kg remaining.

Hey good shot!

[AeroPelican]

I didn\'t say it was impossible, just 'unsafe.' Though honestly, 'impossible' was pretty much what I was thinking so you got me there. But I love to be proved wrong, so by all means go for it. Just make sure to record the screenshot showing how much fuel you have left before taking off again.

The ASAS is there mainly as an aid to my lack of fine keyboard skills, and I wanted to keep this accessible to anyone. The first lander I was going to use even had RCS to help kill that pesky horizontal velocity while only meters from touchdown. Then it occured to me that without a way to measure how much RCS fuel was left, the results would be less than accurate.

I\'m looking forward to seeing what you come up with!

But ASAS only uses winglets and rcs to maneuver, so it does not do anything on the lander stage, put it on a lower stage

[tjoreilly]

Perhaps a little off-topic, but I\'ve found ASAS to be useful if one is doing a near vertical descent as it magically/gyroscopically 'locks' the crafts orientation. I normally land smaller craft like the one in this challenge without ASAS, opting for a less vertical approach to the landing site. I tend also to forego the landing gear, as I have a tendency to snap them off anyway, and just land on the engine.

3rd attempt: Immediate de-orbit to landing site, killed most horizontal velocity at 8km, vertical descent from there using the graciously provided ASAS, and a final vertical burn around 1800m to shave off most of the descent speed. All landing legs intact (in case craft condition becomes a challenge requirement), and 136.3kg remaining.

I am totally done spamming this challenge now. ;D

[Guard13007]

But ASAS only uses winglets and rcs to maneuver, so it does not do anything on the lander stage, put it on a lower stage

ASAS has it\'s own force than can applied without limit (without fuel to power it I mean), like SAS units (else, what would be the point of them?).

[Tarmenius]

I love the smell of results in the morning! Zephram Kerman, I stand corrected and thouroughly impressed that a return to Kerbin has been accomplished! I\'d say Bob should get a stack of lotto tickets. I can just picture the rendezvous crew gearing up, only to have Jeb walk in and say 'You seriously expected me to not to try it?!'

tjoreilly, I\'m glad you enjoyed it! Very nice results, indeed. I don\'t intend to stipulate that the craft needs to remain completely intact. Of course, if someone uses the rocket to break their fall and save fuel, that would kind of violate the spirit of the challenge. Even so, I could see myself making an exception for a particularly Kerbal use of that method. And you hit the nail on the head with why I included an ASAS on the lander. Without it, I have a very hard time keeping the lander oriented upright (even with CAPS-lock on) and I figured others might as well. Didn\'t want them to be turned off to this challenge because of a lander that\'s too difficult to pilot.

[Apotheosist]

This is an excellent challenge. I had thoughts about creating a similar one myself. I shall give this a go right away

[PakledHostage]

Here\'s my entry. I had 139.8 kg fuel remaining upon landing on Mun\'s surface. I went around once from the persistence file\'s start point, and then lowered my periapsis to about 1 km above terrain. I started my landing burn a few seconds before reaching periapsis. I could have probably saved a tiny bit more fuel (and avoided breaking off 3 of the 4 landing legs) if I\'d used my joystick.

Just to prove it is possible, I also flew the return to Kerbin. I didn\'t quite make the landing (splashing down at 22 m/s). The Apollo guys didn\'t get 'do overs' so neither do my Kerbals, but I\'m sure someone who\'s a better pilot than me could do it on the first try.

Edit: I just noticed that Zephram beat me to it, not only returning to Kerbin but also landing successfully. Nice job.

[Tarmenius]

Good work, PakledHostage! Looks like I\'m going to have to gather my courage and try for a return trip myself.

Has anyone tried a 'direct descent' method yet? I\'d be interested to see how it stacks up. I\'ll be trying it, of course, but I\'m not exactly the most skilled pilot around

[Apotheosist]

This is the best I\'ve had so far: 148.7 kg. If I remember correctly, I waited until apoapsis, (the initial orbit was ever so slightly elliptical, probably hardly made a differnece really) then retro burned until my periapsis was about as low as I dared go, then at periapsis I retro burned again to bring periapsis down to about 1k. I then coasted for a while, slowly losing altidude, and all the time I was looking for a suitable landing spot. I went into a crater, and it didn\'t look like I\'d make it over the rim the other side. So I decided to completely cancel my horizontal velocity there, then align myself vertiacally, wait for a bit, retro burn at the very last second, then turn off engines while still moving down with a few m/s of velocity, letting my landing legs absorb the last of my downward velocity.

