Stable Glider Challenge: Make a glider that flies really efficiently, all by itself.

[Snark]

Your mission:  Build a glider that flies as stably, uniformly, and efficiently as possible with no control input.

(Note:  Requires installing a small utility mod to calculate "score" in flight, see below.  But it's a "safe" mod with no footprint on your craft or save files, so you can uninstall as soon as you don't need it.)

 

The rules:

1. Your craft must include a Mk1 Cockpit.
   • This rule is mainly to encourage a uniform payload so that the designs have a similar "starting place".
   • If you come up with something amazing that doesn't have a Mk1 Cockpit, feel free to post it for our entertainment!    It can still be eligible for the "most interesting" category, see below.
2. Stock parts only.
3.  No mods that affect gameplay (e.g. physics).
   • Mods that don't affect gameplay (e.g. informational, or visual F/X) are fine.
   • No, I haven't set up a separate category for FAR.  If there's a lot of demand for it, I could do so, I suppose.
4. Put your craft into atmospheric flight on Kerbin at "reasonably low" altitude (under 2000 meters).
   • If you come up with something awesome that only works at high altitudes, feel free to submit.  Could be a candidate for the "most interesting" category. 
5. It must have no engine power at all-- a purely passive glider.
   • I don't care at all how your glider gets aloft, or whether it stages away equipment before gliding, or anything like that.
   • For example, it's fine if you build an airplane whose "nonessential for gliding" equipment (such as landing gear, engines, etc.) get staged away before you begin your glide.
   • It's fine if it has engines on it, as long as they're not operating during the glide.  (The scoring mod will yell at you if you have nonzero throttle, though, so that should help to keep you honest.) 
6. It must be capable of landing safely without damage after gliding.
   • Water landing is fine, though, as long as you can do it without breaking up.  So this is pretty easy even if you've jettisoned your landing gear or something.
   • By "can land safely" I mean after you take control.  But if your glider will actually land safely while uncontrolled during its glide (especially on land), please mention that!  Nice bragging rights.
7. Score will be measured when your craft is in uniform, stable flight without any control input (either from you, or from an autopilot like MechJeb).  There are separate scoring categories for SAS on or off, detailed below.
8. No "exploits", please.  The purpose of this challenge is to be an exercise in engineering optimization, not "find a clever loophole in the rules."  (For example, the scoring mod, GliderStats, is pretty basic and I wouldn't be surprised if you could outsmart it by finding some edge case it doesn't account for.  Please don't do that.)

 

 

How it's scored

I've written a very simple little mod, GliderStats.  It's low-footprint and has no effect on either your save game or your craft files, so it's safe to install for this challenge and then uninstall when you're done with it.

What the mod does is to display some stats about your glider flight in a little display next to the navball:

Those two numbers (ratio and descent) are what you get scored on.  There are separate scoring categories for "highest glide ratio" and "lowest rate of descent".

Note:  the mod works by evaluating flight parameters over a 10-second sliding window of time.  The goal here is for you to report numbers after you have achieved a stable uniform glide.  This is important, because a glider can easily "porpoise" (a.k.a. phugoid oscillations), i.e. dive and accelerate, then pitch up while slowing down.

Obviously, if you have a glider that's moving artificially fast-- either because it just dived, or because you used engines to boost it up to high speed before shutting down-- then you could have a "glide ratio" that approaches infinity.  This contest is not about that, so don't do it.  The scoring mod tracks how "stable" you are, by looking at the percentage variation in your speed over the last 10 seconds.  If it's over 1% variation, it includes a warning to that effect in the displayed message.  So get that "stabilization percentage" down under 1% if possible before reporting your score.

Note:  the mod is designed to "reward" stability.  When it reports the glide ratio and the rate of descent, it deliberately reports the worst value from the last 10 seconds (i.e. lowest glide ratio, highest rate of descent).  So it's to your advantage to be as stable and uniform as possible, because if you're oscillating, your score will be from the wrong end of the oscillation. 

 

Submission guidelines

Please include a screenshot of your glider in flight, with the navball display so we can see what the mod's reporting.

Also, please share your .craft file somewhere public (e.g. dropbox, Google Drive, KerbalX, whatever) so people can download and try it out.

 

Scoring categories

I'll maintain separate leaderboards for the following:

• Separate categories for "highest glide ratio" and "lowest rate of descent".
• Separate categories for "free" versus "controlled" glide.  "Free" means absolutely no control input whatsoever, meaning hands-off the keyboard and SAS is turned off.  "Controlled" allows that, e.g. if you've got SAS turned on.  "Free" is more challenging than "controlled", which is why it's a separate category.  The mod's status display indicates which mode you're in.
• Plus one completely subjective "most interesting" category that's based on my arbitrary judgment of what's cool.  Feel free to submit a glider that breaks any or all of the above rules; if it's especially fun or interesting IMO, it's eligible for this category. 

