Is Torque Additive?

[957Chatterton]

I know this has been answered but the threads I've read contradict each other. Is torque additive? If so, how much per ton?

Edited April 15, 2014 by 957Chatterton

[TechnicalK3rbal]

What exactly do you mean?

[chemieglennie]

The more reaction wheels you add the quicker you can turn. I've started adding some on the radial attachement points on the sides of big tanks. I usually only add 2 or 4 even on pretty big ships.

[karolus10]

Yes, each gyro apply certain force and they all sum up.

Reaction wheels are pretty powerful for now, and even few are enough to generate a lot of torque.

Proof:

[Kasuha]

Reposting my earlier experiment with pendulum on spring:

Javascript is disabled. View full album

[cantab]

Yup, two reaction wheels gives twice the torque as one.

For how quickly that torque turns you, it's more complicated, look up moment of inertia. The more mass, and also the further it is from the centre of mass, the harder.

[Taki117]

...The more mass, and also the further it is from the centre of mass, the harder.

This is a common misconception with KSP. The position of the reaction wheels has no effect on their turning efficiency. There was some discussion some time ago that talked at length about this, but it was determined that Reaction Wheels placed At (Or close to) the CoM are just as effective as those placed further away. (Contrary to Logic, but when has KSP ever been logical?)

[Claw]

Yup, two reaction wheels gives twice the torque as one.

For how quickly that torque turns you, it's more complicated, look up moment of inertia. The more mass, and also the further it is from the centre of mass, the harder.

This is a common misconception with KSP. The position of the reaction wheels has no effect on their turning efficiency. There was some discussion some time ago that talked at length about this, but it was determined that Reaction Wheels placed At (Or close to) the CoM are just as effective as those placed further away. (Contrary to Logic, but when has KSP ever been logical?)

Taki, I think maybe you misunderstood what cantab said.

He is 100% correct in saying "for how quick torque turns you...moment of inertia. The more mass, and also the further it is from the center of mass, the harder." KSP does, in fact, simulate moments of inertia. Mass distributed away from the CoM is much harder to turn than mass at the CoM. Long, heavy ships are much slower to turn/stop turning.

You are also 100% correct in saying that the position of the reaction wheels has no effect on their turning efficiency. And it's not contrary to reality, it is 100% in agreement. I will say that it's not intuitive, but it isn't wrong.

[cantab]

Claw pretty much covered it. Reaction wheel placement doesn't matter for overall control effectiveness (though it may matter for deformation of the ship, I'm not sure). Overall mass distribution does matter, and I just checked that with my Reaction wheel test ship. Four fuel tanks in a line, two full and two empty, turns more slowly if the end ones are full than if the two middle ones are full. And of course a typical rocket will roll under reaction wheel torque far more readily than it will pitch or yaw.

[Claw]

...(though it may matter for deformation of the ship, I'm not sure). ...

Yes, it certainly does and has broken many space stations.

[StrandedonEarth]

Yes, it certainly does and has broken many space stations.

So any suggestions on the best RW distribution to avoid breaking long ships, and perhaps minimizing wobble? In the middle, at the ends, or spread evenly along? I would guess the last one

[Claw]

So any suggestions on the best RW distribution to avoid breaking long ships, and perhaps minimizing wobble? In the middle, at the ends, or spread evenly along? I would guess the last one

It will depend a lot on construction techniques and the design of the ship itself. Generally speaking, using a mix of torque and RCS is a good start if you're making something large. RCS seems to put a bit less bend into the joints.

Spreading out your torque a bit is better than putting it way out at the ends. Concentrated right at the center is often impractical and isn't ideal either, but is often better than at the ends. Usually distributed about, but mostly near the core (if that makes sense).

The problem is that if you put a torque part on a module that is relatively weakly connected, all that twisting translates into a lot of bending force on the joints. This can also be a problem if the part is relatively long before it connects to significant structure. Even if the joint doesn't break, the long portion of the structure flexes as the SAS controller tries to fine tune, which starts driving it crazy. That's when you start to get into resonating SAS modules that rip the ship apart. (Parts moving away, then toward each other repeatedly.) If you can move the torque module closer to where it connects, that often helps keep the wobble and breaking down.

The nature of the tree construction that KSP uses forces some of this. So be careful about connecting a really long part and putting a torque module at the end.

[m4v]

This is a common misconception with KSP. The position of the reaction wheels has no effect on their turning efficiency. There was some discussion some time ago that talked at length about this, but it was determined that Reaction Wheels placed At (Or close to) the CoM are just as effective as those placed further away. (Contrary to Logic, but when has KSP ever been logical?)

It kinda does matter, is just that for your average rocket most of the time the CoM is aligned with the reaction wheel so you never notice it. But in an asymmetrical rocket is pretty easy to see that if you try to yaw, the rocket will both yaw and roll, after all is the reaction wheel itself the one that's trying to rotate, dragging the ship with it.

[957Chatterton]

See? There's a lot of conflict about this. Could one of the devs, as busy as they are, confirm this?

[Kasuha]

So any suggestions on the best RW distribution to avoid breaking long ships, and perhaps minimizing wobble? In the middle, at the ends, or spread evenly along? I would guess the last one

My experience is that to break a ship, you really have to go to extremes. Or have a really unstable ship. But even without breaking up, just the wobbling may get rather annoying. My approach is to distribute torque evenly over the mass of the ship. Reaction wheel - X tons of mass - reaction wheel - X tons of mass - reaction wheel - etc. As long as the ship is reasonably homogenous and those "X tons of mass" are of similar mass and size the resulting ship usually behaves very well.

[m4v]

See? There's a lot of conflict about this. Could one of the devs, as busy as they are, confirm this?

No, what taki117 said is correct for like all practical purposes, I was just getting "technical", really you aren't going to build asymmetrical rockets. So in short, torque is additive; and most of the time you can place reaction wheels wherever the hell you feel like it, if you have to place many of them is best to distribute it around your ship and not in a single stack so the rotation is uniform for most parts (less wobble).

[Claw]

See? There's a lot of conflict about this. Could one of the devs, as busy as they are, confirm this?

Actually, this thread has the least amount of disagreement I have seen on the subject of torque. There is an old one that went on for 20+ pages if I recall.

- Torque is additive

- Torque module location does not matter for effectiveness of turning

- The amount per ton depends on the shape and size of your ship. Long ships will need a bit more than short ones. Heavy will need more than light ones. It also depends on what you are planning to do with the ship and how fast you want it to turn around.

Unfortunately, there isn't a magic "x torque per ton" like there is with thrust for launching a rocket.

If you are referring specifically about "where exactly do I put torque modules?" You will see a variety of answers. Again, there is no one-perfect solution that matches every design.

[cantab]

You might build an asymmetric rocket if you attached a reaction wheel radially and balanced it with other radially-attached parts. And you're probably more likely than not to build an asymmetric space station.

And I can confirm that in .23.5 at least ships are indeed hard to break.

That's 4 asteroid control sections each with 4 reaction wheels, connected to the rest of the ship by Regular Docking Port-Quad Adapter-Sr Docking Port both above and below. It bent back and forth on attempting to turn as you can see, but didn't break until I deliberately made inputs to make the bending worse.

[957Chatterton]

Okay, I think I finally understand the confusion. Torque modules stacks are too loosely connected.