Wobble. Where does it come from?

[jros83]

Let's assume a smaller rocket with minimal contact points, no struts, symmetrical in both aesthetic and weight distribution, and with no flat surfaces fighting against atmosphere.

In such a circumstance, what causes wobble, especially from lift-off to breaking into upper atmosphere? If airflow is the same at any given point and surfaces are aerodynamic, I could only assume vibration?

I ask because, whereas using a pilot this can be easily remedied, I like to do small missions early in Career with the Stayputnik, and typically I have no SAS module at this point, or if I do, it's the first gen SAS disc, which still rarely helps with wobble.

Most of the time, I end up with a vehicle that tips past critical, or (most commonly) one that starts spinning after separating the first stage, and/or upon engaging second stage engine. I either lose control completely, or spend precious time and fuel trying to right the thing that I do not have sufficient fuel and momentum to complete my goal. Also, when not using a human, I mean, Kerbin, pilot, how can you finely tune movement to correct errors when you can't engage SAS? Because another issue I often face is finding myself nearing the 70km border to space trying desperately to get the thing in a stable orientation, and often failing, because no matter how lightly I tap any direction, it seems the slightest touch is still overcompensating, and I end up in an overcompensation cascade, that is 9 times out of 10 irrecoverable.

Edited December 3, 2015 by jros83

[kiwi1960]

I usually use fins or wings to keep it steady... that is, till it hits orbit... then its either stable still... or not at all.

 

I hate that happening.

[jros83]

I use fins as well. Doesn't always work. Assume fins are included in my example.

[RandomRyan]

is... um i mean... aaare youre rockets tall? if they're tall, woble happens.

they are a few things that causes wobble.

1.tall rockets (i blame the physics)

2.players mistakes

3.SAS.(sometimes SAS becomes unSAS)

4.the design itself

5.Size 3 are (sometimes) wobbly when attached to a size 2

6.etc,etc

to overcome wobble,

1.struts

2.add fins to downer areas of the rocket. never put on top! causes instability.

3.SAS (since you have fins)

4.do simple designs.Yet, sophisticated

5.etc,etc *

i dont know much, but i would be honoured to see youre design

All the Best!and dont explode.

*but wait! theres more! master the grav turn!

 

[jros83]

8 minutes ago, RandomRyan said:

is... um i mean... aaare youre rockets tall? if they're tall, woble happens.

they are a few things that causes wobble.

1.tall rockets (i blame the physics)

2.players mistakes

3.SAS.(sometimes SAS becomes unSAS)

4.the design itself

5.Size 3 are (sometimes) wobbly when attached to a size 2

6.etc,etc

to overcome wobble,

1.struts

2.add fins to downer areas of the rocket. never put on top! causes instability.

3.SAS (since you have fins)

4.do simple designs.Yet, sophisticated

5.etc,etc *

i dont know much, but i would be honoured to see youre design

All the Best!and dont explode.

*but wait! theres more! master the grav turn!

 

1. Tall or short, doesn't matter.

2. Not player mistake, because this happens with minimal to no player input as well.

3. SAS turned on or off, SAS module attached or not attached. Makes no difference. Unless it's a physical pilot's SAS skill. But if you read my original post you'll see I already touched on that.

4. Once again... please reread my post.

5. ONCE AGAIN PLEASE REREAD MY POST and do understand that size connected to equal size is implied in my comment about symmetry.

6. "etc. etc."

to overcome:

1. Makes no difference. Still happens with struts. Unless you care to elaborate?

2. Already discussed fins, and please assume I know enough about placement...

3. We've already discussed SAS...

4. This is an empty platitude...

5. "etc. etc." again...

By the way, the word is "your" not "youre." And "you're" requires an apostrophe.

As for grav turn, that has no bearing on the topic.

I'm sorry, but your reply was borderline trollish, blatantly condescending, and not helpful in the least.

Edited December 3, 2015 by jros83

[Superfluous J]

We can help you troubleshoot a specific rocket (if we could see it). With a general "my rockets wobble" we can't do very much but give generalities.

The Stayputnik is not purposed to fly a rocket off the launchpad into space. Some can do it. I could, but not consistently and don't enjoy it enough (at all) to put the work into gaining that skill.

If your rocket is long, it could easily be becoming back-heavy, which makes it want to flip. If it's short, it could start out that way which makes it want to flip. Fins help, but nothing will keep a rocket that is sufficiently back-heavy from flipping around.

Regarding symmetry, it essentially doesn't matter in this case. You're going to get tiny deviations from mathematical perfection and wobble will tend to get more and more as you fly, gain speed, and fight air.

