How do I stop my rockets from precessing?

[Trekkin]

This happens every single time I fly anything larger than a Mun rocket: Once I get to the upper atmosphere, almost totally ex nihilo, my rocket will start spinning around its z axis with considerable angular acceleration. Then, at a rate not normally considered normal for precession, it starts doing exactly that, until I either end up in a flat spin or have my rocket ripped apart. ASAS doesn't help, fins don't help, RCS and a boatload of nozzles do not help; it's like the rocket just decides it wants to pretend to be a top for a moment, and damn any attempts on my part to control it.

Is there some logical reason for rockets to precess so violently at high speeds when the centers of mass and thrust line up perfectly, and is there some way for me to halt the process, either in design or in flight? I can't get my design for a Duna land-and-return craft small enough for this issue not to pop up.

[SiliconPyro]

I have this same problem on some of my designs. I haven't managed to isolate what design element causes it so I'm very interested in a solution. It's maddening to nearly reach orbit with plenty of fuel and to have your craft tear itself apart.

[EndlessWaves]

my rocket will start spinning around its z axis with considerable angular acceleration. Then, at a rate not normally considered normal for precession, it starts doing exactly that, until I either end up in a flat spin or have my rocket ripped apart.

By a 'flat spin' do you mean with the nose pointing towards the horizon?

If so then I wonder if it's a balance issue, more drag forward of the centre of mass than behind it for example.

A picture of one of your troublesome rockets would probably help as there are several things that can be causing this. If you're flying with mechjeb then try without it as mechjeb doesn't handle unstable rockets very well and you can often control a rocket manually that suffers from precession with mechjeb (although it sounds like it's fairly violent so is more likely to be a design issue).

[phoenix_ca]

Yup, I've had the same issue. It's...disgruntling. Worse, I've been able to get a rocket that fell apart in flight because of this odd behaviour into orbit in a successive launch.

One might try a true bucketload of RCS thrusters. Perhaps the Kosmos versions because they are hilariously overpowered.

[Wayfare]

Uncontrolled rolling is caused by asymmetry. The hard part is figuring out what, ecaxtly, is asymmetrical on your rocket. Common causes I've found on my rockets:

- Uneven struts causing boosters (whether solid or liquid) to torque to one side or another. Be aware that fuel lines also function as struts!

- Uneven thrust due to poor staging design. Make sure the center of thrust remains in line with the center of mass throughout the stages.

- Uneven fuel consumption due to poor fuel line routing. If one side's tanks drain faster than the opposite side's, you're shifting your center of mass.

[Trekkin]

By a 'flat spin' do you mean with the nose pointing towards the horizon?

I mean it tends to rotate purely in the xy plane, taking z as my original direction of travel. Since I'm usually into my gravity turn by then, I end up drawing a circle passing through the horizon. And then ripping my rocket apart and dying.

Here, though, is the rocket: http://i.imgur.com/3dVmt.jpg (linked for size). It's designed around asparagus stalk staging, first with the big towers of tanks, then with the little ones. It usually falls apart once those central 8 stacks lose the first pair of stacks.

And I can't really not use mechjeb, since the lag on launch makes manual control impossible.

[nhnifong]

Turn the ASAS off and on after each decouple or this will happen even to completely symmetrical rockets. It has to "re-calibrate" or something.

[Wayfare]

Here, though, is the rocket: http://i.imgur.com/3dVmt.jpg (linked for size).

Mother of God...

Your rocket might be a bit too big. For, well, anything really. It seems like you're using a lot of rocket to push up a lot of rocket just to push up a lot of rocket. How much of that is supposed to make it to orbit?

The thing about big rockets is that they tend to have both a lot of thrust and a really long distance between the center of thrust and the center of gravity. Now, in theory, this is a good thing. Like the old "balancing a broom on your fingertip with the brush up"-analogue. But what tends to happen is that any course correction overshoots its mark by so much that the counter-correction just makes things worse, and things go downhill from there. The high thrust also multiplies any minor asymmetry issues your design might have.

I know Jeb will be displeased but I think you need a smaller boat.

[Trekkin]

Mother of God...

Your rocket might be a bit too big. For, well, anything really. It seems like you're using a lot of rocket to push up a lot of rocket just to push up a lot of rocket. How much of that is supposed to make it to orbit?

Of the 1250 or so tons at launch, about 120 or so. I spent a while cutting down on engines until I was near the minimum. Back to cutting down, I suppose.

The thing is, going from Kerbin to Duna and back in one go via direct ascent needs a ludicrous rocket, in my estimation of the delta-v involved.

I had thought that since my 33% larger rocket (636 parts!) made it to orbit ok, the size might not be the issue, but I'll see if I can minimize more.

[Bluejayek]

When rockets get sufficiently big, turn off asas for the launch to orbit and do it manually. The coding that seems to exist for the ASAS is hilariously awful in the way wayfare described. It constantly overcorrects and loses control, and even when it doesnt, it initiates these constant vibrations that can easily be damped out with manual control.

[Wayfare]

A 10% launch to orbit mass effiency is not bad. The question then becomes whether you need 120 tons to make it to Duna and back.

Start from the top. You'll notice that a lighter lander stage (which might even drop some tanks to be lighter for the return stage) requires a lighter interplanetary stage, which requires an even lighter orbital stage, which requires an even lighter launch stage. I don't know if the effect is exponential or not (guess it might be) but it shakes down your rocket fast.

[Trekkin]

A 10% launch to orbit mass effiency is not bad. The question then becomes whether you need 120 tons to make it to Duna and back.

Start from the top. You'll notice that a lighter lander stage (which might even drop some tanks to be lighter for the return stage) requires a lighter interplanetary stage, which requires an even lighter orbital stage, which requires an even lighter launch stage. I don't know if the effect is exponential or not (guess it might be) but it shakes down your rocket fast.

Well, what I did was to economize by removing my kerbin-to-duna nuclear engines, and using my one duna-to-kerbin nuclear engine for both trips. That cut down on enough mass that I can get by with only 16 tank stacks, which seems to have solved my precession problem. It looks like the root cause was the engines spreading out from each other under the thrust.

Incidentally, what's Duna's gravity relative to Kerbin's?

Edited October 11, 2012 by Trekkin

[Wayfare]

About 0.3g (1/3rd of Kerbin) at sea level according to the wiki.

And, congratulations on fixing your rocket!

[EndlessWaves]

Twice that of The Mun.

[TheDarkStar]

Duna has just little enough gravity to allow for a Kerbal's RCS thrusters to work.

[MaverickSawyer]

That, sir, was what I call a derp-rocket, because things go REALLY wrong when you try to launch. Facepalms typically are the result. Glad to see you solved the issue. If not, using some fins and using snap to angles would likely clear up many of the issues. Making sure your parts attached to radial decouplers are EXACTLY centered is critical, or else you have all the force of the ejection NOT going straight in: it adds a very slight rotational component, and that screws up the launch. I've had my share of those launches...