[Bluejayek]

103.8 for a direct descent.

[Tarmenius]

Congrats Apotheosist! You\'ve taken the lead.

After attempting a 'direct descent' where I simply cancelled out as much horizontal velocity as I could then fell for 95km, I discovered I\'m not very good at judging distances. It tested my nerves the last few km before I simply could stand it, and I didn\'t quite get the timing right. I ended up having to maintain a small amount of thrust for the rest of the descent, leaving me with about 107kg remaining fuel. So that was no good.

Then I tried something close to what the rest of you have done. I even managed to pass through part of that huge Mare-side ravine. 8)

Unfortunately, I realized too late that my orbit would take me through one of the peaks at the other end (It\'s that teeny, tiny little bump on the horizon directly above my craft). This occured to me soon after the awe wore off from buzzing that slope to the immediate left of the 'little bump.' I think Bob knew long before I did.

So, making sure not to repeat that little accident, I tried again. It was a pretty smooth run, and I ended up with 130.3kg remaining.

Bluejayek: Out of curiosity, did you burn all at once just before touchdown, or keep a steady descent rate?

[Bluejayek]

Managed 141kg with a dropped periapsis as many have done. Botched the kerbal return and crash landed in the ocean at 29m/s.

I supposed I am cheating however, I have the fps pod addon installed which gives me proper altitude.

[Tarmenius]

Good job! I\'ve updated your position on the Leaderboard. I\'m not going to consider it cheating to use that addon, since the true goal of this challenge is to find the ideal approach method. Having an accurate ground-relative altitude reading will only help make more accurate decisions about when to thrust and therefore a more reliable method. Anyone else who wants to use it is free to as well.

[Edit]: In fact, I considered making a persistance file with MechJeb included for comparison. If you guys want, I would be happy to do it.

[semininja]

I\'m actually kinda curious how MechJeb stacks up with its auto-land function...

[Tarmenius]

So am I. And if there\'s one thing I\'ve learned about myself, it\'s that I often can\'t deny my curiosity. So as soon as I get back from watching my local fireworks display, I will get the MechJeb version set up.

[EDIT] MechJeb version now added. By the way does anyone know how I might list the attachments in such a way that makes it easier to distinguish which is which? I feel as though simply saying 'The first is stock; the second has MechJeb' isn\'t good enough.

[PakledHostage]

Here\'s another entry, this time flown using my joystick. Otherwise, I used the same approach as my earlier entry: Lower my periapsis to about 1 km above terrain then perform my pre-landing burn at periapsis, while allowing the spacecraft to settle towards the ground. When horizontal velocity has been eliminated, turn perpendicular to the surface and arrest the remaining vertical velocity for landing.

Edit: By my figures, the spacecraft needs to slow by just shy of 50 m/s to drop its periapsis to 1 km above terrain. It then needs to decelerate by a further ~610 m/s at periapsis, followed by a last little bit of (difficult to quantify) delta-V prior to landing safely. I used a total of 708 m/s delta-V in my best attempt, while Apotheosist used 699 m/s. That doesn\'t leave much room for improvement for this method. Does anyone have any other ideas to try? What about a reverse gravity turn?

[Zephram Kerman]

Has anyone tried a 'direct descent' method yet? I\'d be interested to see how it stacks up. I\'ll be trying it, of course, but I\'m not exactly the most skilled pilot around

I tried direct descent for my first attempt. Complete de-orbit burn at the starting 100k altitude, drift all the way down with a few horizontal bursts to cancel lateral velocity ahead of time, then braking burn like crazy for the last 4 km. The result was so bad I didn\'t even bother with a screen shot.

[Tarmenius]

Zephram, that\'s pretty much the same result I\'ve been getting. With MechJeb, the LAND button does a mostly-direct descent and the results there were equally bad.

I\'m going to play around with the Landing Autopilot a bit more before posting an official entry. My initial attempt was to select a landing site a good distance away, and MechJeb made a pretty shallow approach, touching down with 130kg remaining. I\'m going to see if distance to landing site determines how shallow the approach is and how that effects fuel consumption.

PakledHostage, what do you mean by a 'reverse gravity turn'? Is that where you start out burning horizontally and let the craft slowly drift to vertical? If so, It\'s not a method I\'ve tried yet but can definitely give it a go.