 

Leader boards

Highest glide ratio, controlled glide:

1. @michal.don, 26.526
2. @no_intelligence, 26.308
3. @Steel Starling, 26.192
4. @qzgy, 23.865
5. @adsii1970, 3.620

Lowest rate of descent, controlled glide:

1. @michal.don, 1.750 m/s
2. @no_intelligence, 1.920 m/s
3. @Steel Starling, 1.954 m/s
4. @qzgy, 2.204 m/s
5. @adsii1970, 6.756 m/s

Highest glide ratio, free glide:

1. @g00bd0g, 24.120
2. @AeroGav, 17.230

Lowest rate of descent, free glide:

1. @g00bd0g, 2.115 m/s
2. @AeroGav, 2.123 m/s

Fun/interesting designs:

(This category's not "ranked", it's just in chronological order of submission):

• (awaiting first entry)

Edited October 18, 2017 by Snark
New entries by michal.don and no_intelligence

[Mr. Peabody]

Looks like a great challenge! Will there be a badge? 

[cratercracker]

Just now, The Dunatian said:

Looks like a great challenge! Will there be a badge? 

*sweats*

[Snark]

4 minutes ago, The Dunatian said:

Looks like a great challenge! Will there be a badge? 

Hadn't planned on it, mainly because I've never had anything to do with badges.  (And am generally totally unaware of them anyway, since I leave everyone's sigs hidden to conserve screen real estate.)  Plus I have no artistic talent.

If someone else wants to design a badge, though, I have no problem with the idea.    Mainly would be a matter of deciding what the badge criteria are, I suppose.  How do people usually do the badge thing?  You get one just for participating?  Or for doing better than some arbitrary score level?  Or what?  I'm open to suggestions.

[AeroGav]

Oh my.        There's two issues I struggle with at times - phantom rolling forces that used to occur if you attached parts in a certain order -eg.  put engine nacelles on the side of the fuselage, and put the wings on the nacelles, craft would tend to roll in flight despite dihedral,   being high winged , having huge vertical stabilizers etc.    It's not as bad as it used to be but i still have to make corrections every so often when flying space planes to orbit.      I have never, in KSP been able to make an aircraft that comes out of a slight bank by itself.   They always just keep rolling.

Second problem I have with making a totally hands off plane are phugoid oscillations.   Though i guess jet engines are to blame for these, because their power increases at the bottom of a dip due to the thicker atmo and higher speed, setting up the next cycle.

[Snark]

17 minutes ago, AeroGav said:

 I have never, in KSP been able to make an aircraft that comes out of a slight bank by itself.   They always just keep rolling.

Well, all I can say is that I just started playing with some gliders this week (I'm generally not an airplane guy), and my first three attempts all are fine on roll stability when in completely free mode, so apparently it is possible.    Not that I'm calling you a noob or anything, maybe I just have beginner's luck!

The planes I speak of are the one pictured in the screenshot in the OP, and also a couple of biplanes I experimented with.

Spoiler

Of course, the above plane would not be eligible for this challenge, since it doesn't have a Mk1 cockpit on it.

And in any case, even if you struggle with roll stability, there's always the "controlled" scoring category, where you can just leave SAS turned on.

17 minutes ago, AeroGav said:

Second problem I have with making a totally hands off plane are phugoid oscillations.

Okay, I admit, I hadda go look that one up.    (The term sounded to me like some sort of medical condition.  "I'm sorry, ma'am, I'm afraid your husband has phugoid oscillations.")  New vocabulary word for me, thanks!  I've added it to the OP.

Gliders do that too, but a stable one (such as the ones pictured above) will damp out the oscillations in a little while.

That's why it's a challenge. 

 

[icantmakemodels]

A well thought-out challenge.

 

Hold my beer.

[Snark]

Hmmm, was thinking about how badges might work with this challenge.  One idea would be to have certain "achievements" (quite aside from the leaderboard), which you get a badge for if you pull them off.

Some ideas for such an achievement:

• Make a stable glider.  "Stable" is defined as "uncontrolled, can glide until the scoring mod shows it as stable", i.e. less than 1% variation in speed over 10 seconds.  Must have a glide ratio of better than, say, 4.  (Need to specify a minimum glide ratio; otherwise a capsule on a parachute would qualify.)  Level 1:  with SAS on.  Level 2:  with SAS off.
• Make a glider that safely lands itself on land, while uncontrolled during its glide.  Level 1:  with SAS on.  Level 2:  no SAS.  (This one sounds hard, though I haven't tried it yet.  Perhaps best to wait until someone's actually done it before taking the trouble to make a badge.)

