Transport plane undercarriage collapses

[Hotel26]

I'm a noob in the SPH and working on only my second (2nd) airplane.  This one is a supersonic transporter (named Aquila Ursa)

All was going well with previous development versions including a number of short-field landings and take-offs, which is its primary mission.

Then I extended its body to insert a short cargo bay in which to load the Prospector 1 (stock) rover.  Gross weight is now 75t.

It takes the whole 3km to reach Vr = 120m/sec, but with half tanks it will do it at 40 m/s and climb solidly.  Cruise is Mach 3+ and range is 75% Kerbin circumference.

Problem is that I cannot land it any more without the wings breaking off.  (Maybe I just need to get better at landing but this one is palpably fussy about it compare to my SSTO (which is a good deal lighter)).

I've moved the engine close and equally-spaced to the gear and moved all closer to the wing root.  I tried aligning the main gear as best I could.  I tried using Toggle Snap for alignment as well.

I can accept not being able to land it with full fuel and that's fine.  (Do the mission or don't come back!)  But I want to be able to land it reliably on half tanks or less.

I will really appreciate any help on this because I am so EXCITED: who needs to go into space when there is all of Kerbin to explore!!?

[Hotel26]

(I'm starting to like the idea of just putting a parachute on the rover, dropping the rear cargo door in flight and just pushing it out over the target, James Bond style...)

[bewing]

56 minutes ago, Hotel26 said:

(I'm starting to like the idea of just putting a parachute on the rover, dropping the rear cargo door in flight and just pushing it out over the target, James Bond style...)

If you are flying too fast when you do this, your rover will be autodeleted when you get more than about 20km away from it (if it hasn't landed yet).

Very First impressions on your plane:

It works better to hook your wheels to the fuselage, rather than the wings. Then you use the offset tool to move them to someplace that looks attractive to you. (Hold down Shift, in order to move the parts a long way.)

It also works better to hook your engines to the fuselage, rather than the wings, ditto.
-- Your center of thrust is a bit below your CoM anyway, which can theoretically cause some handling problems.

The Engine Nacelles aren't doing you any good -- you should replace them with fuel tanks, probably.

 

-- I'll do some testing on it now and see what I can see.

Just doing this much tweaking seems to fix the "wings falling off" problem: https://pastebin.com/raw/8PC0P2cZ

I was able to take off, turn around, and land with the fuel it has.

 

 

Edited November 19, 2017 by bewing

[Hotel26]

Wow.  Looking forward to your feedback!

I moved two of the Whiplashes to the fuselage.  I found this counter-intuituve because it is *more* weight on the wing root, but I just tested it with half fuel and landed it first time.  Plus, I have more fuel onboard now.

The twin engines mounted higher on the fuselage resist take-off rotation a tad but it was still close to 40.  And it looks a whole lot more wicked!

Please do continue your analysis, because this is educational!  But thanks already.

 

[Helmetman]

1st of all. A picture of your plane?

I don't know what your level of engineering is. But in the lower left corner of the VAB and SPH are the following 3 buttons.

Center of Mass (CoM)

Center of Thrust (CoT)

Center of Lift (CoL)

Activate these when building your plane. The pics below are not of a plane, I just strapped something together in my own career. You want CoT (pink orb/arrow) to move through the center of weight (CoM)
Sailing boats have there CoT (from the wind) high up against the sails. That's why you have to feather the sails in order not to keel over. Same with rockets and planes, although they use rockets and not sails ofcourse and the term is not to keel over but besides that, kinda similar. This effect is shown on the 2nd pic by offsetting the engine. This is very over the top ofcourse, but sometimes a pixel or two of offset from the CoT to the CoM can in combination with other factors turn into disaster. 

The CoM is the center of weight and thus the pivoting point for the thrust vector and lift itself. The blue orb (CoL) keeps that center of weight (CoM) aloft. Move the CoL to close into the CoM and you'll get stability issues. Unless you fix things with front and rear stabilizers it will be like walking on a ball control wise. It will be a very flippy floppy turn upside down airplane ride.

