Glider Design

[Somerled]

I'm having some trouble making a glider, an unpowered plane. I'm making this in the VAB, and I don't know if that creates any issues. All my tests are below 1000m, too. I'm sorry don't have any craft to show since the design keeps changing constantly.

Any design I try with the common technique of putting the center of mass just in front of the center of lift is not working. The planes just nose down no matter how many wings I add, weight I remove, or how close I put the CoM to the CoL. After experimenting, I found that putting the CoM just underneath the CoL at least keeps me steady. It's like the CoL is acting as a kind of pivot, and the CoM just wants to settle underneath it. Putting it below and just behind/forward makes it easier to nose up/down, respectively.

Does this make sense to KSPers? In a powered aircraft you want the CoM, CoL, and CoT all aligned axially. But, without engines to push the craft, does that change the situation? Could there be some problem in the design with weight vs. lift vs. drag?

A side question: any way to turn off/suppress/work around the infinigliding behavior? On my most successful designs I gain 5-10 m/s any time I pull up with the wings on fine controls, making landings very tedious.

Any insight is appreciated.

Edited August 13, 2013 by Somerled

[BryceSchroeder]

I have a lot of trouble building winged things in the VAB. Nowadays I build my shuttles in the SPH using a decoupler as the root part, and stick them on boosters with a sub-assembly manager.

[Bacterius]

I've found pure gliders do not work well in stock KSP. They just seem to fall like a rock without control surfaces and only manage to glide - barely - in the last few metres before crashing. If your design weighs more than a paperplane you also need a lot of wings to support it, which quickly looks ugly and detracts from the point of the exercise (which is to create good-looking planes). I think the main problem is that the aerodynamics model used is flawed and doesn't take aerodynamism at all into account - every part has constant drag regardless of its geometry, meaning that gliding would seem to be impossible as as a glider with no control surfaces and no engines would have no way of maintaining horizontal velocity while killing vertical velocity to air resistance and aerodynamic lift (which is what gliders are supposed to do) because both are equally lost to drag. It appears it's a bit more complicated than that in KSP but yeah, atmospheric drag effects are not very accurately modeled at the moment. But then I kind of suck at spaceplanes, so yeah, maybe I am just really bad at KSP

I think infiniglide is built into the stock game (basically a hack used in place of proper lift forces) but you can try the FAR mod to have some more realistic flying. I haven't tried it myself but apparently it works well.

Edited August 12, 2013 by Bacterius

[Tex_NL]

First of all; to build a proper plane do it in the SPH, NOT in the VAB. Building planes in the SPH is a lot easier among others thanks to mirror symmetry.

Infiniglide is not 'build-in' nor is it a hack. It is a bug. KSP control surfaces don't function on the basis of aerodynamics as they do in real life. To steer they create an invisible thrust vector similarly to RCS. Adding an excessive number of control surfaces results in enough thrust to push your plane forward and even take off. To my knowledge FAR does not change the way control surfaces function.

[Somerled]

(short version: any ideas on how to minimize infinigliding?)

If builds are transferable between SPH and VAB using the sub-assembly manager, then I'll try that next. It's not too hard getting mirrored symmetry in VAB though. It doesn't need to be perfectly symmetric except near stall speeds.

Below are pictures of the most successful glider I have to date. I've added bits and bobs to give it weight (and purpose, it can eject the wings to act as a mini rover). It can take yet more weight using only the two canards as lift. This is tested entirely on Kerbin, the one picture is just from its latest mission to Duna. The small size may be important, i.e. better torque control, but I haven't tried anything larger yet.

It's not difficult to control, it can even recover from stalls and tumbles, but landing is a huge pain because of the infinigliding. Any pitch up starts adding speed. That means no S-turns, no steep climbs (yes, it gains speed in a climb ), and no gliding just above the surface to slowly bleed velocity without very careful controls. It takes very, very long to get down to a safe speed for landing. Any ideas at all about how to stop this madness? Maybe a joystick? If not, I'll have to find ways to retroburn or add drag (chutes being a last resort).

