Rocket Stability

[Gibster]

For too long have most of my rockets oscillated or flipped, making them rather a pain to fly into orbit. I want to be able to build larger rockets without dreading the launch (whereas most of my smaller rockets are more stable). Teach me the way of the stable rocket, as I have tried many things, fins (and trying to disable certain control surfaces), tinkering with thrust vectoring, simply adding more reaction control wheals, etc. all with limited success. Some just replace one problem with another, for example I've found that adding fins simply replaces flipping with heavy oscillation. How can I make flying a rocket a joy a again?

[swjr-swis]

I am assuming you've already checked the forum for threads that give basic rocket design 101 tips, mentioning things like checking that center of mass stays above/ahead center of lift/drag even when stages emtpy or decouple, etc (if not, hint hint).

Maybe you are trying too hard? I know, sounds silly, but stick with me. A basic rocket should need very little help to follow a natural east-prograde trajectory. So try to build it with as few fins/reaction wheels/gimbal as you can get away with first, and only add to it if it's clear that what is there is too weak to do what it must.

But it sounds like you have already tried all the tools... so, have you considered your launch profile? Perhaps your chosen trajectory (or the way you control the rocket) is causing more of an issue than the design.

Try to keep the rocket as much as possible on the prograde vector, with as little manual control input as possible, for a very smooth gravity turn and minimal fighting with the atmosphere. Many of the problems you mention are only really a problem when you start pointing outside the circle of the navball prograde marker (causing very harsh and hard to fight drag vectors), or because too much manual input oversteers and then needs correcting, causing more and more oscillation.

I find that a well built rocket is capable of doing a near to perfect and very repeatable gravity turn almost entirely by itself, at most needing a tiny nudge a bit after lift-off to start the prograde slide (a few test launches can quickly show at what speed and how much inclination that one nudge needs to be, but usually at between 50-100m/s nudge to about 80-85 degrees east), and maybe a bit of throttling to keep apoapsis not too far ahead to minimize the need of circularizing (depends a bit on the rocket and staging, but 40-60 seconds ahead of your craft until you reach target tends to produce very nice results). Again, depends on your rocket and staging design, but aim to be at 45 degrees inclination somewhere around 10-15km altitude, 20 degree at around 30-35km, and level with the horizon around 50-55km. There's still a lot of wiggle room with that to account for various rocket designs and sizes.

And watch what happens when the speed indicator switches from surface to orbit: if there is a clearly visible jolt in the prograde marker, it means your gravity turn can be improved - a good gravity turn will hardly twitch the marker at that moment.

[Empiro]

There are many causes of unstable rockets, but the keys are proper design and piloting. I never tinker with thrust vectoring or control surfaces, and I rarely add reaction wheels.

When designing your rocket:

• The top of the rocket should be as aerodynamic as possible. Putting a fairing on could be a huge help
• I always have fins at the bottom of my rockets. They're usually ones with control surfaces
• Don't have a TWR above 1.5 on the pad. Speed will amplify any small aerodynamic issues
• The rocket should be aerodynamically stable. This means to test it by flying it, turning off SAS, and giving it a slight nudge in a direction. It should return to pointing to prograde after a few seconds.

When flying your rocket:

• Make only small course corrections (press Caps Lock for fine controls). Don't ever let your nose get pointed far from your velocity vector
• That means that your gravity turn should be gradual and smooth. I use a 5-10 degree turn at 50 m/s for medium to large rockets
• Don't use SAS too much on big rockets -- it tends to cause oscillations. I usually turn it off after my gravity turn, or set it to prograde and turn it on in short spurts
• If your rocket starts to wobble or bend, don't try to correct it. Instead, turn off SAS if it's on, and throttle back slightly

Edited April 14, 2016 by Empiro

[CSE]

7 minutes ago, Empiro said:

[...] turn off SAS if it's on, and throttle back slightly

This. If your small rockets are more stable but your larger ones tumble even though they look just like the small ones scaled up, there are a few factors I can think of that just creep up on you from KSP 1.25m to 2.5m parts, without  deliberate intent on your part to change your designs:

• the scaled-up rockets might be less stiff, because joint strength is subject to the square-cube law,
• the scaled-up rockets might be going faster, because the Mainsail is just a better lifter engine than the 1.25m engines that you start with,
• the scaled-up rockets might be shorter and fatter, relatively, and have worse aerodynamics: with the more capable 2.5m parts you don't always need as much rocket under your payload to get your target performance,

... and actively managing the throttle during the ascent will tend to help with all of this.

[Gibster]

Hmm, so it sounds like I might be over controlling my rocket, which makes sense because for the first 10,000 meters I'm constantly tapping the A or D keys. That's going to be a hard habit to fight. So what I'm hearing is to make a small 5-10 degree turn around 50 m/s and then let the rocket itself do most of the flying, with only minor course corrections and disabled or prograde SAS.