Move the CoL to far from the CoM in any fwt and aft direction and you get control issues and additional problems when going into high speed/altitude or even suborbital flights.

Good

Bad

2 hours ago, Hotel26 said:

It takes the whole 3km to reach Vr = 120m/sec, but with half tanks it will do it at 40 m/s and climb solidly.  Cruise is Mach 3+ and range is 75% Kerbin circumference.

What is the total mass at launch? What is the total mass with tanks half full? Half the weight or thereabout? 
If you get to 120m/s over the entire course of the runway, you simply don't have the thrust. Add more engines.
With fuel tanks drained your momentum is greater because you have more thrust compared to the total weight.  Factors of thrust against units of mass (TWR) and greater momentum with drained fuel tanks scale up. The result is, a much shorter takeoff roll. 

Those figures aren't that strange to be honest.
What is wrong about KSP are the aerodynamics. They fail, simply put. 120m/s you say? That's like 430 km/h. Isn't this the category where bricks with air pockets start flying?
In ksp you can only do so much. Add moar wings?
Put incidence on the main wing. This allows it to catch more air at the same nose up attitude since incidence on the wing creates a greater angle of attack when pitching up during takeoff. 

2 hours ago, Hotel26 said:

Problem is that I cannot land it any more without the wings breaking off.  (Maybe I just need to get better at landing but this one is palpably fussy about it compare to my SSTO (which is a good deal lighter)).

You can use the strut part. Attach it to the wings from the fuselage to reinforce it. It looks nasty though and struts produce considerable drag. So that's not to kindly for a plane that wants to go fast. There is a other strutting solution.

Go the the main menu > settings > General Tab (default) and check "Advanced tweakables" 
Now go back to the SPH (your plane) and right click any part and use the "autostrut" button.
Generally a wing that is autostrutted to "Grandparent part" or "Root part" will be strong enough. 

Oh, and like is already said. Don't put gear on the wings, only on the fuselage which is sturdier. You can still use the Move tool in the SPH to drag the landing gear to a favourable position, so that they look as if they're attached to the wings while they actually aren't.
 

2 hours ago, Hotel26 said:

I will really appreciate any help on this because I am so EXCITED: who needs to go into space when there is all of Kerbin to explore!!?

There's not much on Kerbin, or any planet for that matter. Unless you like mountain views, find easter eggs or build something like a base on a location of your choosing. I do think you'll find it sporty, challenging and fun to go elsewhere. Getting a plane to work is one thing. If this is a game you keep hanging on to you will most likely do all the other stuff and get your rigs all the way across the solar system.

Have fun flying though  

Edited November 19, 2017 by Helmetman

[Hotel26]

Thanks Helmetman.

The mission profile always starts at either KSC or the Inland KSC where a 3km runway is available.  As long as I get rotation before the end of runway, establish positive climb, make a turn onto heading and reach a 20km cruising altitude within 10 minutes (all accomplished), I am cool.  In addition, the mission objective is to be able to do a short-field landing on possibly uneven/unknown terrain using twin chutes and that actually was easily done because of the chutes (until I added the cargo bay extension).  Then a short-field take-off to get out is required and I've been testing that at KSC on the runway with half fuel and the runway use is much much shorter and the climb is nearly 50 degrees.

Thanks for all the tips, especially about strutting.  I'll try to post more/better photos shortly, but there's one above.

Edited November 19, 2017 by Hotel26

[Rocket In My Pocket]

22 minutes ago, Helmetman said:

There's not much on Kerbin, or any planet for that matter. Unless you like mountain views, find easter eggs or build something like a base on a location of your choosing.

I found a pretty neat canyon with a river flowing through it to fly through one time.

Spoiler

Only pic I have of it sadly, you can just make out the start of it on the left there.

