@#$&*(^ airplanes -- how do they work?

[Catbus]

To begin with: Yes, I've read this. And read it again. And again. And again. And then I build planes that, according to that guide, should fly. And they don't. They veer off the runway (despite being perfectly symmetrical), or nosedive into the sea, or suddenly climb at 60 degrees and flame out, or fly straight but won't turn, or fly straight and (eventually) turn but lawn-dart into the ground when I try to land them.

ARRRRRGH.

On the premise that every unlockable aeronautic tech in the game ought to allow me to produce a workable plane, I've been going through each one, one tech level at a time, trying my darnedest to make a plane I can take off in, fly around and land. After several hours of struggle on the second level yesterday (having already fudged the first -- I never did successfully land my Aviation-only plane), I finally gave up.

Here's one thing I'm sure is contributing to the difficulty: I think any amount of fuel in the main fuselage makes an Aviation-only plane too heavy to take off. The plane I finally succeeded with carries fuel only in its two wing-mounted Mk 0 tanks. But I can't be certain, because unlike the handy-dandy rocket formula (Δv = ln (M_start/M_end) · I_sp · 9.81m/s²), which I've used successfully to choose the right engines and amounts of fuel to send craft all over the Kerbolar System, I don't know of any formula that tells you how much thrust and wing area you need to lift a plane of a certain mass.

Is there any way, any way at all, to design a decent plane, other than changing a bit here, a bob there, then crashing them into the ground again and again and again for hours at a time?

I just want to make planes that work [collapses into fetal position, sobbing].

[steuben]

Pics, and maybe some craft files, would help us, help you, help us all.

The stock craft in the came are a good starting point as well.

Don't worry, it's like docking. Once you get it, it's easy to do. But, till then it is a bit of a slog.

Edited July 24, 2017 by steuben

[Catbus]

Forgive my density, but . . . craft files? Stock craft? I don't know where the former are stored, and I didn't know that the latter exist.

[Catbus]

OK, I managed to answer the first question. Here's my most successful recent Aviation-only plane: https://www.dropbox.com/s/eeo5hv89wlds9tm/Konkorde Mk1.craft

It takes off only at the end of the runway, never before. It handles a little too responsively (enviable yaw, though), and I can't get it to slow down enough to land without faceplanting before I reach the runway.

[Rocket In My Pocket]

15 minutes ago, Catbus said:

OK, I managed to answer the first question. Here's my most successful recent Aviation-only plane: https://www.dropbox.com/s/eeo5hv89wlds9tm/Konkorde Mk1.craft

It takes off only at the end of the runway, never before. It handles a little too responsively (enviable yaw, though), and I can't get it to slow down enough to land without faceplanting before I reach the runway.

If you are having trouble taking off, that is usually due to your rear set of landing gear being too far back, make sure it's directly behind your COM so that the tail of the plane has room to rotate downward so the nose can then rotate upwards.

Control sensitivity can be solved a couple of ways, by pressing Caps Lock to enable "soft mode", or by right clicking your control surfaces and lowering the max deflection some.

Slowing down when landing is another issue entirely, if you don't have airbrakes unlocked make your own with two opposing sets of control surfaces that open together to catch the air.

Edited July 24, 2017 by Rocket In My Pocket

[Rocket In My Pocket]

So I loaded your plane up and the first thing I noticed is that your COL is way behind your COM, you had the right idea but you want them a lot closer than that, like inside each other so the crosses are touching.

The rear landing gear are very high up, but once again you had the right idea, getting the nose elevated; you just went too far with it.

I also notice you have your pitch control on the inner wings instead of on tail fins at the rear? You'll have much better pitch control and an easier time taking off with a traditional tail setup. Putting the control surfaces way out there gives them more of a lever action.

Overall not bad, when it actually takes off it does indeed fly, although not terribly well, lol.

Here's a quick edit of your plane: https://www.dropbox.com/s/y86110fi5fvdljh/Konkorde Mk2.craft?dl=0

I added landing flaps tied to the brakes, a tail fin, lowered the rear landing gear, moved the engines to shift COM back, adjusted wing position to balance COL/COM, etc...(I wanted to add more wing area in general but didn't want to change your design too much, but as an added note, future planes should have larger/more wings.) should give you a good idea of what you were missing. Best of luck!

