Large ship pulls to side while being tugged

[SlothPaladin]

I'm not sure if this is a bug or some massive design flaw but I built a large spaceship which I put in orbit in 5 parts for a mission to Jool, now I'm trying to get a 2000m/s burn to start it on it's way there but when piece 1 is ignited (at about 1/4 power) the ship pulls down and to the right hard.

Initially the ship was only section 4 and 5 but the thrust to weight ratio was pathetic and I only had enough delta v to get to Jool, with none for the return. So in a not very science-y or mathematical approach I figured I'd just send up a ton of fuel and some super powered engines. So I sent up two payloads of fuel (section 2 & 3 each with one S3-14400 tank), and new engine section (with two KR-2Ls and 2 S3-3600 Tanks)

On my initial attempt to get out of the Kerbin system I had the engine section between section 3 & 4 pushing section 4 & 5 while pulling section 2 & 3, that way I could drop fuel loads as I finished them, but the thrust from the engines made the craft fold in half (to be fair I'm not really surprised as those two engines create an insane amount of thrust).

I figured if I couldn't push my ship I should be able to pull it. so I took block 1 out from between blocks 3 & 4 docked the fuel section to block 4 and flipped block 1 around and put it in the configuration you see in the screen shot. Block 5 was meant to be the forward facing section and every other section was designed to be pulled behind it. My first thought was maybe I messed up on some fuel flow but I checked and fuel was flowing from 3 to 2 to the center tank in 1 and then out to the radial tanks for 1's engines exactly as planned. I looked through all the tanks but everything seems to be normal, no engines are enabled that shouldn't be. Then I figured maybe reaction wheels don't like moving in the wrong direction, so I disabled all reaction wheels that were not in block 1 but still the same thing.

I thought that pulling the ship would minimize any off center issues I had, not that there should be anything, all ships hanging off of section 4 are identical with 3 way symmetry and the rover and mini-ship for section 5 were weight matched and I launched them into orbit without a problem.

If you have any suggestions, or if there are any ingame issues that might complicate a ship like this I'd like to know. Any other troubleshooting paths I should look into?

The ship is 948 parts and 453 tons.

[Vanamonde]

Your masses docked to segment 5 do not appear to be symmetrical. If that rover-looking-assembly is on the side towards which the ship is deflecting, I think you've found your problem. It may have worked when the segments were assembled in a different order and it was closer to the COM and therefore was getting less of a boost from leverage. Also, check the amount of fuel in the tanks on those side-assemblies, as they may not be draining the way you expect them to, creating an imbalance.

Interesting ship design, by the way. But in interplanetary travel, you'll find that fuel efficiency is much more important than thrust. I'd suggest replacing those big engines with LV-Ns. Or are they modded LV-Ns? Anyway, good luck.

Edited May 19, 2014 by Vanamonde

[phoenix_ca]

Vanamonde's pretty much covered all the bases as to why it's happening, so, here's some ideas on fixing it with mods. :3

The easiest solution may be to use Throttle Controlled Avionics, which should balance thrust so that you don't lose control.

If fuel imbalance is the issue, check-out TAC Fuel Balancer.

[Kasuha]

It is known problem in KSP that engine gimbal control assumes that engines are placed behind ship's center of mass - which means if they're placed ahead of it, they aim exactly the opposite way than they should. If you disable gimbal on any of your engines ahead of CoM and steer the ship just with torque, it will work.

[phoenix_ca]

It is known problem in KSP that engine gimbal control assumes that engines are placed behind ship's center of mass - which means if they're placed ahead of it, they aim exactly the opposite way than they should. If you disable gimbal on any of your engines ahead of CoM and steer the ship just with torque, it will work.

Well that's an evil bug.

[xcorps]

Interesting ship design, by the way.

The crew list is a little different too.

[Technical Ben]

It is known problem in KSP that engine gimbal control assumes that engines are placed behind ship's center of mass - which means if they're placed ahead of it, they aim exactly the opposite way than they should. If you disable gimbal on any of your engines ahead of CoM and steer the ship just with torque, it will work.

Wow, never knew this... will use that in all future builds. That and I find massively heavy ships without sufficient SAS tend to spin anyhow (I'm guessing tiny wobbles are exacerbated with weight).

[SlothPaladin]

Wow, thanks guys, there is a lot of useful info here.

Kasuha: I'll have to try disabling the gimbaling and see if that works better.

Vanamonde: The two vehicles on the 5th segments have the same mass, but the masses center might be off, if that's the root of the problem I'll probably have to build a new ship for the mission. I had initially planned to use the 8 Atomic Rockets on segment 5 to do the exit burn but the burn was going to be something like 30 minuets or something and I got impatient and sent up that stupidly over powered engine segment. Maybe I just need to build an engine segment with 32 of them or something, but that would be a part count nightmare.

[SlothPaladin]

Sorry for the double post but I'm home now and can actually test some of these theories out.

After locking the gimbals the amount of movement that the ship went through was greatly reduced, if I throttled up to 100% the movement was about the same as being throttled up to 25% before. So I think the gimbal issue was the main cause. Ship still has issues but this whole thing has given me a lot of good redesign ideas.

As far as I can tell the masses on the side are pretty close to perfect, I may still try to use this setup on a new ship:

I also figured out a way to quadrupal the thrust while using the LV-Ns while only adding 32 parts by doing some of this silliness:

[cantab]

A couple of other ideas:

The engine thrust could be just hitting some of the structure behind it. If this happened for one engine and not the other, because the ship flexes, that engine would lose thrust (with no visible change) and the ship would veer. The solution would be to add a space between parts 2 and 3.

Check the thrust limiters on the engines. If they're different, there's your problem, make them the same. If they're the same, try reducing thrust on the engine that's "outside" when it veers to balance things.