2 hours ago, bewing said:

If you are flying too fast when you do this, your rover will be autodeleted when you get more than about 20km away from it (if it hasn't landed yet).

Very true, I often circle my payload as it descends for this very reason. Most of my drop-transport planes have an action group setting that delpoys the control surfaces in such a way as to put it into a slow, sustainable turn. I can then switch to the payload if I want and ensure it lands safely, before returning to my plane and flying off. 

Edited November 19, 2017 by Rocket In My Pocket

[Urses]

Other problems with taking of on a plane can be the position of landing gears to far behind of the CoM. Your ailerons have then a realy short lever to heave the AoA to enable a ascend.

One way is to make the forward gears be longer that the plane have build in elevation on runway. The other is to bring the back gears as near as posiible at the CoM. But pay attention not to smack your tail in the ground.

For trust allignment i personaly prefer to use "RCSBuildAid"-mode. You have there a "Engine" rotation force marker to allign your CoT to your CoM.

I was wondering on one my designs why it never will take off on runway, but found this way that my engines did a nose down pitch rotation. After a small nudge down on the engines, the plane took off on 35m/s at 1/4 runway.

For the wings problem you had some wise advices. Grandparent autostrut works best there for me. And inbuilld AoA on the wings helps you to use lift of the MK2 and Mk3 body parts with evading to much drag of them and save on wing mass.

One point more for me, for breacking of the wings on landings was a to weak landing gear. It looked stable but translated to much force to other parts. As i added one more pair, the gear was able to absorb more Impact stress and plane didn't loose parts on landings anymore.

Edited November 19, 2017 by Urses

[Hotel26]

Incidentally, the conversation about CoM is interesting.  This is a long-range airplane and, now especially with Bewing's advice about fuel tanks on the wings, it carries most of the fuel along the wing-line, but there is a short fuel fuselage ahead of the wings and before the Mk-3 cockpit.  It carries 2,500 units of fuel.  Full, this moves the Com quite far forward which is why the long wait until rotation.  Given the mission profile, this is NOT a bad thing.  Once I have positive climb, I turn off flow from all the other tanks.  (In fact, I should do this in the SPH!)  Once this tank has burned off, the airplane becomes quite responsive.  In the earlier versions, I found that pumping fuel back and for'ard was the most accurate way to trim flight.

And did you know, Concorde had no trim tabs.  Its pilots simply pumped fuel around to trim the aircraft, particularly in level flight while accelerating to Mach 1, anticipating the sudden move in center-of-lift that occurs at Mach 1 going supersonic.  (If I recall, the CoL moves aft causing a pitch-down that has to be pre-empted earlier by trimming.  http://www.concordesst.com/fuelsys.html

Here's a screenshot of the modified version: https://drive.google.com/open?id=1XNmue2O98twPiddUCAOjTNqQAdZC2HpG

And the craft file (contains a MechJeb; hand-edit it out if you don't use MechJeb): https://drive.google.com/open?id=1bpXQ5xMW44TyU7UfhHXw2uHNlKiUvNEN

Thanks for all the help.  Further suggestions to improve welcome.

(I just wish I could get that Prospector back into the cargo bay after sending it out!)

Edited November 19, 2017 by Hotel26

[FancyMouse]

It's ok to put gears on the wings. In fact, it has automatic strutting that probably already implicitly did what other people suggested. But this kind of gear-on-wing setup puts a demand on descent rate, i.e. you'll need a slower rate of descent upon landing not to break anything. And if it breaks like you said, you typically need to land more gently (don't tell me you approach at -10m/s or even more steep). A pair of biggest gear is enough to handle half fuel of 75t takeoff weight plane (I have one of 100t takeoff weight on a pair of biggest gear on wings and it still lands normally).

