Spin Stabilization Question

[Wijbrandus]

I have a couple of questions about the modules that are supposed to spin your rocket and supposedly give it control.

1) First, the capsule already has some torque available to it. So when should I use these modules? How do I know if I need one, and how do I know when I need more than one?

2) Where should these be placed in the stack for best effect?

2) All three of the modules (Inline Reaction Wheel, Inline Advanced Stabilizer, and Advanced S.A.S Module) give the same amount of torque. So what is the point of using one over another?

Thanks! I hope to learn something in this thread that will make my Kerbal-naughts survive longer!

Edited February 12, 2014 by Wijbrandus

[Taki117]

I think you misunderstand what Reaction Wheels actually do.

1) You should use the SAS modules when you need more torque, usually with large and heavy craft.

2) Anywhere is fine. The effect is cumulative

3) the IRW and IAS are the same, but the ASAS is a 2.5m version.

[relin]

1.) The capsules provide some torque, but as your craft grow in mass, it won't be enough to maneuver effectively. It's one of those things you'll notice during testing: you're in orbit and attempting a retrograde burn, but it takes two minutes for your craft to rotate. You can survive with low torque, of course, but depending on your comfort level, you may want more.

2.) For maximum effect, put them as far away from the center of gravity as possible.

3.) The Advanced SAS Module is simply a larger-diameter version of the IRW, and the IAS isn't (currently) any different than the IRW.

[Wijbrandus]

Ok, I completely misunderstood their purpose. I thought they were used to spin the rocket as it ascended through the atmosphere. Sometimes my rockets do that, and sometimes they don't, and I came here trying to discover the correlation.

So torque is used entirely like an RCS, although differently - it's used to move the rocket to point it in a specific direction, and not spin it in terms of stabilization.

In that case, would more torque assist my rockets in ascent when I have a hard time keeping them pointed in the wrong direction?

[Kasuha]

Both RCS and torque can be used to turn your ship but RCS uses monopropellant and you may run out of it. It's better to reserve monopropellant (and RCS) for maneuvers that cannot be done with torque as usually there's endless supply of electricity.

Pod torque is good for small ships (e.g. lander which can go from orbit to surface and back). Larger and heavy things usually need dedicated torque modules.

Also torque can be use to fight the rocket spinning. Although if it spins substantially you probably need to use more struts or different design to make it sturdier. Note that while you can turn the ship by thrust and engine gimbal, you can't affect its roll that way - you need RCS or torque for that.

[Wijbrandus]

Ok thanks!

It's amazing. Even after playing for so long, I'm still learning things.

[KvickFlygarn87]

2.) For maximum effect, put them as far away from the center of gravity as possible.

Wrong. Well, at least in KSP. IRL, it'd be right, but for KSP, doesn't matter for reaction wheels. True for RCS though :-)

[Streetwind]

Right. I have no seen people claim on these forums that:

- Reaction wheels are best as far away from the center of mass as possible

- Reaction wheels are best exactly on the center of mass

- Reaction wheels don't care where they are in relation to center of mass

Clearly there is a very large uncertainty in the community as to this topic, and most people are just guessing blindly or repeating what they have seen somewhere else. We need a proper test for this.

[Wijbrandus]

I think that would be very hard to test for. As I read up more on what reaction wheels are for, logic would put them as far away from the center of mass and on the opposite end of the source of thrust.

Maybe with a very long rocket, it could be determined.

[lajoswinkler]

I have a couple of questions about the modules that are supposed to spin your rocket and supposedly give it control.

1) First, the capsule already has some torque available to it. So when should I use these modules? How do I know if I need one, and how do I know when I need more than one?

2) Where should these be placed in the stack for best effect?

2) All three of the modules (Inline Reaction Wheel, Inline Advanced Stabilizer, and Advanced S.A.S Module) give the same amount of torque. So what is the point of using one over another?

Thanks! I hope to learn something in this thread that will make my Kerbal-naughts survive longer!

"SAS" is nowdays a function built into all capsules and probe cores. IRW and IAS parts are just to provide more torque, and IAS is obsolete (basically a heavy IRW), dating from the times when the capsules and cores didn't have SAS function built in, so you'd add it so that your ship could be steered. Squad should reuse the model for something else, it just clutters the part list.

1. Use IRW to gain more torque. Ignore IAS.

2. Center of mass. You want their grip to be in the center, so that your ship is turning evenly. I have no idea where do some people get the ideas that reaction wheels should be away from the CoM.

Where do you want to hold the hammer if you want to rotate it evenly with the least amount of effort? In the CoM.

3. Advanced S.A.S. has an obsolete name. It's basically a large diameter reaction wheel. Use it for sleek looks of your rockets.

Edited February 13, 2014 by lajoswinkler

[Tw1]

Wrong. Well, at least in KSP. IRL, it'd be right, but for KSP, doesn't matter for reaction wheels. True for RCS though :-)

Technically, they should be inline with the center of mass, for even rotation.

Reaction wheels aren't magic, they just spin, causing what they're attached to spin (or stop spinning) because of Newton's third law.

