Radial Decouplers?

[allmappedout]

Hi all,

I've been building rockets for a while now, but i've never known how to solve one issue that has always bugged me - for long sets of fuel tanks (say, 3 or 4 big high), they can only have one connection point when you're using radial decouplers. As a result, they rotate and wobble around this point, which usually results in unplanned disassembly. Placing multiple decouplers up the length of the main rocket doesn't help because they're not connected to the outside stack and when they decouple they tend to explode in hilarious fashion.

Apart from large amounts of strutting, there doesn't seem to be an elegant solution to this issue. Is it just me being stupid? Am I missing something obvious? Is there a solution, such as attaching docking ports to both sides prior to connecting the tanks so when physics loads all the decouplers are properly linked? (has someone tried this?)? Or do I just have to grin and bear it? For reference, I like to build pretty tall rockets, so I have Kerbal Reinforcement installed which makes this easier, but it doesn't seem to solve this issue.

Alternatively, are there any mod sets that have really long decouplers that are, say, nearly the length of an orange fuel tank so the connection is stronger?

Thanks for your help!

Edited November 13, 2013 by allmappedout

[UberFuber]

Sad to say, I think the current solution is, as you said, strut.

Although I think you only need about 1 strut to main structure per fuel-tank. Two for the big, orange ones. So it might not be as many struts as you think.

[M5000]

I feel your problem, and I have a bit of good news, a bit of bad news, and a bit of news that isn't quite one or the other.

The good news: Yes, using docking ports along the sides of your rocket may work for multiple attachment points. While they will NOT be attached in the editor technically, they SHOULD dock almost instantly upon physics enable. Only thing you'll need to watch for is uniform distance from the center stack, so you may need to get rid of the radial decoupler and opt for docking port-only connections. Again, the pieces, if built correctly, will not connect in the editor, but will dock upon physics load. However, this really isn't the best option.

The bad news: It won't matter how long you make the decoupler, unfortunately, since you'll still only attach at one point. The tank would pivot around there and just clip through it as it bows/bends under the weight (just like any other time you'd do this with a teeny tiny decoupler). So, a really long decoupler isn't going to help.

The neutral: It looks like strutting is your only real option, unless something fundamental to the way the game calculates collisions and craft trees is changed. However, strutting doesn't have to be ugly. Put two struts on top and two struts on bottom of each radial tank, connecting to the central stack. Put one strut on top and one on bottom connecting to the next tank over (if you're using more than two tanks). Add some separation motors to push them out of the way (since the struts will kill your decoupler force) and that should be all you need. This method (using the KW heavy struts for me, but I've no reason to believe it wouldn't work for stock ones) holds two orange tanks stacked atop each other to a central 3.75m core. I'm sure you can adapt it in your own way.

So yeah, looks like strutting is going to be your only valid option.

[allmappedout]

Cheers; that's a shame. Like I said, I've been building for a while and it's just always bugged me that there's no better way to do it - I usually use the x-beam strutting idea to keep everything from shaking apart and that works fairly well, but hopefully this'll look like something that could be improved in future releases (obviously I understand the limitations in the parent-child structure of craft, I was talking more about the physics between parts).

[UberFuber]

Hm... perhaps a new type of radial decouplers that's several meters long and much stiffer?

[allmappedout]

Hm... perhaps a new type of radial decouplers that's several meters long and much stiffer?

I think that's a simple solution, i'm sure there'd be lots of uses for it (particularly for launching rockets from aircraft, for example) as well as the one I mentioned above. We shall see....ISWT!! (In Squad We Trust)

[Geschosskopf]

Hmm, well, I don't make lifters that huge so this might not work, but you could possibly do it like this:

Don't physically connect the tanks of the side stacks. Instead, mount each one individually to the core on its own decoupler with small gaps between the side tanks, and connect the side tanks with fuel lines from the top down. Functionally, this would be the same as having all the side tanks directly connected to each other but with the added stability of each tank being directly connected to the core. You'd probably still need some struts per tank but maybe not as many as otherwise.

There's also the possible advantage of being able to jettison the upper side tanks as they go dry instead of having to keep them all until the whole side stack empties. However, doing this would make your staging list a mile long and you'd need beaucoup Sepratrons to throw the upper side tanks clear of those below.

[M5000]

I'll go ahead and post this, from the B9 Aerospace thread. I believe this is still within context.

No collisions ever happen within the craft hierarchy,

That means.. Longer decouplers will not help the situation.

A stronger attachment point? Sure, though.

So there you have it. I believe the only time you get explosions is when a part falls OFF or is no longer "part of the craft." When that happens, it may have been clipping and been just fine if it were still attached, but as soon as it falls off... BOOM.

That's how I've come to understand the craft hierarchy.

Edited November 13, 2013 by M5000

[allmappedout]

M5000, thanks for that, that's good to know and fits in with what i've seen, now I think about it.

Geschosskopf, that's a cool idea, I'll have to give that a whirl!

[Kerbart]

You really shouldn't need that many struts to stabilize your design. One set for every tank in the stack usually does the trick. If your design as very thin and tall... Well yes that's how we like them rockets here on earth. But not in the Kerbin universe. Just as we don't see football-stadium aircraft with wings the size of a tennis field -- as much as we'd like that design to fly, it just doesn't work here on earth. Same goes for "earth like" rocket design on Kerbin; it just doesn't work that way. Short and chubby is the way to build your Kerbal rockets (for now).

[Kasuha]

You don't need ridiculous amounts of struts. You need two struts at bottom and one or two struts at the top of the stack, that's all.

The two struts at the bottom should be mounted along the decoupler, higher on the tank which gets staged off, lower on the tank which stays after decoupling. Their purpose is to press the tank against the decoupler as the staged stack becomes empty and its lifting power grows more than for the unstaged part.

The two struts at the top should be mounted with their lower end on the staged tank and their upper end on the main column. When the stack leans toward the main column due to lifting force, it helps to lift the main column and decrease stress within it.

You also need to make your fuel stacks rigid, meaning you need to strut your tanks within the stack together. Four or six struts between each two tanks should suffice.

And of course, you need to prevent any sudden turns during liftoff. Gravity turn is supposed to be a smooth maneuver.

Edited November 13, 2013 by Kasuha