unstoppable rotation

[AlamoVampire]

Okay, here is the thing. I wanted to see if I could take a NovaPunch 2.5m fuel tank, strap some 1.5m motors on the side and launch it on its side. Simple in theory, but in practice? Well, its rotating uncontrollably. I have the engines spaced as evenly as I can, I have it set to unbroken joints and unlimited fuel. I tried this with fuel consumption and same effect... What I cannot figure out is: I have them placed directly on the side, down the middle so to speak <snap angle ON> and have a probe body on each end. Yet it rotates as soon as the docking clamps let go.

Here it is:

As you can see, the fuel does not change, so no CoM is changing.

Javascript is disabled. View full album

Edited April 1, 2014 by AlamoVampire

[Temstar]

No rocket can be constructed perfectly. Inevitably, the engine's direction of thrust will not be perfectly aligned with the rocket's center of mass, so the rocket will have an inbuilt tendency to turn. In order for any rocket to flight straight it either needs to be aerodynamically stable or have active guidance.

In this case in particular, KSP has a lot of problem flying a vehicle where the direction of the thrust is not aligned with the orientation of the controlling part. Notice how on the ground your rocket's navball indicates the heading as horizontal where as your thrust is actually upwards? KSP can't deal with that situation EVEN IF you use active guidance from that orientation.

[Majorjim!]

Moar reaction wheels!

[MKI]

Make the core in the SPH. The two side sysmetry SHOULD be able to make the engines spaced PERFECTLY. As long as you have an even number of engines (this case you do)

Then save the core(the fuel part with engines) in subassembly

Make the actual ship in the VAB and start with a probe core, and attatched the fuel/rocket sub and attatch another AI core to the other end.

This SHOULD make your design perfectly symmetrical. If not then add more SAS to balance out the small differences.

EDIT:

i just noticed you have two fuel tanks instead of a single one. I suggest choosing a design that is symmetrical to begin with. So if possible make the same design, but with two medium sized tanks if you can instead of a small and large. This way the COM is inbetween and you can use the SPH symmetry to help.

Edited March 31, 2014 by MKI

[AlamoVampire]

Make the core in the SPH. The two side sysmetry SHOULD be able to make the engines spaced PERFECTLY. As long as you have an even number of engines (this case you do)

Then save the core(the fuel part with engines) in subassembly

Make the actual ship in the VAB and start with a probe core, and attatched the fuel/rocket sub and attatch another AI core to the other end.

This SHOULD make your design perfectly symmetrical. If not then add more SAS to balance out the small differences.

EDIT:

i just noticed you have two fuel tanks instead of a single one. I suggest choosing a design that is symmetrical to begin with. So if possible make the same design, but with two medium sized tanks if you can instead of a small and large. This way the COM is inbetween and you can use the SPH symmetry to help.

The stupid thing did the rotation even as 1 single tank. Ill try the sph idea later today but for now, i need sleep lol

[MKI]

The stupid thing did the rotation even as 1 single tank. Ill try the sph idea later today but for now, i need sleep lol

The SPH editing is more important, as it should evenly distribute the engines. I personally don't know if the modded parts would have the same mass to keep the COM where the Center of parts are. And idk if your fuel would flow correctly between those two parts.

Goodluck!

[Kialar]

Maybe you should lock the gimbal of all rockets but the two in the middle. I think it can happen because of counter gimbaling (OMG thats a kind of neologism..., but I hope you get my idea).

Additionally, try to keep acceleration below 3g. That often helps in controlling wild rotation in random directions.

-Kia

[SRV Ron]

The problem is how fuel is drained from the tank. The results, a fuel tank on its side tank gets lighter on one end then the other resulting in the uncontrolled rotation.

[Kasuha]

Hard to say if it is bug or just caused by slightly uneven spacing of engines but it should be possible to remedy it by slight tweak to thrust limiter of one of engines.

[SSSPutnik]

I'd use two tanks, joined like a mirror. Fuel should empty evenly and yes, place the engines in the SPH. SAS unit should counter any minor differences.

[Claw]

I see your problem is different, but in case it helps... I did an experiment in a previous thread (can't find it at the moment). It turns out that sometimes, even with perfect symmetry (4 engines in my case), some ships will rotate. Possibly due to some bug, but I don't know for sure what the problem is.

I built a cylindrical rocket in the VAB and put 4x symmetric engines on it. Some engines, such as LVT-30s and 45s, caused rotation while other engines did not. Even more bizzare, if I rotated the engine itself (and nothing else on the rocket) the rotation behaved differently. This was especially noticable when pitching/yawing the rocket.

So you could perhaps try different engine types or try rotating the engines in 90 deg increments. Adding more reaction wheels may help, but aerodynamic controls might not help based on how it looks like you will fly it.

If I find that thread later I'll link to it. I think I wrote which engines did not cause rotation and which orientation worked best for the 30s vs 45s (the angle was different).

Not sure if that'll help.

Edited March 31, 2014 by Claw

[SSSPutnik]

Yes I've found aligning the engines helps. Usually the skin on the part has some decoration that helps with this. A pipe for instance. Rotate so they all face inwards, or outwards is fine but if you have them like so, you can get rotation.

((

((

What seems to work is

()

()

[Kerba Fett]

Angle all of your engines outward a bit. It will make it more stable. Start with 5 degrees and go from there.

What's happening is that as soon as it starts to pitch forward it falls off the column of thrust it's balancing on. By angling the engines, the ones on the low side become more efficient while the ones on the high side lose efficiency and provide more lateral thrust to make you move. (which is why you pitch or roll in the first place)

You'll also need to use at least 2 large reaction wheels to provide the torque to control it.

