Aircraft Design: Help Needed

[Diche Bach]

ADDIT: see a couple posts down for a list of my mods (screen cap only, sorry) and also a bunch of screen caps.

I have found aircraft in the game to be "touchy" at best. It took me scores of tests to finally get a basic aircraft that was stable on takeoff and which I can land on ocean no problem and have landed on land a few times. I accidentally overwrote this ship file in my save directory, but I did have a backup of the directory from a few days ago; in examining the dates, I _think_ this is the old model (the one that works reasonably well). I'll check in game in a bit and edit this if I'm wrong. ADDIT: yeah this seems to be the "good" one that works fairly well, though weird that the flaps need to be "inverted" not "normal" in order to produce the dramatic lift for a slow takeoff . . .

Seagull 104

I managed to fly this baby half way round Kerbin to the pyramid, land it in the desert, and then taxi up till the terrain got too rough only to realize: hatch was blocked (which strangely does NOT seem to be the case in this specific file as I just did a test flight and Jeb managed to jump out just past the runway! )

So I did a redesign, and figured I'd "improve" it a bit (more delta-v, etc.). When I was done and it had access ladders it was once again, a disaster. So I kept fiddling with it, and below is about the best I could come up with, but it is still "broken."

Seagull 104a

My questions:

1. What is wrong with Seagull 104a that causes it to swerve on the runway terribly (and exhibit bad lift)?

2. What is wrong with both craft that causes them to pull to port (left)?

3. Can someone point me at a resource to help me understand all the variables and values in those craft files? (I can see that a lot of problems can probably be solved by achieving perfect numbers in the craft file that are tricky to do in the in-game GUI build pane)

4. Any general suggestions?

5. Any way to set up roll trim and pitch trim tabs on the ailerons and elevators?

My understanding of aerodynamics is pretty rudimentary, but I find KSPs handling of things not that edifying. For example control surfaces, angle of attack, lift COMMENTS/ADDITIONAL QUESTIONS:

A. In order to achieve stable flight I found it necessary to use the wing configuration in both of these versions of Seagull: a much oversized swept wing attached near the belly of the fuselage and with the fuselage tilted anti-normal relative to the wing. This to me is a weird design (though I guess it does reflect that in modern real world craft like the stealth bomber?) and it is also weird that it seems to be the only wing configuration that I can get to be relatively stable in flight and landing.

B. Along with the above, I found it necessary to use "flaps" (the control surfaces that have all axes set to inactive and with start deployed "inverted" at 38% which are proximal to the fuselage in both the designs above). Any slight change to these flaps has dramatic consequences, though with the values right around where they are in Seagull 104, it seems to provide _roughly_ the sort of flight I'd expect: general stability in roll and yaw, slow translations, a tendency to lift even at low speeds, and the ability to descend even with nose pitched up a few degrees (you know, the way you LAND a freacking airplane!?! EVERY other design [including the stock designs] wants to go NOSE DOWN once a certain low speed threshold is crossed on approach for landing]).

C. This might be addressed by question #2 above, but . . . in both of these craft, engaging the elevators for pitching normal causes roll.

In closing: I sense that, the way flight is modeled in KSP is actually good and exceptional in that it allows a player to DESIGN aircraft (only flight game I know of that does that), but without a more precise understanding of HOW to design, much less fly airplanes, I think I'm at my limit of how much I can gain from ad hoc experimenting, and I think I need guidance.

Edited August 5, 2016 by Diche Bach
clarification

[The_Rocketeer]

@Diche Bach, this is somewhat critical but please take this as constructive criticism.

1. this is in the wrong subforum, it should be in Gameplay Questions and Tutorials
2. pastebin is a really inconvenient way to share .craft files, there are plenty of others that are more convenient and free
3. you built your planes using mods but you didn't tell us which ones, so effectively nobody can test them for you
4. even a single picture in your top post would probably get a lot of casual advice that could point you in the right direction

I did download both planes into .txt files, save as .craft files, paste into KSP saves folder and launch the game to have a look, but your undeclared use of mods after all that is a real kick in the teeth.