Regarding your above reply, we can't assume anything about you. All we know is that you're having some problem with wobble. If we are to assume you know about fin placement, are we to assume you know everything else? That can't be true else you'd not be having this problem. So, how are we to know what is okay to assume and what is not?

Edited December 3, 2015 by 5thHorseman

[kiwi1960]

I have noticed that if you use MechJeb, you can immediately after lift off help matters by pressing R and mechjeb will use RCS to help steady the craft.

However, as you hit orbit, the RCS goes mad because the target constantly moves and never stays steady.

 

In such situations, I turn *R*CS off ...

 

 

[jros83]

Oddly enough the simpler the design the more I get critical wobbling. Larger and more complicated, less wobble. I don't use MechJeb. I'll just keep plugging away at it.

[RandomRyan]

3 hours ago, jros83 said:

Let's assume a smaller rocket with minimal contact points, no struts, symmetrical in both aesthetic and weight distribution, and with no flat surfaces fighting against atmosphere.

In such a circumstance, what causes wobble, especially from lift-off to breaking into upper atmosphere? If airflow is the same at any given point and surfaces are aerodynamic, I could only assume vibration?

I ask because, whereas using a pilot this can be easily remedied, I like to do small missions early in Career with the Stayputnik, and typically I have no SAS module at this point, or if I do, it's the first gen SAS disc, which still rarely helps with wobble.

Most of the time, I end up with a vehicle that tips past critical, or (most commonly) one that starts spinning after separating the first stage, and/or upon engaging second stage engine. I either lose control completely, or spend precious time and fuel trying to right the thing that I do not have sufficient fuel and momentum to complete my goal. Also, when not using a human, I mean, Kerbin, pilot, how can you finely tune movement to correct errors when you can't engage SAS? Because another issue I often face is finding myself nearing the 70km border to space trying desperately to get the thing in a stable orientation, and often failing, because no matter how lightly I tap any direction, it seems the slightest touch is still overcompensating, and I end up in an overcompensation cascade, that is 9 times out of 10 irrecoverable.

ok. i reread the post,and the reply. forgive me for not being so helpful

"how can you finely tune movement to correct errors when you can't engage SAS?"

okay. press the Caps Lock button on your keyboard. you can see that the pitch, yaw and roll arrows turn to blue.

this will activate fine controls instead normal controls.

you end up with a vehicle that tips past critical. either you lose CTRL completely, or spend precious time and fuel trying to right the thing.

from my experiences, my rockets always (but not all of them) spin out of control. especially my soyuz rocket.

heres my tactic. its called; Go With The Flow maneuver

follow the way your rockets are spinning.cut the throttle.once it starts to slow down,(not the vessel, the spin) activate the boosters.

although it has 50/50 chance to work, i say try it in sandbox mode.

 

[Temstar]

The reason for this is the same both in KSP as well as real life: a rocket is a dynamically unstable system.

Nothing in real life is in perfect mathematical precision. Even the most precisely built machines like rockets have imperfections. The surface of the rocket is not perfectly smooth, the fins are not perfectly symmetrical, the center if thrust moves around and is not always perfectly inline with CoM.

In a dynamically unstable system any of these tiny deviation from the perfect model causes the deviation to build up until you end up with a hugh deviation, eg rocket flipping. To get a dynamically unstable system to stay in equalibrium  (eg rocket staying on course) it must be actively controlled so that small deviation are constantly compensated for.

KSP doesn't need to go out of its way to make sure in game rockets are dynamically unstable - the very fact that things like velocity, acceleration are stored as floating point numbers in finite blocks of memory mean calculations cannot be done to arbitrary levels of precision, and so such loss of precision in calculation results in instability - ie dynamically unstable.

[bonyetty]

A fuel line from the top tank directly to the engine is a way of keeping the center of mass forward on a rocket thus reducing flipping. For long rockets run the fuel line through a cubic octag attached to a panel (no fuel flow part) somewhere on the stage between top tank and engine. Feels a bit cheaty for my personal career game play though. 

Edit. oh also place a panel between engine and bottom tank possibly for 1.0.5. 

Edited December 3, 2015 by bonyetty

[sal_vager]

Moved to gameplay questions.

 

Connections between parts and their strengths are handled by the physics engine, parts are not rigidly defined as being exactly xyz from their parent at all times as would occur in games where no or limited physics apply.

[jros83]

5 hours ago, Temstar said:

This should not be quote box. The new forum won't let me edit it out. Please disregard.