Any other ideas?  What do folks think?

[adsii1970]

Hmmm, very do-able

[AeroGav]

OK,  how about this to get started ?

https://www.dropbox.com/s/zgo3dp28l7ut1x2/Pterodactyl.craft?dl=0

I must emphasise this was built as a min-maxing exercise re: glide ratio

but also fairly high on my priorities was trying to make it fly "hands off"

The phugoids never entirely disappeared but were down to 2.5% or so by the end of the glide.

This plane lacks landing gear or any control surfaces !     It comes with a pair of Spark rockets to get it to altitude, which decouple.  I used "hack gravity" to get it off the runway without damage.

I was really pleased with the lateral stability yet it is gradually turning north the whole flight.   We started our glide at 090 and despite the roll angle  being a constant 0.2 deg.   with negligible sideslip,   we were at 045-040 heading by the time it landed !

Design philosophy  - I tried to minimise non-wing parts.        I used editor extensions redux to put an incidence angle on the wings.

2 degrees incidence on the main wing.    The strakes and canards are at 3 degrees, which gives it a tiny bit of nose up trim, to stop it gliding nose-low.  The wings are above the fuselage and have 4 or 6 degrees of dihedral,  and the twin strake tailbooms provide extra yaw stability.

[Snark]

30 minutes ago, AeroGav said:

OK,  how about this to get started ?

Nice!    We have our first entry.

[Galane]

Airdrop, like an M2-F1? https://www.nasa.gov/centers/dryden/history/pastprojects/Lifting/index.html

[Scarecrow]

I'm having a problem with this.  Have a pod and boom traditional style glider that flies quite happily hands off.  The little app shows the glide ratio is still going up slowly and the sink rate is decreasing slowly, so I am holding off taking my screen shot until they reach max values.  Unfortunately, as anyone knows who has done long distance flights on Kerbin, as you progress in your flight, your nose starts to rise relative to the horizon and this seems to be causing my craft a problem.  While I am waiting for my figures to reach peak, the nose is getting close to the horizon and then suddenly, with absolutely no input from me (and none for several minutes previously) the stabilizing message comes up and starts increasing slowly.

I don't think this is the phugoids mentioned above, as the craft is settled in a steady flight, so curious if anyone else has experienced this.

[herbal space program]

I've spent a fair bit of time on a similar problem to this before,  around a "paper airplane" challenge that somebody posted a couple of years ago, so I think I've got a bit of a lead on this one. The key to it IMO will be to dilute the weight of the cockpit with as much pure wing area as possible, while maintaining sufficient stability and control. With just an ECS, a couple of control surfaces, and wings, it was possible to get 40km downrange from a 4000m apoapsis. for  glide slope of 1:10. I think the air is more favorable since I did that, so I can probably do better now. Of course diluting that cockpit enough is going to require a whole lot of wing area, so the winning craft could be pretty amusing. I'm definitely going to give it a try after I get done posting my entry for the Retrosolar Rescue....

[AeroGav]

6 hours ago, Scarecrow said:

I'm having a problem with this.  Have a pod and boom traditional style glider that flies quite happily hands off.  The little app shows the glide ratio is still going up slowly and the sink rate is decreasing slowly, so I am holding off taking my screen shot until they reach max values.  Unfortunately, as anyone knows who has done long distance flights on Kerbin, as you progress in your flight, your nose starts to rise relative to the horizon and this seems to be causing my craft a problem.  While I am waiting for my figures to reach peak, the nose is getting close to the horizon and then suddenly, with absolutely no input from me (and none for several minutes previously) the stabilizing message comes up and starts increasing slowly.

I don't think this is the phugoids mentioned above, as the craft is settled in a steady flight, so curious if anyone else has experienced this.