[Rodyle]

"SAS" is nowdays a function built into all capsules and probe cores. IRW and IAS parts are just to provide more torque, and IAS is obsolete (basically a heavy IRW), dating from the times when the capsules and cores didn't have SAS function built in, so you'd add it so that your ship could be steered. Squad should reuse the model for something else, it just clutters the part list.

1. Use IRW to gain more torque. Ignore IAS.

2. Center of mass. You want their grip to be in the center, so that your ship is turning evenly. I have no idea where do some people get the ideas that reaction wheels should be away from the CoM.

http://www.racetomars.ca/mars/ed-module/artificial_gravity/images/hammer2.jpg

Where do you want to hold the hammer if you want to rotate it evenly with the least amount of effort? In the CoM.

3. Advanced S.A.S. has an obsolete name. It's basically a large diameter reaction wheel. Use it for sleek looks of your rockets.

The idea with putting them as far away from the centre as possible comes from the idea that you have more leverage that way, I think.

[Kasuha]

Right. I have no seen people claim on these forums that:

- Reaction wheels are best as far away from the center of mass as possible

- Reaction wheels are best exactly on the center of mass

- Reaction wheels don't care where they are in relation to center of mass

I'll add myself to the side of those who say it's irrelevant where you put them.

Difference between reaction wheels and RCS is, reaction wheels provide rotation. RCS provides translation. Rotation via RCS is achieved by providing equal translation of opposite ends in opposite directions. Leverage plays role in it. For torque, leverage does not play any role and so doesn't distance from CoM.

When the reaction wheel starts acting, the part itself just tries to rotate around its own CoM. It skews its connections to other parts and applies force on them in opposite directions. The resulting speed of rotation of the whole ship does not depend on distance from CoM because it always needs to move the same mass.

[Vim Razz]

2) All three of the modules (Inline Reaction Wheel, Inline Advanced Stabilizer, and Advanced S.A.S Module) give the same amount of torque. So what is the point of using one over another?

For the two smaller wheels? Fashion! Use whichever looks best on your rocket.

For the larger wheel? It provides a better structural connection when sandwiched between two large segments than trying to put a smaller part in the same location.

[SSSPutnik]

The only time it matters (in the KSP universe) is if you attach them on something flimsy I think that breaks off when you try and use them.

In the real world, they should be at CoM IMHO.

Watch the weight of them, despite same strength some weigh more than others.

[Clockwork_werewolf]

So I was curious about where you should place reaction wheels as I always placed them at the center. So I sent this into space. Then I sent another almost exactly the same but with the wheels at the ends of two of the tanks instead of at the center.

The results:

Very near the center of mass 30-40 seconds for a 180 spin and 180 somersault.

Far from the center of mass 30-40 seconds for a 180 spin and 180 somersault.

Notes: The 10 seconds error was in my timing of the "thrust". I switched off ASAS and held down Q or W for 2 seconds then timed until it had turned 180. The Timing varied a bit but on both far and center spins without much noticeable difference. If someone want to go and test this over and over for more data be my guest but I got board.

As far as I am concerned, reaction wheel placement makes little to no difference.

[Streetwind]

That may have been because you chose a symmetric second layout. And in that layout, despite the fact that the reaction wheels are far away from the center of mass, the center of torque is still exactly congruent with the center of mass. Both setups are basically identical, and as such, it's no surprise that the results are identical.

Try again with:

- both reaction wheels where the above screenshot shows one, and none elsewhere

- both reaction wheels at the far end of a single tank, and none elsewhere

This will contrast a case where the center of torque is close to the center of mass with a case where it is well away from the center of mass. This should result in a very noticable difference - unless placement truly does not matter.

[numerobis]

You'd probably want to be flying with a reference next to you (or surrounding you) -- otherwise, spinning about the SAS unit and spinning about the CoM will look pretty similar.

[Stargate525]

An easier way to run that testing rig is to put two in the center, and one at each pole. Flip them on and off in patterns until results occur.

[Claw]

Torque and angular momentum can be tricky and defy intuition.

Reaction wheels make the object rotate about the center of mass. Because you are in space, the Center of Mass acts as the fixed point of your rigid body since nothing else is holding it still. (i.e. it's not attached to a wall)

Torque is not a force, so it does not react the same way as a force to distance from the CoM (or point of rotation).

Take this beam as an example. Imagine the CoM marker is a fixed point on the beam. In other words, it's held solid and not allowed to rotate around the center. The sum of moments is zero. So the moment acting around the center must be the same magnitude as the torque acting at the end of the beam.

Now, if the beam is allowed to rotate, it will rotate with a moment imparted by the torque at the end of the beam. This can be difficult to conceptualize because the torque at the end of the beam is allowed to rotate freely with the beam. Typically you're going to imagine the beam rotating around the torque, but the torque isn't a fixed object and it has no mass.

You can try to replicate this with a pencil. Pinch the center between two fingers in one hand. Use two fingers from the other hand to twist one end of the pencil. HOW you twist it is important. You can't push up or down, but twist around the tip. The pencil will spin easily as long as you allow your twisting fingers to rotate with it. This demo can be susceptible to how you twist, and if you're imparting forces and not torque. (If you have mechanical pencils with clips, you can interlock the clips to help provide rotation without imparting force. Try doing this with the clip at one end, then near the center.)

