Why does my ship tumble in atmosphere?

[SpaceK531]

Hello everyone. I have been trying to get back into KSP after a hiatus. I am trying to design a moon fly-by vehicle with room for 4 tourists as a 4-stage rocket. Shown below are the top 3 stages:

 

 

When the fourth stage is discarded and the third stage is ignited, the rocket encounters a violent instability and becomes completely uncontrollable. Eventually, the ship will find settle down if the SAS is set to Orbital Prograde, but by then the ship can't break out of the atmosphere.

The 3rd stage uses 3x reliant and 1x swivel motors.

Is there a smoking gun issue that this design has, or do I need to completely redesign the payload and 3 stages?

[HebaruSan]

Try removing that initial stage in the VAB and then turn on the center of mass marker.

If the CoM is in the bottom half of the rocket, that's the cause of the instability. You can compensate by moving mass around or adding fins.

[SpaceK531]

To be clear, the low CoM is the reason for the instability, and I want it to be as high on the rocket as possible?

 

EDIT: without the payload, the CoM looks like this:

 

Edited January 9, 2018 by SpaceK531

[HebaruSan]

41 minutes ago, SpaceK531 said:

To be clear, the low CoM is the reason for the instability,

There are always exceptions, but yes, this is the most common reason. If your CoM is high up already, then we would have to look at other possibilities.

Quote

and I want it to be as high on the rocket as possible?

Yes. Whether it's darts, arrows, or rockets, the heavy part should be at the front and the draggy part should be at the back for stable atmospheric flight.

Quote

EDIT: without the payload, the CoM looks like this:

 

Keep the payload on, but remove the bottommost stage that works OK. Then you can see how the upper part looks after staging.

Edited January 9, 2018 by HebaruSan

[Plusck]

11 minutes ago, SpaceK531 said:

To be clear, the low CoM is the reason for the instability, and I want it to be as high on the rocket as possible?

 

Yes, precisely.

The crew cabins and lander can are largely just air. Much lighter than a similarly-sized fuel tank.

If it rotates at all, your craft will always spin around its centre of mass. That CoM being quite far back, the instant you go off prograde you have a long lever arm for drag to push on up top, pushing it further off prograde.

The simple solution generally is fins. You put fins on the back, and as long as the atmosphere is thick enough to pose an instability  problem, it will also be thick enough to apply a lot of self-correcting force as soon as you go off prograde. The trouble is that as your ship currently stands, putting fins high enough to help correct the fourth stage (which I assume is going to be a Swivel) might well put them above the CoM for your third stage, which is really really bad.
 

So I'd suggest a redesign. You don't really need that extra Swivel between your lifter stage and upper stage. It's probably causing more of a dead mass penalty than it helps by letting you shed fuel tanks. Get rid of it. Shorten the central section a touch (make it only two FL-T800s long), and add an extra FL-T100 tank to each of the side boosters. Make sure the boosters are feeding fuel to the core stack. And put small steerable fins (I generally use the first elevons in the tech tree) on the bottom of the craft.

The only disadvantage of this setup is that you'll have a longer period of relatively low TWR when you ditch the side boosters. You should probably therefore make a slightly more gentle gravity turn, aiming to get the nose down to 80° at 150 m/s just above the launchpad, so that it drifts down to 45° at about 15km altitude.

[Snark]

@SpaceK531,

As folks have said, your CoM is way too far in the back.  You want it up front.  Those big, lightweight, fluffy crew cabins up front are just killing you.  In addition, your craft has very little control authority-- basically all you've got going for you is the gimbal on the Swivel, and it doesn't have much lever arm to work with since the CoM is so far behind that the engine's not very far behind the CoM.

So, what can you do?  Several things.

• Add fins.  You want the fins to be as low down as you can possibly manage.  Put some on the bottom of those three radial boosters, and also put some around the bottom of your central stack.
• Make sure your fuel drains from the bottom up.  When you have vertically-stacked tanks like that, by default they all drain together, which is not what you want.  Make sure "advanced tweakables" is turned on in the game options, then click the "+1" on each bottom-most tank in each vertical stack.  That will make the bottom tank drain before the one above it.  Why does this help?  Because it rapidly moves the CoM upwards as you drain fuel, which will help your stability.
• Consider 4-way symmetry instead of 3-way.  Three-way symmetery can be a bit flaky for stability, sometimes.
• Asparagus, if you're not already.  It's kind of hard for me to tell, but you've got 3 Reliants and one Swivel there at the bottom.  Are those all one stage?  If they are... don't.  Mount the three radial stacks on radial decouplers, and enable crossfeed on the decouplers.  This will have the result that the radial stacks drain before the central one does, so you can stage them away as soon as their tanks are empty.  Not only will that help your dV, but it means you can hang on to that Swivel-and-its-gimbal for longer.
• Make sure you stick to all the way up.  What's your flight path?  Are you going straight up and them manually muscling it around eastwards?  If you are... .don't.  Start a gravity turn to the east practically right off the pad, and as soon as you build up just a little bit of speed, click the "hold " button for SAS.  That'll do a much better job than your poor human reflexes can at holding you straight prograde, which means even if you do have a little bit of inherent aerodynamic instability, your engine gimbal can (hopefully) keep up with it.

 

[Aegolius13]

Ditto what other people said, but you might also want to revisit that lander can.  It is, well, about as draggy as it looks given how it sticks out from a 1.25m stack (in other words, pretty draggy).  That's probably not the root cause of your issue, since at least it's sorta near the back of the stage in question, but I doubt it's helping things either.  And that drag is probably hurting your fuel efficiency as well. 

A promising redesign might be to replace the lander can with an Mk 1 command pod at the very top of the rocket (which would also allow you to sub out the nosecone).  Then put a small nosecone or parachute on top of the pod.  

Yeah, the command pod is a little heavier, but the difference is small, and having that extra mass up top would help your stability issue. The command pod has other perks as well - it's quite a bit cheaper, has more reaction wheel torque, etc.    I find it's a better part in most cases, but almost always when the part is going to be exposed to atmosphere, as here.  

 

[SpaceK531]

Thanks for the help everyone. I've been away long enough I didn't know about using SAS prograde for the gravity turn.

[foamyesque]

I'm not seeing any reason for the radially-mounted girders. What's the story there?

[Snark]

21 hours ago, foamyesque said:

I'm not seeing any reason for the radially-mounted girders. What's the story there?

If you look at the original screenshot up near the top of the thread, he's using them as attachment points for struts that reach up higher in the stack, presumably to stiffen the long skinny rocket so it won't go all floppy-noodle on him.

There are definitely better solutions for that problem out there ...but that's pretty clearly what he's after.