Radial Decouplers still not fixed???

[biohazard15]

The best solution would be Tweakables Everything mod - with it, you can set decoupling force to 0 and then safely use sepratrons.

[KasperVld]

are you sure this outwards rotational force is not a result of how they are placed that is causing a lever action? For example, take this very quick, ugly drawing of mine

The decouplers (in red) are placed far below the center of mass (yellow circle) of the empty fuel boosters. This causes a lever force which pushes the boosters outwards at the bottom which causes a rotation around the center of mass (blue arrows) and pushes the booster inwards on the top (green arrows) causing them to collide with the rocket. I ask this because personally I haven't had any issues with the decoupler force itself. The ideal place for the decouplers would be on or just above the center of mass of the empty booster (since the rocket accelerates upwards the bottom of the boosters can safely be pushed in a little when you decouple them).

P.S.

SepMotors are king when you're building with very heavy boosters

Edited October 16, 2014 by KasperVld

[Padishar]

That may be relevant when your boosters are made of multiple fuel tanks with an engine at the bottom but for solid boosters it is simply not possible to attach them in that way. When you attach a booster to a decoupler it uses surface attachment and the surface attach node on a solid booster is basically in the middle (height-wise) so the middle of the booster has to be somewhere on the decoupler. Also, the CoM of a solid booster doesn't actually move as the fuel is drained and is always roughly around the middle so they always line up pretty well.

This is a confirmed bug in the public bugtracker.

[Claw]

Kasper, I am absolutely certain that what you describe is not the problem in this case. I'd post pictures now but I'm mobile.

This problem was introduced in either 0.24.1 or 0.24.2 (I forget which off hand) when the stack decouplers were fixed. Technically the stack decouplers are also broken, but it's much less apparent than with the radial decouplers since they don't tip in and destroy the rocket.

I'm actually working on / testing a new fix for this, but it's not done yet.

-Claw

Edit: Here we go, found my pictures from July. This is a known issue and is in the tracker.

Well, this was done in 0.24.2. By the way, the radial decouplers act normally in v0.24.0.

I did a test with a pretty simple rocket in both 0.23.5 and 0.24.2 (pictures split below so you can compare side by side). The decouplers are functioning, but not like they used to. It's causing problems for previously well established rockets. Also, it's making things a bit difficult for constructing new rockets because radially attached boosters are coming off crooked. In the example below, I'm unable to attach the BACC solid boosters low enough to prevent the tops from toeing in.

This was done with the exact same .craft file. I copied and pasted it around.

Here is the rocket in 0.24.2:

- Note the top of the solid boosters tipping in. There is also two pictures of the large decoupler with a 3 second time lapse.

Javascript is disabled. View full album

And images of the same rocket in 0.23.5:

- Note the solid boosters separate cleanly. The 3 second time lapse of the large decoupler shows a much greater separation, and that's with a half full giant tank attached.

Javascript is disabled. View full album

Edited October 16, 2014 by Claw

[Claw]

Okay. It's not perfect, but here is a link to the stock bug fix add-ons I have been working on. Not the perfect answer for the stock purist, but there's a radial decoupler fix in there.

http://forum.kerbalspaceprogram.com/threads/97285-0-25-Stock-Bug-Fix-Modules

Cheers,

~Claw

[ansaman]

Thanks Claw. I tried it and it seems to work. I appreciate your efforts.

[Kerano]

Kasper, I am absolutely certain that what you describe is not the problem in this case. I'd post pictures now but I'm mobile.

This problem was introduced in either 0.24.1 or 0.24.2 (I forget which off hand) when the stack decouplers were fixed. Technically the stack decouplers are also broken, but it's much less apparent than with the radial decouplers since they don't tip in and destroy the rocket.

I'm actually working on / testing a new fix for this, but it's not done yet.

-Claw

Edit: Here we go, found my pictures from July. This is a known issue and is in the tracker.

Well, this was done in 0.24.2. By the way, the radial decouplers act normally in v0.24.0.

I did a test with a pretty simple rocket in both 0.23.5 and 0.24.2 (pictures split below so you can compare side by side). The decouplers are functioning, but not like they used to. It's causing problems for previously well established rockets. Also, it's making things a bit difficult for constructing new rockets because radially attached boosters are coming off crooked. In the example below, I'm unable to attach the BACC solid boosters low enough to prevent the tops from toeing in.

This was done with the exact same .craft file. I copied and pasted it around.

Here is the rocket in 0.24.2:

- Note the top of the solid boosters tipping in. There is also two pictures of the large decoupler with a 3 second time lapse.

http://imgur.com/a/l3irB

And images of the same rocket in 0.23.5:

- Note the solid boosters separate cleanly. The 3 second time lapse of the large decoupler shows a much greater separation, and that's with a half full giant tank attached.

http://imgur.com/a/rS3Qn

Thank you so much for documenting this. I thought I was going insane when I spent hours designing an Eve return vessel in v0.24.2 only to have the parts fall inwards and destroy the rocket. No matter how much mass I attached below the radial decouplers, they always defied the laws of physics and turned inwards. I was so sure an identical design from a previous version of KSP had worked perfectly, and I couldn't work out what the heck I was doing wrong. Turns out it was a bug all along... well, that explains it.

I'm surprised this bug hasn't been highlighted more since it's a very serious one which cripples basic rocket design. I really hope this is fixed with the next update.

[Starwaster]

Anyone who thinks that this bug can be worked around by tweaking decoupler forces, through the use of sepratrons or that it has to do with decoupler position relative to center of mass of the separated part hasn't truly experienced the bug.

The bug generally strikes hardest when separated parts are strutted to parts other than the part that the decoupler is connected to. So if I have a long booster connected via decoupler to a fuel tank and I strut to another fuel tank stacked on top of the base fuel tank.... that's a recipe for disaster.

The reason sepratrons are ineffective is that an inwards momentum has already been imparted when the sepratrons fire. The momentum was immediate and the part has to decelerate before the sepratrons can even start to push the part away.

From a construction standpoint, don't strut anything other than the two parts that are attached via the decoupler. Fewest possible struts are best in this situation. MOAR struts bad.

[Claw]

The bug generally strikes hardest when separated parts are strutted to parts other than the part that the decoupler is connected to. So if I have a long booster connected via decoupler to a fuel tank and I strut to another fuel tank stacked on top of the base fuel tank.... that's a recipe for disaster.

I've learned some more with one of my recent bug fix modules. If you could provide a stock example craft of what you're talking about, I might be able to modify the Decoupler bug fix add-on that I've already created.

Cheers,

~Claw

[Claw]

I believe I have finally cracked this one. I've just uploaded a new radial decoupler fix module to my Add-on thread. As far as I can tell, it seems to restore the radial decoupler's functionality. I think I've also fixed the problem where struts were stealing all of the decoupler's momentum.

Give it a try. If you have problems still, please post me some pictures!

Cheers,

~Claw