A weird drawback of the low CoM: this baby has no problem flying sideways. That means even the tiniest roll and it's heading off course. Not too great.

I've run into one more funky problem. If the CoM is vertically aligned with the CoL, the control surfaces act very weird. Sometimes they'll turn opposite directions when trying to pitch. Sometimes one or both will twitch violently for no reason. So, setting CoM slightly forward/backward of CoL is necessary in this setup.

And here it is in the VAB. You can just barely see that the CoL is above and slightly forward of the CoM.

[antbin]

If forward/back pitch is your problem, and if you're designing gliders with more than one wing (rather than just canards), you could try 'baking in' some lateral stability.

Do that by tilting the wing at the front of your aircraft 1 small notch 'lead-edge-up'. Alternatively, you can tilt a tail / control surface at the back of your plane 1 small notch 'trailing-edge-up'. Then move the wings / body to match your CoL and CoM as usual (feel free to put the CoM a little *behind* the CoL, even!).

What this does is make the front leading wing 'stall out' sooner than the rest of the wings. This provides some restoring forces so that your glider will settle eventually at *some* angle of attack. Tweaking the location of the CoM can fine-tune the angle of attack of your choice.

[Somerled]

antbin, does that require the adjusted wing to be entirely or mostly forward of the rest? I haven't tried multiple wings along the axis yet (it's just not long enough). Honestly, I forgot how useful it is to have a forward pair. Maybe not having control surfaces balanced in front and back of the CoM is problematic.

[Anton P. Nym]

I don't know if it'll help, but here's a glider I made with ions for "kicker" engines that right now gets a ~30:1 glide ratio. It's slightly nose-heavy, but that's easily compensated for by using SAS.

It probably could've gotten further but it, er, got dark about the time I got to 800mASL so I aborted the test. Haven't gone back to it since.

-- Steve

Edited August 12, 2013 by Anton P. Nym

[Dispatcher]

Are you using your glider as the final stage of a rocket or are you just jettisoning your glider from a guided rocket? The game may require a command part to be incorporated into the glider in order for meaningful physics to be applied to it. I hope that made sense.

[bsalis]

As others have said, best use the SPH to build a glider, or any plane for that matter.

Be mindful that landing gear does not have any mass when flown, but the SPH includes it in the CoM marker.

Which is why I removed the landing gear to take this pic of a working glider...

So yes, I have found that for a stable glider with no trim, you should have the CoL slightly in front of the CoM. However, if you have something you want to bring down from orbit, you probably still want to have the CoL slightly behind, then use trim to pitch up.

You also want some of this for full hands-free flight...

http://en.wikipedia.org/wiki/Dihedral_(aircraft)

Then you can have some fun chasing your glider in another plane...

[Somerled]

Okay. Thanks to all of the very helpful and intelligent people who have posted here, I've improved my understanding and fixed my designs.

Having wings supplying lift both forward and behind the CoM fixed most controllability issues. All my pitch stability problems seemed to be caused by a confluence of poor balancing of lift vectors (and not just cumulative lift) and how KSP treats control surfaces. Oh, and a little research lead me to discover that in aerodynamics a "canard" is actually a small, forward wing. Neat!

For future aeronauticists:

http://en.wikipedia.org/wiki/Longitudinal_static_stability

http://en.wikipedia.org/wiki/Canard_(aeronautics)

...and all the posts above, which are top-notch.

[Brevillance]

im not sure if itll work, but could you not use a tank of RCS, a RCS port and use the counter thrust to bring the plane down for a landing? Given that the size may be small, the thrust/weight will be good and wont need a lot of counter thrust so the RCS may last a while. It could be used to bleed the speed and bring it in for a landing, without needing to touch the controls.

Oh, and also couldn't you use a action group to shut off wing controls (using the torque from the command portion, wether pod or probe) to twist it around, as well as RCS. At max on a small plane 2 RCS ports with some RCS fuel can give you every direction in thrust, and turning off the ability to use the canards would counter the infini-glide bug. wouldn't it?