Edited July 24, 2017 by Rocket In My Pocket

[Spricigo]

Don't feel bad.  There is so many ways to build a plane wrong,  you need a fair amount of persistence (or luck)  to finally find the some way to build it right. 

3 hours ago, Catbus said:

 Is there any way, any way at all, to design a decent plane, other than changing a bit here, a bob there, then crashing them into the ground again and again and again for hours at a time?

How cute... It took me more than a week to build my first "aircraft" that "flew"  well enough to crash in the ocean. (let alone the time before that where I didn't dare to open SPH).  

The problem is that without guidance or experience you don't know if that bit there will make things better or worse.  And given there is more ways to make things worse.... 

If your immediate goal is to have a aircraft that can fly around and complete some mission, you can take a look at KerbalX or the spacecraft exchange and look for something that fit your needs

<self-promotion>KarlJatho  this one is easy to fly,  low tech,  with science equipment and space for 2 crew (pilot & scientist) </self-promotion>

If you really want know what is wrong with your current design,  post pictures. Picture of the craft in the SPH with CoM and CoL.I displayed.  Also a brief description of the intended purpose (e.g. Collect science [flying low]  and [landed] around KSC). 

Craft file may be useful,  but it requires people to download it and fire up KSP to provide advice.  Experienced players may spot novice mistakes from a good picture while not able to open KSC. 

Don't feel bad.  There is so many ways to build a plane wrong,  you need a fair amount of persistence (or luck)  to finally find the some way to build it right. 

3 hours ago, Catbus said:

 Is there any way, any way at all, to design a decent plane, other than changing a bit here, a bob there, then crashing them into the ground again and again and again for hours at a time?

How cute... It took me more than a week to build my first "aircraft" that "flew"  well enough to crash in the ocean. (let alone the time before that where I didn't dare to open SPH).  

The problem is that without guidance or experience you don't know if that bit there will make things better or worse.  And given there is more ways to make things worse.... 

If your immediate goal is to have a aircraft that can fly around and complete some mission, you can take a look at KerbalX or the spacecraft exchange and look for something that fit your needs

<self-promotion>KarlJatho  this one is easy to fly,  low TECH,  carries science gears and have space for 2 crew (pilot & scientist) </self-promotion>

BTW : for help with a particular craft design, pictures are often more useful than craft files.  Pictures are more accessible (no need to fire up KSC)  and experienced player can often easily spot issue that are elusive for novices. 

[bewing]

As far as runway veering goes, there are many possible issues, and many solutions. That question is even in the FAQ on this forum. I personally prefer to modify the friction settings on the wheels to fix it.

 

It is relatively easy to make a plane with basic aero tech that flies with 200 units of fuel. And they actually fly rather well. This picture is one of my early examples -- it is not very efficient.

[Kerbal101]

@Catbus LY-01 gear is for me personally so bad, that I even wrote a minipatch to make it as stiff as LY-05, because its impossible for me to enjoy..
Stock LY-01, while realistic ("kinda").. has a 3t-per-wheel max weight and approx. max 2-5m/s vertical speed on landing requirement.. but has been by far the most unreliable plane gear, at least for me.

If your issue is not LY-01, then you have just to remove more fuel.
If the plane does not handle right, make sure that its nose is a bit up by adjusting back/front gear position.
If it flips after taking off - keep the main wings near center of mass, and center of lift in the middle - or slightly behind of the center of mass to trade maneuverability for stability (regardless how full the tanks end up), like a "dart".
Since planes work essentially by putting their weight on the main wings - and balancing their weight around them,... like on "swing", so main wings don't even need elevons there, and fins/tail/elevons/canards work best - the further they are away from the "swing".

Edit: you could also possibly overlooked requirement to dedicate the yaw/pitch/bank roles on each plane control surfaces. Or perhaps not activating SAS.
As of TWR calculation, I use KER to get the numbers (for Kerbin only) - subsonic's are 0.3-0.6, supersonics - 0.8-1+.

Edited July 25, 2017 by Kerbal101

[GDJ]

4 hours ago, Catbus said:

I just want to make planes that work [collapses into fetal position, sobbing].

I just want to cry with you. 
Well, maybe not.