 

[Red Iron Crown]

Moved to Gameplay Questions.

[Diche Bach]

Thanks Red Iron Crown for moving to correct sub-forum.

Thanks The_Rocketeer for taking time to help. Sorry about the lack of info , let me see if I can rectify those problems . . .

My mods (best way to list them I know of, since CKAN cannot list them unless you installed everything with it as far as I know)

I've noticed a lot of serious performance issues (gigantic RAM use by the game), so I'm betting I've got some mods that are improperly installed else fighting one another somehow . . . a separate issue, but I might be opening that can of worms . . .

If anyone can suggest a better way to list mods than a screen cap of the directory (and without reinstalling everything with CKAN, which I do have installed, but which I found to be less than helpful).

A few screen caps of the "Seagull 104" (the one that is a basically functional design that only suffers from the pull to port and the rolling when elevators are engaged):

Seagull 104 Nose

Seagull 104 Port/Belly Oblique view

Seagull 104 superior plan view

Seagull 104 Port cross-section with Aero centers overlay

Seagull 104 Aft with grid and Aero centers overlay

The control surfaces:

Wings are set to MM symmetry. The surfaces proximal to the fuselage (middle lower selection pane in image) are acting as flaps. The ones near the wingtips (far left selection pane in image) ailerons, then of course the rudder (middle top selection pane) and the elevators (far right selection pane in image). on the tail.

And finally the tail with detail on the engine (specs are visible in all the pics via Kerbal Engineer, etc.)

The engine is the 0.625m MRS engine tweaked up to 0.875m.

[Diche Bach]

You know what . . . I examining those images more closely (the aft with grid in particular) I think I see what _might_ be causing the pull to port: the forward gear seems to be shifted ever so slightly to starboard.

Still, this pull problem occurs both on land and in air, so I'm not sure that is it. I appreciate any feedback or suggestions.

[The_Rocketeer]

@Diche Bach wow that's a lot of mods... I'm afraid that's too many for me to install just to test this for you.

Looking at your pictures, I agree that the nosewheel looks a touch off-line - I guess u accidentally attached it to one of the wings instead of the fuselage. I also note that none of your gears is attached in a true 0 position. Any angle of any kind can have undesirable effects, so I would make sure that your gears all point straight forwards and deploy straight down without any angle at all. In my experience, using any of the widgets on gears is basically asking for issues.

Veering on the runway can be caused by a lot of other things, including bugs. I would recommend swapping the LY-10s for LY-50s (if you can) which seem a lot less susceptible to such issues. Another advantage is you won't need (tho tbh u probably don't anyway) to double up on the rear wheels - 3 gears should be enough.

Another tip for the gears - attach them under the wings instead of on top. If u need to move them higher, move the whole wing - clipping thru the wing when the gear lowers can cause phantom forces.

Regarding flap deployment direction, I would try removing them and placing the same part (flipped and mirrored) on the opposite wing. However it could be another issue, or maybe something to do with the game's auto-detection of where the flap is being attached?

You can trim your aircraft's pitch, roll and yaw with alt-wsadqe during flight, tho precision-trimming is tricky and SAS overrides it anyway. An alternative would be to use a mod like Pilot Assistant which does a great job of trimming your plane and holding or changing course for you (but if forgotten/misused may crash your plane).

Other things to try:

Strapping the wings with Struts. On larger aircraft with multi-part wings I've witnessed strange deflections during phys-warp, which to me suggests that mirroring is not truly symmetrical. Anyway, moar struts helps stop this happening.

Giving the wings a little attack- or anhedral/dihedral-angle might also help null out any imbalance. Generally it's better to angle the wings than angle the whole plane.



 

Also if u can narrow down the specific parts mods that u've used in that craft I'd be willing to try it out for u, but I don't know what I'm seeing in your pictures.

[Alshain]

For starters, do not clip any part of a landing gear through any other part, even the wings (some of the housing will naturally clip through, that's ok, but you seem to have the whole housing moved up through the wing).  This is problematic in 1.1, it may or may not fix the problem of veering off the runway, but you will not be able to fix it as long as they are clipped.