Thanks all for the replies.

Edited December 3, 2015 by jros83

[cephalo]

Flipping is one thing, wobble is another. When you say 'wobble' I think you mean that your craft just shakes apart, even in space. I failed my first attempt at the Jool 5 challenge because of this. All my components were pie in the sky and tested to completion, but when glued together with big docking ports, I could NOT get them to Jool. The floppy joints couldn't handle the escape burn.

This happens because every craft has some small amount of torque, when you initiate a burn there will be some drifting off target that will need to be corrected. When a correction happens, either manually or with SAS, the craft will try to flex. If it is not rigid, it will bend and then try to straighten out. As it does this, different parts of the ship are acting on each other, changing the perception of the crafts direction of movement and making it less accurate. The front end throws the back end, and the back end retaliates by throwing the front end. Using SAS of course will shake the ship apart as it tries to compensate for these exaggerated errors. Even if you are not using SAS, your indicators will show a drift that will not represent your total craft, as the different parts are not moving together. It's impossible then to really know what you should do with the controls.

You can stop wobble by making your craft more rigid with struts etc. or by not caring where your craft is going. If you just burn and let it drift, you won't know where you're going, but at least it won't wobble!

[jros83]

Not shaking apart. Wobbling to the point of not being able to get it on course again.

[cephalo]

9 minutes ago, jros83 said:

Not shaking apart. Wobbling to the point of not being able to get it on course again.

Heheh, I think I know what you mean. I suppose if your craft is light enough, you can have a situation where you can't actually shake it apart without an honest effort, but the problem is the same. Your front doesn't know what your back end is doing, and your indicators are way off as to your direction of travel. In theory it should be possible to correct for such things, but the game currently only cares about the command part in terms of reporting the ships state.

[Snark]

Just to be clear, my understanding is this:  When you say "wobble" you mean "can't hold prograde without SAS", and this is specifically a problem you have in atmosphere, yes?  If you mean something different, then the following answer will be unhelpful. 

It boils down to this:  is your rocket's CoM in front of or behind its center of drag?

If it's in front, then you're aerodynamically stable.  Your rocket wants to point straight forward, like a lawn dart or a badminton birdie.  You'll have no wobble problems at all.

If it's behind, then you're aerodynamically unstable when you're pointing forward.  Your rocket wants to point backward.  It's like trying to throw a badminton birdie backwards.  Even the slightest deviation from perfectly straight forward will rapidly magnify itself until you flip around.

The reason so many people run into this is because one of the heaviest components in a rocket is the engine, which is down at the bottom.  This is often compounded in early career game because you haven't unlocked the taller fuel tanks yet, so you build a stack of smaller tanks, which is bad because your engine drains the topmost tanks first, which moves your CoM downward.

The answer is the usual-- put fins on the bottom of the rocket, as far down as you can manage.  Streamline the front, if you can.  If you have a stack of liquid fuel tanks, consider disabling the topmost tank or two, at least until the bottom tank is burned-through; that will help keep your CoM high.  Anything you can do to move your CoM forward is good.  Starting with an SRB is useful because it doesn't shift its CoM as it burns (in fact, your rocket's CoM will move upward as the SRB burns, which is good), and also an SRB is relatively light relative to the amount of thrust it provides.

Edited December 3, 2015 by Snark

[AeroGav]

I've seen something I call hysteresis (a big word, not sure I'm using the right technical term) that've seen when using SAS on my own space planes and in youtube videos of other people's rockets.  Basically SAS continually overcorrecting and having to compensate for last correction.  It's average trajectory does match prograde. but it constantly jerks either side of it with the control surfaces flipping back and forth like mad.  Creates drag, wastes electric power, bleeds your RCS dry before you get anywhere.  Saw a mod that helps you tune that out?  Can't remember that name.

 

I'm guessing the physics / rounding errors described in previous posts apply to aircraft too.   I've never built a completely stable aircraft no matter how hard I try.    Centre of mass well forward, check, mahoosive tail fin, check,  gull wing design with dihedral on outer sections - check,  wing attached to top of fuselage so centre of mass below wings - check.  This trainer plane felt really easy to fly compared to anything else I've flown in KSP (stock or self created), but on one of those Kerbin observation missions, involving flying halfway round the planet, I noticed I still couldn't let go the joystick completely for long periods, or use time acceleration comfortably.    The problem was roll axis - let go the stick in a 1degree left bank, and despite the dihedral, high wing design etc, it gradually rolls further and further (about 1 degree per second) into the low wing, rather than righting itself, like such a design surely should?