Best Placement for Reaction Wheels

[-root]

I tried searching but couldn't find anything, where is the best spot to put reaction wheels?

I'm thinking near the COM would be the best place, but I'm not that good at physics to know for sure. I do know placing them off centre can create a "worm" effect in the rocket as one part moves and then the other whips around to follow it.

Also, scaling up, would multiple wheels near the COM be better or would you be better off putting one at each end of the rocket (or spreading them throughout)?

Reason I ask, I've got some absolute beasts heading into orbit now and they are a copper plated to orientate effectively. Couple of times I've missed apoapsis because I simply couldn't get the rocket around in time

Edited July 5, 2014 by -root
Answered

[Claw]

I tried searching but couldn't find anything, where is the best spot to put reaction wheels?

Let's see if I can get this post done before getting ninja'd.

I'm surprised you didn't find anything, since this particular topic usually spurs a LOT of conversation...

I'm thinking near the COM would be the best place, but I'm not that good at physics to know for sure. I do know placing them off centre can create a "worm" effect in the rocket as one part moves and then the other whips around to follow it.

Well, I can say what you are thinking is correct. If you place the Torque modules in bad spots, it can cause the ship to flex and bend around. Doing the "worm" as you so put it.

One thing to understand first though, is that placement of the torque modules does NOT matter in terms of how well they turn the craft. Whether they are at the center or at the ends of the craft, they will apply an equal amount of "turning ability" on the ship. And the ship ALWAYS turns around the Center of Mass.

Also, scaling up, would multiple wheels near the COM be better or would you be better off putting one at each end of the rocket (or spreading them throughout)?

As far as scaling up, you are generally better off spreading the torque modules around a little. The reason is because all that bending and flexing gets spread out over the ship. If you put all the torque modules in one place, then all of their combined bending gets concentrated into one (or a few) joints on the ship. This is usually a bit painful for those few joints.

Reason I ask, I've got some absolute beasts heading into orbit now and they are a copper plated to orientate effectively. Couple of times I've missed apoapsis because I simply couldn't get the rocket around in time

Yep, you can add more torque modules to help turn around faster. Adding RCS out on the extents of your ship will also help turn faster while placing less bending stresses on your ship...

However, for Torque, it's more important that your ship is fairly stiff and the torque a bit spread out. Also, place your "control from here" part near the center of the ship, or on in a place where the ship won't be "bendy." What happens is that once the ship starts to wobble, the SAS reacts to the control part bouncing around, which inputs more torque, which causes the ship to bounce around more, which wiggles the control part, which causes the SAS to react... Hopefully you can see where that is heading.

Hopefully that gives you a good start.

-Claw

[Red Iron Crown]

If your ship is reasonably rigid, it doesn't really matter where the wheels go. Put them wherever it's convenient.

If your ship is fairly flexible (common for things assembled in orbit), you're better off evenly distributing the wheels throughout the ship.

[Kasuha]

There are two ways how to place reaction wheels reasonably.

1/ Either spread them evenly throughout the ship, making sure about equal mass is there per a reaction wheel.

2/ Or put them all around your control point (command pod or probe core).

The method 1/ works well for all reasonable ships and keeps working as you stage, as long as the ship is not too flexible. It reduces bending of the ship and in general provides stable results.

Method 2/ works when your ship is flexible or consists of two or more considerable masses separated by flexible joints. It does not try to keep the ship stable - it tries to keep your control point stable to provide you reasonable reaction to your inputs. You will need to wait for the ship to stop flexing after each maneuver but it will allow you to rotate it to desired position.

[SRV Ron]

I just place an extra SAS on a mid stage if needed for extra control. Usually, one extra is more then enough. That way, it can be staged when it is no longer needed.

If you need more then one extra, check over your rocket design. Most SRBs are not vector able and should only be used in the first stage. Any control problems past that is usually a design problem, something has broken during flight, or the result of flexing.

[LethalDose]

I think it's been stated repeated that the location of the torque wheels in the craft doesn't matter. But this is due the code, not the magical physics behind how gyroscopes work.

Somewhat related question: does anyone have any experience with changing the "control from" point to somewhere near the center of the ship to prevent wobble? I'm thinking that if the orientation of the "control from" location isn't changing much, it should mitigate the wobble. Haven't had a good chance to test that yet.

[Claw]

I think it's been stated repeated that the location of the torque wheels in the craft doesn't matter. But this is due the code, not the magical physics behind how gyroscopes work.

No, seriously. Torque works like this in the real world too. Torque is not an intuitive thing like regular levers and point force because it is rarely encountered in everyday life. This is not some happenstance property of KSP.

Placement doesn't matter for adding the potential to turn. But since spacecraft generally aren't rigid bodies, placement does matter for being able to turn without ripping things apart.

I'm thinking that if the orientation of the "control from" location isn't changing much, it should mitigate the wobble.

It's not necssarily that the Control part is at the center. Just that it's at a spot that doesn't wobble much. Especially if that wobbly spot is somewhere away from the torque, which can get out of sync with the torque and cause the SAS to resonate. Orientation can play into it depending on the resonance, but a less wobbly location has a much bigger impact.

[LethalDose]

It's not necssarily that the Control part is at the center. Just that it's at a spot that doesn't wobble much. Especially if that wobbly spot is somewhere away from the torque, which can get out of sync with the torque and cause the SAS to resonate. Orientation can play into it depending on the resonance, but a less wobbly location has a much bigger impact

This is the reason that I was saying put it near the center of mass; locations near the CoM should have a shorter lever arm and therefore less experience less wobble. When I wrote "the orientation [...] isn't changing much", I meant changing due to the wobble. I thought that was clear, but apparently it wasn't.

Certainly, there are other locations may not experience wobble, but they may not be as reliably wobble free as the CoM, and after the craft is launched there's usually not much you can do about, unless the design was modular.

Regardless of all that, the question wasn't answered: does it work in practice?

[Claw]

Regardless of all that, the question wasn't answered: does it work in practice?

Yes. Generally speaking (for most typical designs) the CoM tends to wobble less when the rest may wobble. Because of that, using a control part near the CoM will reduce the chance of resonance.

For the same reason, putting torque modules near the CoM can have the same effect, which I think is the root of the "torque at the CoM is best" claim. But torque concentrated near the CoM can still cause the ship's "arms" to bend excessively.

It's all very conditional on the design, which is why it's hard to give an absolute rule.

[GoSlash27]

IMO the most important thing is to have adequate torque while avoiding excessive oscillation. I'm a big fan of having a reaction wheel right at the COM, but it's just as important to have your controlling pod rigidly linked to it. If you can't put your controller right next to your reaction wheel, then you either must make sure that your rocket is rigid or else move your reaction wheel closer to your controller.

HTHs,

-Slashy

[-root]

Let's see if I can get this post done before getting ninja'd.

Congrats

Well, I can say what you are thinking is correct

Excellent, but now I know why I was confused:

As far as scaling up, you are generally better off spreading the torque modules around a little. The reason is because all that bending and flexing gets spread out over the ship.

I had tried loading up the middle of the ship and spreading them out. Theoretically placing them at the centre was the right thing, but I got better results spreading them out.

Thanks for the answers everyone

[cantab]

Regarding control point location, I like to have it close to the engines. After all, when I'm making a burn it's the direction they're thrusting in that matters. It's also likely the engines are attached to the fuel tank stack which is the most massive part of the ship when full.