Wheels in 1.1 are very specific to their contact area.  Don't put them on backward, don't angle them, don't clip them.

Are the wings angled downward from the fuselage?  That creates downforce, it puts air over the top of the wing and kills your lift.

To answer the question, the trim can not be set in the SPH.  KSP's trim implementation is lacking, it can't even be assigned to a flight stick.  It's really bad.  You can adjust the deployment angle but that affects everything.

Edited August 5, 2016 by Alshain

[Diche Bach]

Thanks guys! I'll have a look at what parts I'm using in that plane and list the mods involved. I'm a bit of a mod whore

Off the top of my head, the mods used in that plane that might well be causing issues:

1. Tweakscale

2. KW Rocketry (the fuel to weight ratios are slightly better) I'd have to question if this is it though.

. . . . hmmm, that is all I can think of . . . not sure if the radial attached air intake is stock or not . . . there are several science parts on it, as well as some KIS stuff, but I don't think any of that should matter.

I figured the heavier landing gear would improve things but don't yet have that teched.

When you guys say "no clipping" on the gear, what do you mean? Do you mean 1. the top of the gear housing should sit RIGHT under the wing, just barely touching the underside surface of the wing? Or do you mean, 2. the top of the housing should at worst, be completely enclosed inside the wing? If the former I'm skeptical (the stock velocitize design has the them clipped in both ways if I recall) the latter I can believe is an issue, so I hope that is what you are talking about.

Also, when you say "no angle" what do you mean by that? Reason I ask is, the gear themselves do not form a 90-degree when they are fully deployed, so which plane should be parallel to the horizontal? Or for that matter, should it be that the gear is parallel to the surface it is attached to? Or, is it that the actual strut should be perfectly vertical right at the point where it intersects the wheel bogie? I figure it is the latter but be good to clarify.

Anyway, I really appreciate your feedback. Now that we've identified it, I'm almost positive the forward gear is not quite right so I'll fix that and see what it does.

ADDIT: re the issue of the angle of attack . . . honestly it is not something I feel I have masterful understanding of in real life, so add in that this is a game modeling that dynamic and I honestly am pretty much lost. I know that in general the point is to get air traveling more rapidly under the wing than over, but I was under the impression that a sizeable portion of that dynamic in real life comes from aerofoil cross-section . . . whereas in KSP our options seem to be: angle of the wing.

I came up with this design (the relatively massive  "wing sled" with the fuselage sitting on top and fuselage tilted) through trial and error. Like I said, every other design I've tried (including the stock ones like Velocitize) exhibit horrific nose dive (stall speed I guess that would correspond to?) when I attempt to land them and a certain low speed threshold is crossed. This one doesn't suffer that issue: it practically just glides in for a landing and one almost has to point the nose DOWN and kill the thrust to GET it to drop at all!

All of this would be much less of a issue if the stock designs in the game files were better performers. Based on my limited understanding of aerodynamics, and what I've gathered from trying my own designs as well as a couple of the stock designs, and then glancing at the rest of them, every single one of the stock designs is laughably bad. Some of them might look like real aircraft, but none of them function worth a damn.

ADDIT: and one last comment for this post: I wonder if my horizontal stabilizers are a bit undersized relative to the wings and that might be accounting for the roll during pitch up maneuvers?

It was so painful just to get to this design (which is functional, but not yet 'perfect') that I've been loath to ask questions or to try too many different things in game. But stretched out over a longer real life time span, I can see how fiddling with this stuff is very engrossing and look forward to trying more variations.

Edited August 5, 2016 by Diche Bach

[Alshain]

1 hour ago, Diche Bach said:

When you guys say "no clipping" on the gear, what do you mean? Do you mean 1. the top of the gear housing should sit RIGHT under the wing, just barely touching the underside surface of the wing? Or do you mean, 2. the top of the housing should at worst, be completely enclosed inside the wing? If the former I'm skeptical (the stock velocitize design has the them clipped in both ways if I recall) the latter I can believe is an issue, so I hope that is what you are talking about.