BTW here's one of my problem rockets: http://imgur.com/a/Oohxz

 

[Empiro]

4 minutes ago, Gibster said:

Hmm, so it sounds like I might be over controlling my rocket, which makes sense because for the first 10,000 meters I'm constantly tapping the A or D keys. That's going to be a hard habit to fight. So what I'm hearing is to make a small 5-10 degree turn around 50 m/s and then let the rocket itself do most of the flying, with only minor course corrections and disabled or prograde SAS.

BTW here's one of my problem rockets: http://imgur.com/a/Oohxz

 

Yep, that sounds like it could be the issue. Your rocket looks good -- you can try to take out the reaction wheel. It might be weakening the joint of your rocket a bit. If you take that out, and directly connect the tank to a larger fairing base, I bet your rocket will be rock solid. The only other minor issue I could see is that the TWR is a tad high on both your first and second stages. The easiest way to fix that would be to add more fuel to the second stage, but that's completely optional -- I think the rocket should fly great either way.

[Gibster]

3 minutes ago, Empiro said:

Yep, that sounds like it could be the issue. Your rocket looks good -- you can try to take out the reaction wheel. It might be weakening the joint of your rocket a bit. If you take that out, and directly connect the tank to a larger fairing base, I bet your rocket will be rock solid. The only other minor issue I could see is that the TWR is a tad high on both your first and second stages. The easiest way to fix that would be to add more fuel to the second stage, but that's completely optional -- I think the rocket should fly great either way.

Ok I'll try that, I just unlocked the 2.5 meter base fairings anyways. Will report back with my launch to Minmus.

[moogoob]

I concur with the others - try the different flight profile you described and let us know.

I've flown weirder rockets well enough.

[the_pazter]

Kerbal joint reinforcement defiantly will get rid of most of the oscillations. and the flipping means, either A: Lack of fins or B: Too Quick, Too Low

[KocLobster]

You've already got some really good advice, and rather than beat a dead horse, I thought I'd add some very minor observations regarding the particular rocket you posted that will help. I would suggest using less fins, that rocket looks like it could get away with just basic fins. Use 4 fins, not 6, and make sure they're on the N, S, E, & W sides of your rocket. Most of the time less is more with fins/reaction wheels/control surfaces. It's hard to say with your particular rocket, but I've found that a lot of times, fairings tend to be counter-productive or just plain unnecessary. You *might* try not using that fairing. Other than that, definitely try altering your ascent profile to the gravity turn methods everyone has already recommended.

8 hours ago, Empiro said:

Your rocket looks good -- you can try to take out the reaction wheel. It might be weakening the joint of your rocket a bit.

+1

[Foxster]

Sometimes I paint myself into a corner when designing a rocket and then struggle to get it not to flip during ascent and the usual methods either don't work or aren't a viable option for some reason. Then you can tackle a common cause of flipping by keeping the fuel mass up-top during the lift.

By default, your craft will often lighten at the top/front, with engines at the rear drawing their fuel from the front first. This shifts the CoM back behind the lift and the craft flips. To fix this you can do something like this...

 

The engines are isolated from the tanks by a decoupler that has staging and crossfeed disabled. Then add a couple of fuel ducts as you can see. This means the fuel flows from the lower tank, through the upper, through the top fuel duct, the main body, the lower fuel duct and finally to the engine. The fuel in the lower tank then being used first, keeping the CoM forward. 

[KocLobster]

@Foxster

God you're impressive...that thing is crazy

[mrclucks]

On 4/14/2016 at 4:31 AM, Gibster said:

Hmm, so it sounds like I might be over controlling my rocket, which makes sense because for the first 10,000 meters I'm constantly tapping the A or D keys. That's going to be a hard habit to fight. So what I'm hearing is to make a small 5-10 degree turn around 50 m/s and then let the rocket itself do most of the flying, with only minor course corrections and disabled or prograde SAS.

BTW here's one of my problem rockets: http://imgur.com/a/Oohxz

 

You might want to swap out the mail sail for the slightly weaker 2.5 m engine(name escapes me) for more DV and such.  If your twr is 1 or a bit below, then add boosters for an initial kick.  If you still need more thrust after that then try adding liquid fuel boosters with reliants on the end.  Once you're at 20km or so, a .8 twr is fine. 

[AeroGav]

Foxster   that's the most Kerbal thing i've seen all day,  nice !

Gibster those pix you shared seem to show the centre of gravity and thrust axis (pink indicator) but where's the blue centre of lift indicator?

Those fins look very small in relation to the size of the rocket.    I'd probably rather four larger ones than six small, you can control pitch and yaw independently.

Also it's very tall and thin.   Should still be possible to get that working but i do find shorter fatter rockets nicer to fly.  More frontal area = aero lossses but sometimes it's worth it for an easier life.

Short  and fat -

[KocLobster]

@AeroGav

It's a little off topic, but I'm curious...what exactly is going on with that rocket? It's cost is over 4 million funds, it's over 300 tons, but it ultimately only has under 5k dV?

[FullMetalMachinist]

2 minutes ago, KocLobster said:

It's cost is over 4 million funds

Looks like it's got at least 12 Vectors, maybe 16, so there's that.

