[Solved] Airlock collider strangeness.

[StarVision]

Javascript is disabled. View full album

So I've run into some problems creating an airlock/ladder for this command pod. From what I know I set it up correctly, but then when I go into KSP, no dice...

What's worse is I've done this plenty of times before with no problems, but for some reason it's a problem now. Also an interesting bit about this, is that when I move the colliders down to the wider part of the command pod, it works fine.

I was wondering if anyone has any clues to what the problem is.

Here's the project file: Download

and the part file: Download

Edited September 21, 2014 by StarVision
Solved

[philotical]

TT wrote once this:

Hatches Box colliders typically MUST overlap the collider for the model, and the models collider must overlap the ORIGIN of the model (0,0,0). You might get away with either of these but if you get "Hatch obstructed" bug, not following those is probably the cause.

It didn't help me with my hatches, because KSP is not consistant in that..

In one model it worked (shuttle), for th next, I had to make the hatch collider that big, that it'selfe overlaps the origin (belerophon)..

no clue why - but just try around with the collider size..

The code that checks if it's a valid hatch is buggy and inconsistant to say the least..

[JeffersonFlight]

Try moving the root obj down on the Y.

[Porkjet]

After you've tested the above suggestions, test if it works if you delete the bottom collider (so you'll have only 1 col left). I've had cases where having more than one collider near the hatch caused problems.

[nli2work]

I'd try a capsule collider; after the first pod where I used a box collider, I only used capsule colliders for Airlock and never looked back.

as last resort you can try moving everything lower so the hatch collider is closer to 0 on the Y; then use CoMOffset in config to move the COM back to where you need it.

[StarVision]

Thanks everyone, I think I managed to get it working. I ended up switching to the capsule collider, moving the model down (and offsetting the CoM), moving the hatch further up the surface, and making the collider extend a bit more into the mesh. I'm not quite sure why it wasn't working before, but as Porkjet said, it was probably something to do with several colliders overlapping.

It's not quite in the center of the actual hatch, but it works and it's not too noticeable. Again, thanks for the help.