I won't bore you with the minutia details since you probably read alot of tutorials already, but a quick and dirty way to get something off the ground:

-Keep your COL behind the dead centre of the COM, but keep the blue sphere inside the yellow one
-build it with your fuel tanks empty, then fill them up. Look at the COL/COM positions. Sometimes it's best to mount fuel tanks in the wings, in line with the COM. That way nothing changes as the tanks run dry.
-Nothing wrong with having too much lift. Better to have too much and go slow than having too little and trying to do Mach on the runway just to take off.
-Sometimes having the COT (centre of thrust) slightly below the COM helps you get off the ground. Don't go nuts, just a tiny bit.
-Keep it simple.
-Double check the actions of the flight surfaces and flaps. The rear ones should handle pitch and yaw only. Front (main) wings should be roll. 

If your plane needs adjusting, small tiny steps.

Edited July 25, 2017 by GDJ

[Spricigo]

25 minutes ago, GDJ said:

-Nothing wrong with having too much lift. Better to have too much and go slow than having too little and trying to do Mach on the runway just to take off.

There is a big difference between a lesser evil and a good thing.  A slow plane uses more fuel and, much worse,  eat more player time. 

 

[GDJ]

17 minutes ago, Spricigo said:

There is a big difference between a lesser evil and a good thing.  A slow plane uses more fuel and, much worse,  eat more player time. 

You have to learn to crawl before you run. Going slower in a high lift design is far better to learn on than a mach buster.

[Catbus]

4 hours ago, Rocket In My Pocket said:

So I loaded your plane up and the first thing I noticed is that your COL is way behind your COM, you had the right idea but you want them a lot closer than that, like inside each other so the crosses are touching.

The rear landing gear are very high up, but once again you had the right idea, getting the nose elevated; you just went too far with it.

I also notice you have your pitch control on the inner wings instead of on tail fins at the rear? You'll have much better pitch control and an easier time taking off with a traditional tail setup. Putting the control surfaces way out there gives them more of a lever action.

Overall not bad, when it actually takes off it does indeed fly, although not terribly well, lol.

I made my own modifications based on your reply. Behold: https://www.dropbox.com/s/yvzl3j6knadd28k/Konkorde New and Improved Mk1.craft?dl=0

It flies! It even lands! (Though sometimes at the cost of a wing or two. That may be more a reflection on my piloting skills than on my design skills.)

Here's the next one that needs fixing, using parts from the Landing and Aerodynamics techs: https://www.dropbox.com/s/fcuapqk4rtzzoug/Konkorde Mk1A.craft?dl=0

I originally had swept wings on it, but they screwed up the functioning of the elevons, because KSP thought they were in front of the CoM when in fact they were behind it. :-\

Edited July 25, 2017 by Catbus

[Catbus]

And here's the one that works! https://www.dropbox.com/s/u52pz3z0r2qbh92/Konkorde New and Improved Mk1A.craft?dl=0

OK, I think I'm starting to get the hang of this finally. Thanks for the pointers.

Though I'd still like to see an equation for how much lift is needed per ton of mass, or how much fuel it takes to circumnavigate Kerbin . . .

[artwhaley]

Since nobody mentioned it again after you asked, I'll add - 

The game comes with more than a dozen stock craft already built.  I believe there's now a menu setting that controls if they're available to you in a career game, and you'd still be limited by what parts are unlocked, but you can always just start a sandbox save and play with the examples that are magically in the space plane hanger's 'saved craft' folder then!   Doing so will let you learn to fly on known good hardware - because it IS difficult when you're starting out to know 'did I crash because of the build, or because of my flying.'   The answer is usually both - and that if you'd been better at EITHER building or flying you might have not crashed.  But - that's all the more reason to learn to fly so that when you are test piloting your own creations you'll know what it SHOULD have handled like and be able to have a starting place on fixing it.

[Vanamonde]

Quote

Is there any way, any way at all, to design a decent plane, other than changing a bit here, a bob there, then crashing them into the ground again and again and again for hours at a time? 

No.

That is the way to make everything else in KSP, and planes are no different. 

[Spricigo]

2 hours ago, GDJ said:

You have to learn to crawl before you run. Going slower in a high lift design is far better to learn on than a mach buster.

Different people learn in different ways. Based in experience I can say that mach busters with quasi-vestigial wings are pretty effective learning tools. (and fun as... mach buster with quasi-vestigial wings,  I suppose) 

In any case that is not the point.  A plane that fly poorly may be better than one that crash while attempting to take off,  but is still much worse than a balanced design that fly properly. High or low lift don't matter if that is balanced with the other characteristics of the vessel. 