[AeroGav]

10 hours ago, Hotel26 said:

t takes the whole 3km to reach Vr = 120m/sec, but with half tanks it will do it at 40 m/s and climb solidly

Oofya,  that's a fast takeoff speed.     A craft that weight needs more wing, you'll have to get creative joining non-modular wing parts together to give it more lift.    Did you say you wanted this as a STOL transport ?    Most of my spaceplanes can take off below 40 m/s

one of mine - 

bir

not one of mine, but i like it - 

[AeroGav]

7 hours ago, Hotel26 said:

Here's a screenshot of the modified version: https://drive.google.com/open?id=1XNmue2O98twPiddUCAOjTNqQAdZC2HpG

And the craft file (contains a MechJeb; hand-edit it out if you don't use MechJeb): https://drive.google.com/open?id=1bpXQ5xMW44TyU7UfhHXw2uHNlKiUvNEN

I downloaded this version and rotation was about 130-135 m/s at the very end of the runway.    There is a big gap between the CoL and CoM so it takes a lot of force to get the nose up.  Your tailplane is having to push down very hard to make the tail drop and the nose go up - it is working against the wing, generating downforce instead of lift.    The ailerons are also attempting to pitch the airplane, by pushing down.   Since they are close to CoM they have very little effect on pitch for all the negative lift they are making.

I made a canard version by of your airplane by swapping the tailplanes to the front (attached them to the cockpit) and deleted those little stubby whiskers it has atm.  Then moved the wings and engines waay back (attached them to cargo bay) to give us the same distance between centre of lift and centre of mass as the original version.     I disabled pitch control on the ailerons.

This brings rotation speed down to 115 m/s.    With more wing and  closer col/com you could lower that waay further.

If you can arrange the plane so it's CoM don't move much full versus empty you can safely bring CoL and CoM closer together, making a more efficient airplane.  Cargo bay over centre of mass perhaps, then add some mk1 tanks around rear fuselage?

EDIT - Bedtime for me,  this version of your ship holds 10,898 units of fuel,  rotates at 85 m/s (just as you're reaching the control tower)

https://www.dropbox.com/s/6922w259tzsc21s/Transport 2d.craft?dl=0

Edited November 19, 2017 by AeroGav

[GoSlash27]

If your wing mounted gear are unequal to the loads, you can always put another gear inline with then under the fuselage.

I have used this technique for large rough field capable aircraft before.

Best,
-Slashy

[AeroGav]

10 minutes ago, GoSlash27 said:

If your wing mounted gear are unequal to the loads, you can always put another gear inline with then under the fuselage.

I have used this technique for large rough field capable aircraft before.

Best,
-Slashy

Yes, I just actually tried to land my version of his plane and despite the shorter takeoff length,  wings still come off if you touch down with much more than 5m/s vertical.    So  I made a version with body and wing gears -  updated the dropbox craft again !

[Hotel26]

Thanks.  The tip about the inner ailerons in pitch mode being useless is noted.  I'm going to remove them.

Remembering that the problem description is landing impact breaking the wing roots (seems to be solved but still testing), the CoM issue is secondary.

7K units of fuel are over the wings and make no difference to the CoM in flight.  It's the 3K units of fuel in the short fuel fuselage ahead of the wings that make the issue.  Simply removing this module and shifting the wings back to match the CoM would "solve" that issue.  Then I'd have an airplane that could take off in 25% of the runway at KSC, climb hard and be simpler to fly.  The only penalty would be losing 3K units fuel of flight range. 

EDIT: So I did that and the fully-laden take-off certainly works better.  A shorter aircraft length might improve unladen performance (yet to be tested).  And I guess it's a 25% reduction in range.  (I'll have to weigh the decision.)

For anyone who has downloaded the craft file, take it into the SPH and set the fuel in the for'ard fuselage to 0 and watch where the CoM moves.  Presto!  STOL machine.  This is the way I test: for'ard tanks full and for'ard tanks empty.

Lots of great lessons in all of the above replies to my post; thank you all.