Attach it however it attaches native.  Do not move it.  You can rotate it as needed but do not move it inside the wing any further than it does when you place it, and definitely do not place it on the fuselage and then put wings that clip into it.

1 hour ago, Diche Bach said:

Also, when you say "no angle" what do you mean by that? Reason I ask is, the gear themselves do not form a 90-degree when they are fully deployed, so which plane should be parallel to the horizontal? Or for that matter, should it be that the gear is parallel to the surface it is attached to? Or, is it that the actual strut should be perfectly vertical right at the point where it intersects the wheel bogie? I figure it is the latter but be good to clarify.

The angle I was referring to was on the wing itself.  It's mostly a flight issue, the crafts heading relative the artificial horizon on the navball is based on the cockpit position.  So on your plane design if the your heading sits at a perfect 0 degrees, the wings will be angled down, this will remove lift and the plane will descend as if the nose were pointed down.

As for wheels, this is one of the biggest issues in 1.1, which has broken wheels.  The game only registers a specific point on collide.  Wheels in this version are only a visual illusion, in fact there are no wheels as you know them.  If you slightly turn them at any angle to the collider, the wheel is no longer making contact with the ground.  The angle built into the model is insignificant, the wheel colliers have been placed to account for that, just don't rotate it yourself and if it is rotated due to it's parent part, try your best to rotate it back.

[bewing]

OK, you've posted a lot of good questions -- so let me take my shot at some answers.

9 hours ago, Diche Bach said:

1. What is wrong with Seagull 104a that causes it to swerve on the runway terribly

Generally, this sort of veering has two causes. One is the wheels not facing perfectly straight forwards. The second is "too much friction in front of the CoM and not enough behind the CoM".

To fix the first, you use "Absolute Rotation Mode". You click the "Rotate" gizmo, then hit F. You only have to do this once. Select "Angle Snap Mode" also. Then you click on a wheel, rotate it one click sideways, and rotate it one click back. Then click "Place". The wheel is now perfectly straight.

To fix the friction problem: Lower the friction on your front landing gear a lot (below .4 -- and the more your plane veers, the lower you should set this value). On your rear wheels: move them toward the back of the plane a bit, and set their friction much higher (1.4 at least, and higher than that is fine).

9 hours ago, Diche Bach said:

though weird that the flaps need to be "inverted" not "normal" in order to produce the dramatic lift for a slow takeoff . . .

Control surfaces that are ahead of the CoM work inverted from ones that are behind the CoM. So the game always has to guess whether you intend the surface to work inverted or not. If the center of a wing is attached ahead of the CoM, but the wing is swept so that the control surfaces are behind the CoM, the game can sometimes guess wrong. Then you have to fix it by hand.

9 hours ago, Diche Bach said:

2. What is wrong with both craft that causes them to pull to port (left)?

Very hard to say. Could be those fuel ducts. They happen to cause excessive drag in this version, and they are completely unnecessary. Try removing them and see if that changes things.

General notes on stability:

Having your CoL above your CoM is more stable than either having them inline or having it the other way (like you do). The way you have it makes your plane very unstable in the roll direction, which is probably why you are experiencing roll when you pitch.

Your horizontal stabilizers are probably too small. If you switch them for Tail Fins, things might work better.

You are turning on SAS to help you fly these planes, right?

9 hours ago, Diche Bach said:

EVERY other design [including the stock designs] wants to go NOSE DOWN once a certain low speed threshold is crossed on approach for landing]).

Yup, that's called "stalling". You need to land going faster than that.

 

I understand that you've found all this trial and error to be painful and frustrating. One thing I can say is that canard wing planes are much easier to design and fly. You may want to try a few of those. And the forums here are full of example planes that fly quite well -- since you don't like the stock ones (for good reason).

Edited August 5, 2016 by bewing

[Diche Bach]

Thanks Bewing! Some good stuff.