[cantab]

Well I just tried a sideways orange tank, two mainsails underneath, two large SASes on each end, QBE probe core on top. Built in the VAB. Launches were at 50% thrust limiter (because TWR was way too high otherwise) and full throttle all the way.

Note that the orange tank makes equal spacing easy with the two silver bands near each end.

Hands off, no SAS: Veers consistently to the east and spins into the ground.

Hands off, SAS: Initially veers east but swiftly corrected. Climbs vertically under control.

Hands off, SAS, no torque: Initial east veer takes slightly longer to correct. Climbs near-vertically, a touch to the east with a little roll but stays under control.

Hands off, SAS, no gimballing: Similar. Perhaps a little stronger veering initially.

Hands off, SAS, no gimballing, no reaction wheel torque just the probe core: Uncontrollable.

Hands on flying: Handles fine on just torque, just gimballing, and both.

The tank inclined at launch: Flies straight and true.

Shifting one engine inboard by about the width of a cubic octagonal strut or two: Completely uncontrollable spin.

Same shift, with the more outboard engine set to a 46.5% rather than 50% thrust limit: Smooth ascent.

The more outboard engine on a 46 or 47% thrust limit: Leaving the SAS to its own devices, loss of control sooner or later. I didn't try hands-on flying for these tests.

To conclude, your problem is almost certainly asymmetric engine placement. It doesn't take much of a difference to make things spiral out of control. Fixing it by thrust limiters is tricky, fixing the engine placement is better. If the engines you're using don't gimbal, you may also have insufficient control authority, so try adding reaction wheels.

CoM shifting as a single tank drains is not the problem, at least not with a Jumbo 64 it isn't. With multiple tanks as you have though it may be, but your infinite fuel tests show that's not your only issue at least.

Edited March 31, 2014 by cantab

[SSSPutnik]

Nice work, saves a lot of testing.

[m4v]

Look! no hands!

Okay, I cheated, but really, without a plugin I can't imagine getting that thing balanced.

[ScottyDoesKnow]

Wings made it work for me. Here's an image with just SAS on, mechjeb not in use:

Here were the steps I took:

2 gimbaling engines on outside, 4 non on inside, controllable surfaces on all 4 sides: out of control immediately, even SAS made the surfaces on the sides always point in one direction and flip the ship

same engine setup, controllable surfaces on ends: out of control

all gimbaling engines, no more controllable surfaces: out of control (I think)

all gimbaling engines, 2 large ASAS on the ends: in control

all gimbaling engines, no large ASAS, wings: more in control

It tilts slightly at the start but corrects and stays straight (with SAS). Mechjeb does not do well with it though.

All in all there's definitely something strange going on here. I don't know why the controllable surfaces on the side went crazy under SAS, and I don't know why mechjeb can't fly it.

[AlamoVampire]

wow, just wow! I am completely blown away! My issue cropped up last night when I decided to just screw around see what could be done, and it presented me this rotational issue. I came here, thought I would get 1 or 2 replies, but to see the effort each of you guys went through to help make a flight of fancy fly, well, I am humbled. I want to thank each of you! I present you fine fine FINE folks a fruits of your help, I have only 2 pictures, but, they say more than I can!

The first here, is the craft on ascent, HUD is shown 2 2.5m reaction wheels on each end + an ASAS unit as well on each end. the two central engines are gimbal locked and turned 180 degrees from the outer two. The 2nd is just a close up glamour shot of the craft a few seconds later!

@scottydoesknow I am not sure whats going on either, but, I know the nav ball with this thing cannot be trusted. I aimed at the prograde and wound up changing inclination, aimed retro and added altitude and yet ANOTHER inclination change. All in all, this is one odd craft, but, still, was fun to try and fun to see all yalls help

should be noted, the central portion <tank/engines> was as I think MKI suggested maded in SPH and saved as a subassembly

Edited April 1, 2014 by AlamoVampire

[Depth]

Looking at the pictures in this thread I was thinking to myself "why would anyone want to do this", but then it hit me.

You should add those contraptions to radial connectors/detachers and use them as boosters

[AlamoVampire]

Looking at the pictures in this thread I was thinking to myself "why would anyone want to do this", but then it hit me.

You should add those contraptions to radial connectors/detachers and use them as boosters

my thought was: can i?

[lammatt]

add wings and control surfaces and a coulpe of reaction wheels if the asymmetric thrusting isnt that bad

[SSSPutnik]

This type of setup is also nice for landers... Your cockpit is easy to place low to the ground.

[ScottyDoesKnow]

@scottydoesknow I am not sure whats going on either, but, I know the nav ball with this thing cannot be trusted. I aimed at the prograde and wound up changing inclination, aimed retro and added altitude and yet ANOTHER inclination change. All in all, this is one odd craft, but, still, was fun to try and fun to see all yalls help

You'll notice in my screenshot that I have a probe on the top that I'm controlling from. On yours you can't trust any of the navball markers because your probe body is facing in a different direction from your thrust.

[cantab]

All in all there's definitely something strange going on here. I don't know why the controllable surfaces on the side went crazy under SAS, and I don't know why mechjeb can't fly it.

I know SAS can have problems with oddly positioned engines and control surfaces. IIRC it can usually handle winglets or RCS forward of the centre of mass, but not engines - it gimbals them the wrong way. Of course the designs here have engines aft of the CoM, albeit not very far aft.

[Depth]

my thought was: can i?

Looks like you can