Any advice on stage separation?

[CobraA1]

Any advice on stage separation? I've been having troubles. Dunno if it's MechJeb or my designs, but it seems pretty difficult to design a clean separation with the large components.

[sal_vager]

Hi CobraA1, being able to see and test your craft would help immensely, how about you upload a pic to Imgur, and your .craft to Mediafire, then post the links to both here?

Don't forget you can edit your original post

[Randox]

I had a problem after I upgraded to a better large seperation ring (that won't crumple like a tin can during liftoff) where my ejected radial components liked to clip things on the way down, causing a few incidents, so I ended up ordering some mega explosive bolts and upped the ejection force on the large radial decoulplers to nearly double the vanilla standard, which gives them enough bang to not hit things on the way down. Sequencing things so that parts above other parts eject from the bottom up (I had a rocket dump it's spare fuel tanks on top of its boosters rather than drop the boosters) is also key. It would help greatly if we could see your ship though.

[CobraA1]

I'll post the ship design after work (9 or so hours after this post), but the issue I'm having largely seems to stem from rotation of the outer stages after their release, which tends to collide into the inner stages. All too often, I've had the innermost stage's engine cut off before it has a chance to fire. MechJeb's gravity turn probably isn't helping, as it adds some rotation as well.

I'll probably upload to Imageshack/Dropbox, as I have accounts with them already.

so I ended up ordering some mega explosive bolts

That a mod or something?

[Ziff]

I'll post the ship design after work (9 or so hours after this post), but the issue I'm having largely seems to stem from rotation of the outer stages after their release, which tends to collide into the inner stages. All too often, I've had the innermost stage's engine cut off before it has a chance to fire. MechJeb's gravity turn probably isn't helping, as it adds some rotation as well.

I'll probably upload to Imageshack/Dropbox, as I have accounts with them already.

That a mod or something?

He edited the part.cfg file to increase the decoupler ejection strength. I only have problems if there are too many tanks stacked vertically. A photo would be immensely helpful.

[CobraA1]

Okay, so here's the dropbox link, includes screenshots and one of the many designs I've tried:

https://www.dropbox.com/sh/6ftj73unsiyp8su/EX33alEKbu

Hopefully the dropbox link works, looks like they recently changed how stuff is shared.

[Ziff]

Okay, so here's the dropbox link, includes screenshots and one of the many designs I've tried:

https://www.dropbox.com/sh/6ftj73unsiyp8su/EX33alEKbu

Hopefully the dropbox link works, looks like they recently changed how stuff is shared.

Yep. Stacking 4 of those large tanks on top of each other, without creating a large space between them, is almost a surefire way to cause a separating stage to crash into another part. I never stack more than 2 of those tanks vertically when they are laterally mounted, and only use 3 at most for a center stage. If you are just going to the Mun or Minmus, your lander is overly large, by a lot, which is requiring you to build a fantastically large ascent stage. Best advice, try to to make your rockets look like rockets. Wider at the bottom, and getting narrower as you get up towards the top. Having all the same size lateral stages is difficult.

[nhnifong]

I had a problem after I upgraded to a better large seperation ring (that won't crumple like a tin can during liftoff) where my ejected radial components liked to clip things on the way down, causing a few incidents, so I ended up ordering some mega explosive bolts and upped the ejection force on the large radial decoulplers to nearly double the vanilla standard, which gives them enough bang to not hit things on the way down. Sequencing things so that parts above other parts eject from the bottom up (I had a rocket dump it's spare fuel tanks on top of its boosters rather than drop the boosters) is also key. It would help greatly if we could see your ship though.

Hold q or e to spin up to at least 0.5 Revolutions per second before breaking off the radially coupled junk. their centrifugal momentum will throw them away from the ship.

[CobraA1]

If you are just going to the Mun or Minmus, your lander is overly large, by a lot

Well, I am trying to send a three kerbal capsule to the Mun (later Minimus). I suppose I'll revisit the design. I actually have a working one kerbal capsule design using the smaller components. It's the larger capsule/components that have been giving me a headache.

I probably *do* tend to overbuild, as I never did gravity turns before I installed MechJeb. In fact, it seems to save me an entire stage's worth of fuel doing that gravity turn.