 

 

2 hours ago, GDJ said:

You have to learn to crawl before you run. Going slower in a high lift design is far better to learn on than a mach buster.

Different people learn in different ways. Based in experience I can say that mach busters with quasi-vestigial wings are pretty effective learning tools. (and fun as... mach buster with quasi-vestigial wings,  I suppose) 

In any case that is not the point.  A plane that fly poorly may be better than one that crash while attempting to take off,  but is still much worse than a balanced design that fly properly. High or low lift don't matter if that is balanced with the other characteristics of the vessel. 

 

 

[Rocket In My Pocket]

Glad to hear you are on the right track @Catbus

You'll be building planes like a pro before you know it!

[Catbus]

9 hours ago, Vanamonde said:

No.

That is the way to make everything else in KSP, and planes are no different. 

Totally disagree. As I mentioned, I've been successfully using the rocket equation and delta-v maps to design multistage rockets that go exactly where they need to with just enough fuel to get there, if not on the first try then on the second or (at most) third. (Usually the fault isn't in the basic design but rather is because I forget to put on solar panels or RCS thrusters.) Maybe I'm spoiled by the predictability, but the amount of sheer guesswork involved in designing planes, compared with the near-total absence of guesswork in designing rockets, has been frustrating to me.

Edited July 25, 2017 by Catbus

[Kryten 2X4B 523P]

Theres been some great advice in this thread and you seem to be on the right track now Catbus, which is good.

 

As to equations - I dont know of any myself, but I put some rough figures together for Light Aircraft (i.e. under 5 Ton)  using the Juno engine that I found successful, as follows:

1.6 lift per ton, with a very rough 5 to 1 ratio on lift surfaces between front and rear (and thats including elevons etc)

landing gear - using the LY01 / LY05 with a max of 1.5T per wheel (so if using these, try to stick to 4.5 tons max weight for your plane for max landing gear reliability)

Engines - for the Juno I use 1 engine per 3 tons as a rough guide.

 

After that, the main points as previously mentioned:

COM / COL is in the right spot (COL just behind COM, as a guide try to put the COL right on the edge of the COM to start)

Make sure your wheel positioning is right (and that differs between tricycle and taildragger setups) - assuming tricycle setup start with this lined up with the back of the COM

make sure your wheels are dead straight (rotate in absolute mode)

Mount the wheels on the fuselage, not the wings

Make sure your Thrust is in line with your COM - under will generally push the nose upwards, above will push the nose down (pivoting around the COM)

Tweak your control surface settings as needed

Make sure your jet exhaust isnt blocked / restricted in any way 

Veering - this has a number of causes and can be quite complex to sort out especially on bigger planes, but as a rule, do the above on wheel setup and then set the nose gear friction to half of its starting value (assuming a tricycle setup) . After that you can explore things like fuselage flexing etc

 

It also doesnt help these some issues with low end parts as well (cockpit too heavy, swept wings has issues etc) which makes things at the early stages even more frustrating.

[easytarget]

As a complete newb I took a stab at making a plane last night for the first time and had a great time crashing my creations, one thing I discovered I wasn't expecting is the original airstrip is quite bumpy to the point at least for me of being unusable, so much so that I've abandoned working on planes until I upgrade it.

[Spricigo]

7 minutes ago, easytarget said:

As a complete newb I took a stab at making a plane last night for the first time and had a great time crashing my creations, one thing I discovered I wasn't expecting is the original airstrip is quite bumpy to the point at least for me of being unusable, so much so that I've abandoned working on planes until I upgrade it.

That issue its easy to solve, take off and land in the grass. Much longer and perfectly flat. OTOH the instability you are experimenting may be more about your design than you are aware, wheels in particular are a bit tricky.

[easytarget]

Probably design, like I said, first stab at it.  Still find it ironic kerbals are so challenged in how to properly pave a runway.

[Catbus]

It just costs money they don't have. Remember, they bootstrapped this space program using supplies they found behind a dumpster.

[Rocket In My Pocket]

1 hour ago, easytarget said:

Probably design, like I said, first stab at it.  Still find it ironic kerbals are so challenged in how to properly pave a runway.

You should have seen it before 1.3.

They actually flattened it out considerably.

It used to look like a Mario Kart track.