[KocLobster]

3 minutes ago, FullMetalMachinist said:

Looks like it's got at least 12 Vectors, maybe 16, so there's that.

True, but aren't those only roughly 20k funds? That's still 3.5million unaccounted for that I can't see. Maybe I'm just mistaken; maybe he has 4 million funds, and that isn't the cost of the rocket but how much he has?

[FullMetalMachinist]

3 minutes ago, KocLobster said:

True, but aren't those only roughly 20k funds? That's still 3.5million unaccounted for that I can't see. Maybe I'm just mistaken; maybe he has 4 million funds, and that isn't the cost of the rocket but how much he has?

Hmmm, true. And the rocket definitely costs 4 million. Now I'm really curious as well...

[AeroGav]

3 hours ago, KocLobster said:

@AeroGav

It's a little off topic, but I'm curious...what exactly is going on with that rocket? It's cost is over 4 million funds, it's over 300 tons, but it ultimately only has under 5k dV?

Someone wanted a rocket that could put a 30T lander on Minmus.    I don't have his mods so i used ModularFuelTanks to swap the contents of the top Kerbodyne tank to 30 Tonnes of Xenon.    That's probably why.           All I need to do is add one Ox-stat and a Dawn thruster and i'd definitely win the Delta V race.  TWR on the other hand...

[Martian Music]

I feel you. I got to a point where I was so desperate that I started getting designs from real-life rockets and youtubers. I had good designs (mostly like yours) and still it always went wrong. Turns out it's mostly in how you do your ascent. Yeah as others explained, the trick is mostly in your early 5-10 degrees turn, and also your TWR. I miss the old "gravity turn at 10k meters". I'm not sure about this, but more fins makes it worse. Sometimes 2 are enough.

On a side note, KJR really helps so I highly suggest it. No, it shouldn't be considered "cheats" or whatsoever.

Lastly, I'm really bad at ascents so if I said anything wrong, pls dont kill me.

[Hodari]

Especially for 2.5m or larger rockets, the biggest problem seems to be the fact that when you have multiple tanks stacked on top of each other, the top tank is emptied first and then it works its way down.  This causes your center of mass to shift too far back.  Try splitting it up so each tank has its own stage.

[Wemb]

On 17/04/2016 at 2:49 AM, KocLobster said:

@AeroGav

It's a little off topic, but I'm curious...what exactly is going on with that rocket? It's cost is over 4 million funds, it's over 300 tons, but it ultimately only has under 5k dV?

... and a TWR of 5.64? Is that for calibrated for Kerbin, cause if so, that really is overbuilt!

Wemb

[KocLobster]

On 4/17/2016 at 10:14 PM, AeroGav said:

Someone wanted a rocket that could put a 30T lander on Minmus.    I don't have his mods so i used ModularFuelTanks to swap the contents of the top Kerbodyne tank to 30 Tonnes of Xenon.    That's probably why.           All I need to do is add one Ox-stat and a Dawn thruster and i'd definitely win the Delta V race.  TWR on the other hand...

I don't understand though; you could easily accomplish that w/o mods for dramatically cheaper. Why use a dawn and xenon? I really struggle to grasp why you'd design a rocket like this. You could easily achieve your goal for less than 200k.

[AeroGav]

5 hours ago, KocLobster said:

I don't understand though; you could easily accomplish that w/o mods for dramatically cheaper. Why use a dawn and xenon? I really struggle to grasp why you'd design a rocket like this. You could easily achieve your goal for less than 200k.

I'll rephrase this again -

I PUT THE XENON IN THE KERBODYNE TANK IN ORDER TO CREATE A 30 TONNE BALLAST TO SIMULATE THE WEIGHT OF THE LANDER OF A  PERSON WHO WAS HAVING TROUBLE PUTTING HIS 30 TONNE LANDER ON MINMUS.      That user hadn't linked a craft file for his lander and was using mods, including a large nuclear reactor for electricity generation, which I did not have installed.   Creating that Kerbo-Xenon tank was the quickest way for me to create a 30 tonne ballast that was about the same size as his lander, so that I could spend more time designing the launcher itself, rather than first trying reverse engineer his Minmus base from a picture.

I'm sure it's still a rubbish rocket but it's the most powerful i've ever built by some margin.  It was done more to challenge myself than help the OP, other more experienced rocketeers had already posted designs.  I normally concentrate on Spaceplanes.  Pretty sure that person was also in a Sandbox game, so cost was irrelevant.

If I remove the Xenon payload, cost drops from 4 and a half million Kredits to 432,000.   Delta V also goes up to 6700.

Xenon is clearly a VERY expensive propellant.  It's also much heavier than an LF/O tank the same size.

Edited April 18, 2016 by AeroGav

[Hodari]

7 hours ago, AeroGav said:

Xenon is clearly a VERY expensive propellant.  It's also much heavier than an LF/O tank the same size.

And, as I found out when I got bored and stacked something like 100 large tanks of it on the launchpad, it's also very highly explosive, especially for a noble gas...