Perhaps a better example is the hammer example put forth above. Find the center of mass of your hammer (which won't be at the center of the length of the handle). Lay it on it's side on the floor.

-Use your fingers to rotate the hammer at the CoM.

-Use your fingers again to twist the hammer handle at the base and allow your hand to rotate with the hammer. You'll find it's easiest to rotate it this way when the hammer rotates around the CoM. If you're not rotating around the CoM, then you're also imparting a force which is going to make it harder to turn. Torque doesn't impart a force like this at the center of rotation.

If you manage to do this right, the amount of twisting you have to do at the CoM is the same as the twisting you have to do at the base of the handle.

APPLICATION

Now we all know in KSP the spacecraft are rarely rigid. So when you apply all that torque to one end of the space station to flip it around, it will rotate around the CoM. However, the poor module at the end where the SAS modules are located will take all of the strain of bending that torque around the CoM. If you distribute the SAS units, or put them near the center, there isn't nearly as much shear force concentrating at any point on the craft.

You can try out the fuel tank experiment above by assembling a fuel "tube" in the SPH and launching it. Then hack gravity and give it a try. The runway provides a visual backdrop to watch the center of rotation.

This thing is completely symmetric in weight from one end to the other. It flips around the CoM at the same speed whether I use the three SAS units at either end (one end at a time), or the three in the middle. It takes about 10 seconds to do a 360. The big difference is that using the ones on the end causes noticeable bending.

[problemecium]

Okay so I came here because today I was puzzling over the exact same thing. I used to think that lever mechanics applied, then I realized that was dumb and that any device that produces torque (rather than linear force) should be in the middle. But I wasn't completely sure so...

what's the consensus now? Does KSP add torque to the whole ship, or to the SAS module?

[Kasuha]

Does KSP add torque to the whole ship, or to the SAS module?

KSP definitely adds torque to the part, not to the whole ship. If you want to see a proof, make a stack of 10 small SAS modules, put it at the end of orange tank, put it on Launchpad and watch the stack bend as you apply torque.

Regarding where you should put it, physics says it is irrelevant.

With translation, lever reduces required force but increases the trajectory on which you need to apply it. Applying less force on longer trajectory means spending the same energy.

With torque it is the same. To achieve certain angular momentum, you need to spend appropriate energy and it is irrelevant whether you apply it at center or at ends.

[Claw]

I used to think that lever mechanics applied, then I realized that was dumb and that any device that produces torque (rather than linear force) should be in the middle. But I wasn't completely sure so...

That isn't dumb, it feels intuitive. But I generally find that torque and moments are not necessarily intuitive. It sort of defies the standard mass, fulcrum, lever concept that most people are used to.

what's the consensus now? Does KSP add torque to the whole ship, or to the SAS module?

Your question can still be misinterpreted. As Kasuha said, the torque is definitely applied at the point where the SAS modules are. However, the torque causes the whole ship to move around the CoM. All of that stress of moving around the CoM passes through whatever part the SAS units are attached to on the way to the CoM. Weak parts = bad results. If you spread out the torque or put it near the CoM, you will have less bending to deal with.

Just be thankful that you don't have to contend with all the actual physics. Otherwise we would have a limit to how much angular momentum you can generate before you have to bleed it off. (Please, no...)

[Vanamonde]

This is one of those details there was never any need for me to clarify, but this discussion made me curious, so I built a test rig.

That's 4 SAS modules and a probe core to exert torque, 2 of the long strut thingies to make something flimsy enough to show deviation, and the mass of a Mainsail and half-full orange tank to serve as something massive to resist the torque. Sure enough, when I apply WASD, the lump of reaction wheels jerks suddenly, and a fraction of a second later the rest of the structure twists to the side and follows along reluctantly, with the spars at a slight angle from the centerline as long as I'm holding the control input.

(The deviation is quite small, and wouldn't show in pics well. You'll have to take my word for it or duplicate my test.)

Edited February 14, 2014 by Vanamonde

[Clockwork_werewolf]

You'd probably want to be flying with a reference next to you (or surrounding you) -- otherwise, spinning about the SAS unit and spinning about the COM will look pretty similar.

One is rotating the navball around a single "true north" point, the other is rotating the whole navball on its head. Trust me no reference point apart for kerbal is needed.

And in that layout, despite the fact that the reaction wheels are far away from the center of mass, the center of torque is still exactly congruent with the center of mass. Both setups are basically identical, and as such, it's no surprise that the results are identical.

An easier way to run that testing rig is to put two in the center, and one at each pole. Flip them on and off in patterns until results occur.

In hindsight yes, but at the time I just wanted to get the test done. Considering the responses to this I'm not doing it again.

Thanks Claw, I did this stuff in school but that's now a decade in the past. Also I wasn't sure whether the reaction wheels would in game logic create a force sideways (which could "move" the COM) or whether they created torque. As its the latter your post is very helpful.