Using Alshain's comments about the wheels, I made a new one and did as he said: did not move the wheels up from where they natively wanted to lock into place. That seemed to help tremendously on the swerve on the runway.

So the fuel ducts, I was thinking that was a way to insure that the fuel in the front of the plane emptied into the back, so that the plane's CoM wouldn't get further forward as it flew making landing (and nose diving) more difficult. If I were to clip fuel tanks mounted on externally inside the fuselage a bit and then link forward fuel tanks to rear with the ducts inside the fuselage, would that be a good practice or as you say, just unnecessary?

So what is a reasonable ballpark for a landing speed? That Seagull I can land at like 45m/s which is nice, but slow to descend.

[bewing]

35 m/s to 60 m/s is generally safe and effective for landing. If you have a hell of a lot of wing on your plane, you can get it below 20 m/s. If your plane is really heavy, then you may need to land faster.

Fuel ducts are for rocket engines. Jet engines can draw fuel from anywhere on your plane. In ver 1.1.3, with rocket engines, there are many orientations of tanks that a rocket cannot draw fuel from. If you want to draw fuel from them anyway, then you put a fuel duct to a part that the rocket can draw fuel from.

 

[Alshain]

One more thing to add, 1.2 is supposed to fix a lot of the wheel issues.  You are probably running into issues that aren't normal plane construction issues (such as the above mentioned clipping and collider position).  Hopefully this goes away after the next patch, the wheels are very buggy at the moment.

You can also deploy those flaps on landing to stay in the air at slower speed, how much depends on the plane.  Not really necessary on the KSC runway, but it helps if you are shooting for the much shorter island runway.

Edited August 6, 2016 by Alshain

[GoSlash27]

Dichie Bach,

 It looks to me like the leading edge of your wing is lower than the trailing edge when at rest; "negative static incidence". This is going to press the aircraft down into the runway as speed increases. That, in turn, will bow your wings and put your main gear out of alignment and in addition force you to have a higher takeoff speed/ longer takeoff run... which compounds the problem.

 I recommend clipping your wings higher into the fuselage and rotating them to give positive incidence. With all that wing area, you should be able to float off the runway at a very low speed. You should be airborne before the gear has a chance to give you problems.

 Best,
-Slashy

 

[Diche Bach]

Wow, you guys are awesome! Thanks to your feedback and suggestions, I've managed to get a plane that I find more than just "functional;" I might even say its performance is 'good' and it certainly seems on a par with how planes in other flight sims I've played handle.

No time to go into details now, too giddy with having a basic understanding of this stuff and too many kerbin based missions to complete! But here is one screen shot of my new Seagull 104-A in flight.

Happy to provide more screen caps and/or description of what I did differently if anyone (perhaps a fellow beginner like me) is interested; and of course, I already have still more questions that I'll want to pose eventually. But for now gotta play with this baby!

Thanks folks!

Edited August 6, 2016 by Diche Bach

[swjr-swis]

On 8/5/2016 at 4:29 PM, Diche Bach said:

If anyone can suggest a better way to list mods than a screen cap of the directory (and without reinstalling everything with CKAN, which I do have installed, but which I found to be less than helpful).

A pitch for a site I really like: if you upload your craft to KerbalX.com, the site has code that automatically recognises mods used in the craft and lists them all (and the ones that it doesn't, you can help with to get them listed as well). As a second benefit, it's a very nice site by a good maintainer specifically for showcasing and sharing your craft files.

 

23 hours ago, The_Rocketeer said:

I would make sure that your gears all point straight forwards and deploy straight down without any angle at all. In my experience, using any of the widgets on gears is basically asking for issues.

The '0' position for the gear leaves a slight angle forwards for the gear. And many parts will actually cause the gear to not be in zero angle position, because the game editor always assigns the first anglerelative to the part you are attaching to. So it's good to remove any rotation by placing the gear first, selecting the rotate gizmo, and setting it to absolute mode (which is not the default), then it's usually just a tiny tick to zero out any angle it 'inherited' from its parent part.