Edited November 19, 2017 by Hotel26

[Rodhern]

3 hours ago, AeroGav said:

 

The crew seem a tad .. uhhm .. apprehensive though. 

[AeroGav]

5 hours ago, Hotel26 said:

Thanks.  The tip about the inner ailerons in pitch mode being useless is noted.  I'm going to remove them.

Remembering that the problem description is landing impact breaking the wing roots (seems to be solved but still testing), the CoM issue is secondary.

7K units of fuel are over the wings and make no difference to the CoM in flight.  It's the 3K units of fuel in the short fuel fuselage ahead of the wings that make the issue.  Simply removing this module and shifting the wings back to match the CoM would "solve" that issue.  Then I'd have an airplane that could take off in 25% of the runway at KSC, climb hard and be simpler to fly.  The only penalty would be losing 3K units fuel of flight range. 

EDIT: So I did that and the fully-laden take-off certainly works better.  A shorter aircraft length might improve unladen performance (yet to be tested).  And I guess it's a 25% reduction in range.  (I'll have to weigh the decision.)

For anyone who has downloaded the craft file, take it into the SPH and set the fuel in the for'ard fuselage to 0 and watch where the CoM moves.  Presto!  STOL machine.  This is the way I test: for'ard tanks full and for'ard tanks empty.

Lots of great lessons in all of the above replies to my post; thank you all.

Obviously I am struggling with the structural/landing gear side of things, I'm better at aerodynamics.    You didn't use the "rigid attachment" option for the wings did you? That makes them stiffer but more brittle.

I don't see why you need so much fuel though.    What altitude are you cruising at ?   Fuel burn is less at altitude.   Best lift/drag and glide ratio at supersonic speed is when the nose is 5 degrees above prograde.   So,  assuming you are cruising at full throttle (no reason not to), keep raising the cruise height until the nose is 5 degrees above the horizon just to make enough lift to get level flight.

This thing can go around Kerbin 3 times at this altitude.   The only fuel tanks are in the wings - the fuselage is entirely taken up with crew cabins.

Note - on this airplane the nose is directly on prograde, because the wings are already angled up at 5 degrees where they attach to the fuselage, but on yours, with flat wings, you will cruise best with nose 5 degrees over prograde.

 

[Hotel26]

I cruise at 20km at about 1140 m/sec.  I don't' need 75% circ; 50% is enough to get anywhere.  I do like the idea of "reconnoitre range", though.

"Rigid attachment" is another new term to me.  Thank you!  I'll look into that one.

Note this is before I deleted the short fuel fuselage (with MechJeb and solar panel) and the inboard ailerons.  This works and it's pretty clean.

NB: I haven't completely finished testing its rough field performance.

Edited November 20, 2017 by Hotel26

[Hotel26]

On 2017/11/19 at 7:02 PM, bewing said:

 (Hold down Shift, in order to move the parts a long way.)

 

 

Gotcha.

[Daniel Prates]

On 19/11/2017 at 8:21 AM, Hotel26 said:

Problem is that I cannot land it any more without the wings breaking off.  (Maybe I just need to get better at landing but this one is palpably fussy about it compare to my SSTO (which is a good deal lighter)).

 

Except for the runways, all surface in kerbin is pretty rough, making landings difficult. However you can circumvent this somewhat, with four easy steps:

1 - Do use some autopilot, like 'atmospheric autopilot'. It is well worth having. With it you can fly straight, which allows for a finer control of touchdowns. Doing it manually with the keyboard is just asking to have a too high rate of descent, which is a sure way to break everything in your craft. Actually this is true even for the  runways, more so for landing on rough ground!

