Return of the Kraken?

[Jaevko]

First, I apologize for the cross-forum post. I initially posted this in support, but upon reading the stickies and reviewing that forum, I realized it was more for crashy-sort of issues. This, while technically a bug, is probably more of an physics engine-limitation/ship design question, so I'm going to repost here.

Youtube Link

The only mod I'm using is MechJeb.

Repeated issue with this ship exploding or falling apart. Sometimes it also "accelerates" (acts as if an engine is on) and the orbit changes when it's not supposed to. It happens under acceleration or freefall, with autopilot or SAS on or off, with torque wheels on or off. It has happened in Kerbin orbit, sun orbit, Ike escape trajectory into Duna orbit, in atmo during Duna aerobraking.

It is actually two ships: the nuclear tug in the back and the lander/science ship in the front. The lander/science ship seems to be the problem: it has happened once or twice just to that with the tug undocked.

How can I design the ship so this won't happen? KSP is pretty much unplayable with this relatively uncomplicated ship, it took countless quick-saves to get to Duna. I did use the terrible barely-working "Rockomax HubMax Multi-Point Connector" piece, and the four "arms" on the lander attached to that piece wouldn't work with symmetry. I created one arm, and then had to copy/rotate it to create the other 3. Is that the problem? Or is it something entirely unrelated to the hub max piece?

Edited August 30, 2014 by Jaevko

[Taki117]

Is there any part clipping going on? (And I do mean ANY)

[Claw]

Yeah, I would ask the same question as Taki. Especially if there is any clipping near your docking ports. Docking ports really do not like clipping at all.

That being said, I've seen other similar disassembles with similar, X-shaped fuel tankers. While I couldn't figure out what was causing the shake/explosion, it was controllable by toggling off the torque modules immediately upon switching to the craft. That allowed enough time to move fuel around. For some reason, moving fuel around caused enough of a change in the mass distribution that the shaking didn't happen after that.

Good luck,

~Claw

[Jaevko]

Part clipping means that I have to actively open a menu and allow it, right? Nope, never did that (although I definitely did my best to make the parts fit as closely as possible).

[steve_v]

Not always, parts will often clip without using the debug window (usually just not how you want them to). Strut clipping for eg. is notorious for creating 'phantom torque' and other strange issues.

Often you can spot the offending part by carefully inspecting for wobbling / vibrating parts on the launchpad, If stuff is too close together it will sometimes clip momnetarily as the physics loads, then try to spring back.

What's happening in the video looks exactly like a clipping issue, yet I can't se anything that would clip?

What happens if you use say, a round tank with 4 radial adaptors rather than the hub part?

Edited August 30, 2014 by steve_v

[Sirrobert]

Part clipping means that I have to actively open a menu and allow it, right? Nope, never did that (although I definitely did my best to make the parts fit as closely as possible).

You can enable part clipping yourself yes, but it can also happen on accident after docking (for example).

The on accident cases are usually what cause.

In the video you can also see that there is reaction wheel input on the craft, dispite SAS being turned off. Where are the reaction wheels located?

[Roadkill]

Jaevko, would you please upload the craft file somewhere? I'd like to look at this a little closer.

[Jaevko]

Wow, thanks for all the replies and help, this seems like a really great community!

Here's the craft file:

https://drive.google.com/file/d/0B1MXNWoMTkwNTGNaUmswRlpTNlk/edit?usp=sharing (Let me know if there are issues with that, this is the first time I've shared a craft file. The only modded pieces on there are the two mechjeb controllers).

I see some potential part clipping on the tug, but don't see much on the lander. (Maybe the parachutes next to the radial engines?)

Like I said, I'm pretty sure the problem is the lander, not the tug: The first time it happened was in low Kerbin orbit with the two undocked (gassing them up) and it happened to the lander/science ship.

Note that in the video, some of the fuel tanks on the tug have been detached, and the two different ships are probably rotated compared to how they are in the craft file (since they undocked and redocked a couple times).

[Jaevko]

What happens if you use say, a round tank with 4 radial adaptors rather than the hub part?

Haven't tried that yet, but I'll definitely try that in future designs! Stupid hub part

I just noticed that the Pegasus Mobility Enhancer is definitely clipping a little with the Telus Mobility Enhancer. Also, Pegasus ME's might be clipping up and down the length of the lander. Surely those tiny pieces aren't the problem? Looking at how it shakes itself apart, I would guess it was something with the arms.

I am, however, doubting the whole "clipping" theory a bit. This Kraken only happens at certains times during different parts of orbits. For example, during a "Krakenless" part of an orbit, I can save, reload, and do anything and everything trying to get it to attack the ship, and nothing will happen. During a really bad part of an orbit, I am unable to maneuver for even a second without it striking hard and fast. Reloading it multiple times doesn't help: when this happened orbiting the sun trying to perform an intercept orbit to Duna, I had no choice but to put time acceleration on and wait another whole year until there was another opportunity to perform the burn (at which point the Kraken decided to go away and completely stop causing trouble for a bit). For that reason, I am having doubts about the whole part clipping theory (why would it come and go like that depending on the orbital situation?)

Edited August 30, 2014 by Jaevko

[Roadkill]

I think Claw is right about the craft, Jaevko. When I hyper-edit the top portion of your craft into 100,000 km Kerbin orbit, it's as stable as a rock. So, it has to be something related to fuel balancing or it could simply be kraken related errors associated with an SOI change. There is some part clipping going on also, which does lead to stability issues after SOI changes. Here are some screenshots and some pointers to go with them :

Attach the fuel lines to the fuel tank, not the goo canisters, The goo canisters have animations involved that could lead to problems.

center the fuel tanks on the decouplers. This will make the craft more stable

This parachute and the engine are clipping. Be careful when using symmetry in the VAB, Make sure there is no clipping.

The parachute and the goo cannister are clipping, again caused by placing parachutes with symmetry in the VAB.

Parachutes on the materials bay make me nervous again because of the animations involved with the materials bay.

Finally, I would place an adapter of some kind under this sas module to give the goo canisters a little more room to operate.

Hopefully these tips will help you to avoid the dreaded kraken.

Edited August 30, 2014 by Roadkill

[Jaevko]

Thanks for your help. I'll go ahead and redesign with all of that in mind. I really appreciate everyone's help, especially Roadkill's. Thanks!!