In my opinion, 'absolute' should be the default mode for the rotate gizmo, like it already is for the offset one, essentially because there's no real consistency in what the editor sometimes decides to assign angle-wise.

 

20 hours ago, Alshain said:

The game only registers a specific point on collide.

Agreed on the rest of your explanation, but not on this particular point. In my experience, it appears to be a 90 (effectively more like 88) degree angle pointing straight 'down' from the perspective of the wheel. So as long as any point within 45 degrees forward/backwar is visibly touching the ground, it still works, but beyond that its like the wheel is not even there.

 

16 hours ago, Diche Bach said:

So the fuel ducts, I was thinking that was a way to insure that the fuel in the front of the plane emptied into the back, so that the plane's CoM wouldn't get further forward as it flew making landing (and nose diving) more difficult. If I were to clip fuel tanks mounted on externally inside the fuselage a bit and then link forward fuel tanks to rear with the ducts inside the fuselage, would that be a good practice or as you say, just unnecessary?

In general, try to design your plane such that fuel is distributed -and subsequently used- equally from behind the CoM as from the front. Turn the CoM indicator on, and start playing with the fuel quantity sliders of your tanks; the less the CoM moves when you do that, the better.

Easier said than done because engines tend to contribute way more mass than cockpits, but a small little trick that may serve you: if you attach a small fuel or ore tank as far front of the front of the plane as possible (the Oscar or the radial ore tanks work well, and they can usually be clipped entirely into most cockpits), you can lock their resources and then use them just as balancing 'ballast' to offset a less than optimal fuel mass (or engine mass) distribution.

 

17 hours ago, Diche Bach said:

If I were to clip fuel tanks mounted on externally inside the fuselage a bit and then link forward fuel tanks to rear with the ducts inside the fuselage, would that be a good practice or as you say, just unnecessary?

In general fuel ducts should be unnecessary for jet engines, and would only add a rather disproportionate amount of drag. There are times when the game decides an engine cannot draw from certain tanks, but this is rare and tends to be an indication that a rethinking of the design is necessary anyway.

 

I've not been able to test your craft due to mods, and for some reason I'm not getting the images you posted, so can't give you any advice based on those yet. If I manage to make it work I'll revisit this.

 

[Diche Bach]

23 hours ago, swjr-swis said:

A pitch for a site I really like: if you upload your craft to KerbalX.com, the site has code that automatically recognises mods used in the craft and lists them all (and the ones that it doesn't, you can help with to get them listed as well). As a second benefit, it's a very nice site by a good maintainer specifically for showcasing and sharing your craft files.

Thanks swjr! That is an excellent site. I uploaded it and will upload the "more successful" version I built thanks to this thread next.

https://kerbalx.com/Diche_Bach/Seagull-104

ADDIT: here is the one I built with the help you guys provided.

https://kerbalx.com/Diche_Bach/Seagull-104-AB

 

Edited August 7, 2016 by Diche Bach
ADDIT

[DaMachinator]

Protip: RCS Build Aid adds CoM markers for tanks half full and tanks empty. It is usually not very difficult to balance your fuel tanks and engines in such a way that the CoM does not move much as tanks empty.

Also, this tutorial by @keptin is very very nice.

[Diche Bach]

On 8/11/2016 at 8:12 AM, DaMachinator said:

Protip: RCS Build Aid adds CoM markers for tanks half full and tanks empty. It is usually not very difficult to balance your fuel tanks and engines in such a way that the CoM does not move much as tanks empty.

Also, this tutorial by @keptin is very very nice.

oh hell yess, that tutorial is fantastic! Will have to revisit that when I'm not on a: Mun/Minmus mission kick.

[Alshain]

3 hours ago, Diche Bach said:

oh hell yess, that tutorial is fantastic! Will have to revisit that when I'm not on a: Mun/Minmus mission kick.

Be aware that tutorial is old, and the parts about KSP's Drag Model are now obsolete.  In fact KSP's Drag Model is now more like FAR was back when that tutorial was made.

So fairings and nose cones, swept wings and angle of incidence all have effect in stock now.

Edited August 14, 2016 by Alshain