I thought the problem of the nose slowly rising relative to Kerbin was due to flying with SAS in Stability Assist mode ?  Because SAS tries to hold the nose at a constant angle in absolute terms, it fails to take account of the curvature of the planet.      If you're going for an award in the "free glide" category,  you need to fly with SAS off.   You can adjust the nose angle it glides at by either

• using the fine rotation tool in the SPH on your tailplane or canard to give some built in pitch trim
• In flight, click "Deploy" on the tailplane or canard so it makes the nose go up.  If it makes the nose go down instead, try toggling the "deploy direction" from "normal" to "inverted".  This probably sent your airplane shooting up to a stall - right click on the "limit authority" slider so you've got just the right amount of nose up for a good glide.  Whilst Snark's App sees this as a "free glide", i'm not sure it's within the spirit of what constitutes "free gliding"

Note, if you're happy with a controlled glide instead,  the above two tricks work really well with SAS set in Prograde Hold mode (just make sure the Navball is set to surface) since that damps out other disturbances.   You need to apply a lot more deflection to the control surface to get the same amount of nose up however, since the SAS is going to fight back.  If youi're using the "Deploy / Limit authority " trick, bear in mind this achieves nothing if that surface also has pitch control enabled and is in fact your only pitch control device.

 

Incidentally,    it gets hard to calculate glide ratios at hypersonic speeds, when orbital freefall effect is still having a major influence.    

Val was watching the challenge from orbit, and thought it looked so much fun, she wanted a go -

 

 

Don't think that's quite in the spirit of the challenge though..

@Snark   BTW the Pterodactyl landed without damage in the water while uncontrolled in the glide - not hard to do mnid you.

[cratercracker]

14 hours ago, Snark said:

Hmmm, was thinking about how badges might work with this challenge.  One idea would be to have certain "achievements" (quite aside from the leaderboard), which you get a badge for if you pull them off.

Some ideas for such an achievement:

• Make a stable glider.  "Stable" is defined as "uncontrolled, can glide until the scoring mod shows it as stable", i.e. less than 1% variation in speed over 10 seconds.  Must have a glide ratio of better than, say, 4.  (Need to specify a minimum glide ratio; otherwise a capsule on a parachute would qualify.)  Level 1:  with SAS on.  Level 2:  with SAS off.
• Make a glider that safely lands itself on land, while uncontrolled during its glide.  Level 1:  with SAS on.  Level 2:  no SAS.  (This one sounds hard, though I haven't tried it yet.  Perhaps best to wait until someone's actually done it before taking the trouble to make a badge.)

Any other ideas?  What do folks think?

How can I help you, mr. @Snark?

[Scarecrow]

3 hours ago, AeroGav said:

I thought the problem of the nose slowly rising relative to Kerbin was due to flying with SAS in Stability Assist mode ?  Because SAS tries to hold the nose at a constant angle in absolute terms, it fails to take account of the curvature of the planet.      If you're going for an award in the "free glide" category,  you need to fly with SAS off.   You can adjust the nose angle it glides at by either

• using the fine rotation tool in the SPH on your tailplane or canard to give some built in pitch trim
• In flight, click "Deploy" on the tailplane or canard so it makes the nose go up.  If it makes the nose go down instead, try toggling the "deploy direction" from "normal" to "inverted".  This probably sent your airplane shooting up to a stall - right click on the "limit authority" slider so you've got just the right amount of nose up for a good glide.  Whilst Snark's App sees this as a "free glide", i'm not sure it's within the spirit of what constitutes "free gliding"

Note, if you're happy with a controlled glide instead,  the above two tricks work really well with SAS set in Prograde Hold mode (just make sure the Navball is set to surface) since that damps out other disturbances.   You need to apply a lot more deflection to the control surface to get the same amount of nose up however, since the SAS is going to fight back.  If youi're using the "Deploy / Limit authority " trick, bear in mind this achieves nothing if that surface also has pitch control enabled and is in fact your only pitch control device.

 

Incidentally,    it gets hard to calculate glide ratios at hypersonic speeds, when orbital freefall effect is still having a major influence.    

Val was watching the challenge from orbit, and thought it looked so much fun, she wanted a go -

 

 

Don't think that's quite in the spirit of the challenge though..

@Snark   BTW the Pterodactyl landed without damage in the water while uncontrolled in the glide - not hard to do mnid you.

OK.  Didn't realise SAS needs to be switched off, as in the image posted by the OP it's switched on.  Actually having just re-read the rules it would appear there are two categories - for SAS on or off.

Edited September 15, 2017 by Scarecrow

[Snark]

On 9/15/2017 at 0:05 AM, AeroGav said:

Val was watching the challenge from orbit, and thought it looked so much fun, she wanted a go -

Don't think that's quite in the spirit of the challenge though..

Yeah, I'm not gonna put that up on the leaderboard.    This goes back to,

On 9/14/2017 at 10:38 AM, Snark said:

the scoring mod, GliderStats, is pretty basic and I wouldn't be surprised if you could outsmart it by finding some edge case it doesn't account for.  Please don't do that.

This would be one of those edge cases.  It's pretty easy to get an insanely high "glide ratio" if you're just skimming the top edge of the atmosphere, where there's practically no drag and you're getting all your "lift" from the fact that you're going at orbital speeds and therefore effectively don't weigh anything.