I'll look at a redesign, perhaps a smaller capsule without the extra tanks, and a rocket that has more classical staging. Although I seem to recall having issues with that type of design a couple versions back with KSP.

[Khrissetti]

I tend to largely ignore the new large tanks and decouplers and just build my lander like a mushroom, with the 3 man CM sitting on top of the usual 1m wide parts.

[Ziff]

Well, I am trying to send a three kerbal capsule to the Mun (later Minimus). I suppose I'll revisit the design. I actually have a working one kerbal capsule design using the smaller components. It's the larger capsule/components that have been giving me a headache.

I probably *do* tend to overbuild, as I never did gravity turns before I installed MechJeb. In fact, it seems to save me an entire stage's worth of fuel doing that gravity turn.

I'll look at a redesign, perhaps a smaller capsule without the extra tanks, and a rocket that has more classical staging. Although I seem to recall having issues with that type of design a couple versions back with KSP.

Honestly, all you need for a Mun landing is the command module, parachute, the FL-T16 tank, 3 of the new landing legs, and the new 2m LV-909 engine. That gives your lander a delta-v of 2300m/s. Enough to do a full landing and return to Kerbin. Start with that first. That saves you a ton of weight. The injection stage only requires 900m/s (less if you time it right) for a transfer to the Mun. So if you build an injection stage with a fuel efficient engine with say, 1500 delta-v, you can use some of that fuel for the start of your landing, thus saving more fuel in the landing stage, just in case. I threw this together just to show you what it might look like.

This is the lander with the injection stage attached. The lander has 2300m/s delta-v, enough to do a full landing/return to kerbin, and the injection stage has about 1600m/s of delta-v. Enough to transfer to the Mun and even do most of the landing. Now it just needs an ascent stage.

Anyway, back to your original rocket. I would try to space the tanks farther apart so they don't hit when they drop off. It's really tough when they are that tall though. I basically try to get up to 9-12km with 1 stage, and eject it there, and then have the rest of my ascent stage last until I am at 30-40km, where I eject that too. Then I coast up to 75km and have an orbit insertion stage to burn me into orbit and just before the periapsis hits 30km I drop that stage as well. Then I use my TMI stage to finish the job. It's a little bit more complicated but it is efficient and saves me from having to drop stages while in that critical gravity turn where things like to bang together.

[CobraA1]

Humm, just uploaded Apollen 4, my latest attempt. This time I went with the following:

-Large 7 rocket bottom stage, two tanks high. Lots of power for lower atmosphere. Added some fins to prevent excessive rotation.

-Top stage is essentially eight tanks, with six that separate when empty. Initially, I had only a single large engine, but it proved to be a tad weak, so I added aerospikes to the bottom of the breakaway tanks.

-Top stage also has some RCS, which is an unsuccessful attempt to deal with an oddity I'm encountering.

-The payload is what I *hope* is enough for a mun landing and return.

You have to set MecjJeb for the gravity turn at 12 km - if you start the gravity turn at the normal 10km, it blows up.

This particular design makes it to orbit fine, but when I set MechJeb for a mun burn, I get the strangest behavior. The rocket seems to start rotating out of control and lose speed. It's almost like a strange sort of gimbal lock? Is KSP or MechJeb not using Quaternions?

EDIT: Maybe I simply don't have enough control? RCS doesn't help, but turning on the vectored rocket does. I suppose it's a massive machine, but there still seems to be some sort of mysterious force acting on it.

Edited August 9, 2012 by CobraA1

[CobraA1]

And success! More RCS fixed the problem. Kinda odd, using RCS for one of the stages but not needing it for the Munar (and Minimus!) lander. But it works, the final problem did indeed turn out to be that I simply needed more control over my turns.

Thanks for all of the advice, and I've posted the final craft and some Minimus screenshots - turns out it works fine for both. And the lander is indeed good for a return trip for Kerbin .

So now I've got a rocket that can reach the Mun and Minimus under the control of MechJeb - and I'm kinda surprised it was a bit more tricky than under manual control. I guess I'm smart enough to keep an "onion" design still while releasing the stages, while MechJeb isn't.

I've learned a lot - about both ship design and how to save fuel. Thanks for all of the advice!