2 - Strengthen your craft, specially the wings and main fuselage. I am sorry, I did not see your craft, but as a rule if you have a too-long fuselage, made of several connected parts, it will be wobbly. Same goes for your wings, specially if they are too big or heavy (and specially so if laden with fuel). The trick here is to abuse of the strut connector. Slide your parts around to insert the truss between them, then slide them back into position to hide the truss. It will make the whole airframe much more solid. Some of my larger planes can have 15 or 20 pairs of trusses spread all around, carefully  concealed in between the parts. In real live a connector would have to be in a correct length, position and angle to really provide strength, but the game doesn't know that - it just works like a 'superglue' of sorts.

3 - Also, do consider this: the slower you are touching the ground, the better. So if your plane is intended to be a cargo plane, it makes sense to lower the stall speed. It may  require you to add more wing surface, which will increase weight and consumption, besides decreasing top speed, but it will allow you to fly slower without plunging into the ground, so safer landings in difficult terrain become that easier. 

4 - Landing gear placement. Make them larger, add more pairs, fiddle with the spring rating. That goes a long way towards doing more comfortable landings. 

[swjr-swis]

On 11/20/2017 at 12:27 PM, Hotel26 said:

Note this is before I deleted the short fuel fuselage (with MechJeb and solar panel) and the inboard ailerons.  This works and it's pretty clean.

You could still remove the other Mk3 tank, which will give cockpit access to the bay and make balancing much easier. Also, it really doesn't need 4 engines when it can do the same job with 2.

Here's a leaner version that still offers the same cargo capacity, and can circumnavigate Kerbin in an hour with 900LF to spare:

Spoiler

No mk3 tanks, only two jets.

CoM fully fueled.

CoM empty. With the default mini-rover, the CoM barely moves. Heavier cargo will see it move a bit more, but not enough to affect flight characteristics significantly. Still, by just adding an Mk1 LF tank between the tail cones and adding a bit more fuel, this can be resolved too.

Flaps help it to take off by itself

Cruising at 20-23km, 1150-1250+ m/s.

Back at the KSC for an easy landing one hour later, with plenty of fuel to spare.

Reworked the coupling and the rover as well, now it can be redocked after deployment.

Due to less parts and much less total mass, it's good even without any (auto)strutting or rigidizing.

 

Full imgur album here.

Craft file on KerbalX: Transport-3a

 

Edited December 13, 2017 by swjr-swis
fixed forum and imgur disagreeing again... :(

[Hotel26]

Daniel Prates: thank you sir.

swjr-swis: swesome, I am not worthy!!

[Hotel26]

I guess I am going to have to bite the bullet and upgrade to 1.3!

I am on 1.2 and had sworn never to upgrade again.  Oh well.  (It took me a year to get over the last upgrade after I had found all the mod versions.)

My mods are 1. Kerbal Alarm Clock, 2. Docking Port, 3. MechJeb and 4. Haystack.

For this: it will be worth it!

[swjr-swis]

14 minutes ago, Hotel26 said:

I guess I am going to have to bite the bullet and upgrade to 1.3!

I am on 1.2 and had sworn never to upgrade again.  Oh well.

If you are doing this only because of my craft file, don't. Give me 30 mins and I'll rebuild the craft in 1.2.1 (which is it specifically: 1.2.0, 1.2.1, or 1.2.2? Never mind, saw 1.2.1 in your craft file), it'll be less hassle than having to upgrade for just the craft file.

Edited November 21, 2017 by swjr-swis
nvm

[Hotel26]

You are a champion.  I'm pretty much speechless.

I am on 1.2.1.1604 (LinuxPlayer) 64-bit

(In the very early 1980s, Microsoft pretty much invented single-handed Version Incompatibility and made a business out of Planned Obsolescence.  I know because I worked for a large computer manufacturer in the 80s and Compatibility was our God.)

[swjr-swis]

Ok, craft done, uploaded to Dropbox instead of KerbalX to avoid confusion between the two: link here.

 

Took a bit longer... keep forgetting all the little things that were added in 1.3.x that make things easier/faster.

Edited November 21, 2017 by swjr-swis