Thanks for sharing, though! 

On 9/15/2017 at 4:02 AM, Scarecrow said:

Actually having just re-read the rules it would appear there are two categories - for SAS on or off.

Yep.  There are really two challenges, here:  building something efficient, and building something stable.  The "SAS on" category is to open the door to people who want to focus on the former without having to worry too much about the latter.

[Scarecrow]

The problem I have is that my design will only fly with SAS on, but it will do so for several minutes completely hands off.  Unfortunately it isn't able to reach it's max LD and minimum sink before the stabilizing message appears, counting up from 1%, which makes it look as if there has been some control input.

[herbal space program]

30 minutes ago, Scarecrow said:

The problem I have is that my design will only fly with SAS on, but it will do so for several minutes completely hands off.  Unfortunately it isn't able to reach it's max LD and minimum sink before the stabilizing message appears, counting up from 1%, which makes it look as if there has been some control input.

I spent a few hours on this a few days ago, and I can say with confidence that building a glider big enough to effectively dilute the mass of that cockpit and stable enough to fly straight without control input is a pretty tough task.

[AeroGav]

53 minutes ago, Scarecrow said:

The problem I have is that my design will only fly with SAS on, but it will do so for several minutes completely hands off.  Unfortunately it isn't able to reach it's max LD and minimum sink before the stabilizing message appears, counting up from 1%, which makes it look as if there has been some control input.

Mine continued gently porpoising the whole way down,   at times the glide would be shallower than trend, with airspeed decreasing, and other times steeper with acceleration.   Thus, glide rate as displayed by the plugin did fluctuate,  I think 17:1 was the only ratio displayed for longer than a split second.   "Stabilizing" appears when the airspeed is changing rapidly enough, in other words when the glide had gotten too shallow or the dive too steep, the plugin notices i'm not in a sustainable situation and shows the message.

@herbal space program  My glider consisted of a pair of FAT 455 airliner wings, flipped to sweep forwards (otherwise CoL ends up too far aft with such a short fuselage).   It also had a pair of Big S strakes orientated vertically as the tail fins,  and another pair horizontally ahead of the main wing, to which a Swept Wing Type B is mounted.    I suppose that's quite a lot of wing to support a mk1 cockpit and tailcone  but you could certainly get more extreme than that !

I noticed on Aerodata GUI  2degrees AoA is actually higher than optimal at the very low speeds we're dealing with.  Seems optimal L/D gets less, the slower you go.      Finding the best incidence angle is going to be key to getting a good glide angle, as well as just maximising wing.

[herbal space program]

1 hour ago, AeroGav said:

My glider consisted of a pair of FAT 455 airliner wings, flipped to sweep forwards (otherwise CoL ends up too far aft with such a short fuselage).   It also had a pair of Big S strakes orientated vertically as the tail fins,  and another pair horizontally ahead of the main wing, to which a Swept Wing Type B is mounted.    I suppose that's quite a lot of wing to support a mk1 cockpit and tailcone  but you could certainly get more extreme than that !

I probably had even more wing area on mine than that, but of course I never managed to make one that would fly stably without any control input. I stayed away from the wet wings too because I thought they might have more drag than regular ones. Do you know if that's true or not?

Edited September 18, 2017 by herbal space program

[AeroGav]

7 minutes ago, herbal space program said:

I probably had even more wing area on mine than that, but of course I never managed to make one that would fly stably without any control input. I stayed away from the wet wings too because I thought they might have more drag than regular ones. Do you know if that's true or not?

All wing parts are identical in ksp, i once built a plane with multiple different types of wing part attached to fuselage, enabled the aero data on right click menu, pinned them in place then went flying and took screenshots.  When you did the maths,  all wing parts generated lift and drag at the same ratio (being part of the same airplane, they all had same mach number, air density and aoa) - but it scaled in accordance with lift rating.    The only inferior wing part is the basic swept wing on the aviation tech node, whose lift:rating to mass ratio is half what all the others have.

Control surfaces like elevons have half the lift rating to mass ratio of pure wing surfaces, but still generate lift and drag at same ratio.   Some canards and fins are not 100% control surface, so have slightly more lift rating for their mass than pure/all moving control surfaces.

For my glider  i chose the wet parts because they were the biggest available, and helped keep part count down.  

[PaperAviator]

[Redacted]

Edited September 18, 2017 by PaperAviator

[Scarecrow]

Maybe I should do a video to show that mine isn't porpoising or anything like that, just gliding serenely along and then the stabilizing message pops up.