Does placement of Reaction Wheels matter?

[Tyko]

Do Reaction Wheels work better if they're placed further from the CoM in the same way RCS does?

[Greenfire32]

There was some debate about this awhile back and unless things have changed since then (and they could have!) I believe it was found that there's no difference in performance regarding placement.

[bewing]

But from an IRL physics perspective, they should work best at the CoM.

[Tyko]

1 hour ago, Greenfire32 said:

There was some debate about this awhile back and unless things have changed since then (and they could have!) I believe it was found that there's no difference in performance regarding placement.

Thanks!

27 minutes ago, bewing said:

But from an IRL physics perspective, they should work best at the CoM.

haha...I'll remember that when NASA asks me for my .craft files

[Leftotian]

47 minutes ago, bewing said:

But from an IRL physics perspective, they should work best at the CoM.

Should they? There should be no difference, where they are, torque doesn't care.

[bewing]

Imagine you are holding a steel I-beam out at the tip. How hard is it to turn the thing? Now you are holding it at the balance point. Is it easier to turn? Yes. Yes it is. The moment of inertia drops by half, I think (without doing the calculation).

Edited July 22, 2016 by bewing

[Leftotian]

No, not easier. No difference.

[Anth]

In reality doesnt there have to be 3 reaction wheels...spin, pitch and yaw?

 

[Rocket In My Pocket]

18 minutes ago, Leftotian said:

No, not easier. No difference.

Are you going to back that up with any sort of evidence or argument?

[Leftotian]

4 minutes ago, Rocket In My Pocket said:

Are you going to back that up with any sort of evidence or argument?

Do i have to? He didn't. From my two years of theoretical mechanics in university i remember, that there should be no difference. To be fair, it was more than a decade ago and i don't remember much, but just imagining forces applied to the stuff it doesn't look like there will be any difference. 

[bewing]

Jeez. Moment of Inertia of a Rod: http://hyperphysics.phy-astr.gsu.edu/hbase/mi2.html

I was off by a factor of two -- the moment of inertia is a QUARTER when you are applying the torque at the center of mass. So your theoretical mechanics didn't sink in too far.

[The Optimist]

12 minutes ago, bewing said:

Jeez. Moment of Inertia of a Rod: http://hyperphysics.phy-astr.gsu.edu/hbase/mi2.html

I was off by a factor of two -- the moment of inertia is a QUARTER when you are applying the torque at the center of mass. So your theoretical mechanics didn't sink in too far.

Is that a savage roast I smell?

[Kerbart]

21 minutes ago, bewing said:

Jeez. Moment of Inertia of a Rod: http://hyperphysics.phy-astr.gsu.edu/hbase/mi2.html

I was off by a factor of two -- the moment of inertia is a QUARTER when you are applying the torque at the center of mass. So your theoretical mechanics didn't sink in too far.

You're talking about the moment of inertia, which is the equivalent of mass for rotation. The formula for angular acceleration is τ=I×α (similar to F=m×a). If the position of where torque is applied matters, then why does the formula for angular acceleration not take that into account? Torque is force×arm length; for the force to be applied the position matters. But if you apply torque (without any net forces), the place where it gets applied really doesn't matter.

I had my teachers at college, when I was studying mechanical engineering, explaining this to me. I'd like to see what you can back up your claims with.

Edited July 22, 2016 by Kerbart

[linuxgurugamer]

The Reaction Wheels On The Game are so overpowered, it doest matter. But in real life, it does

[Kerbart]

4 minutes ago, linuxgurugamer said:

The Reaction Wheels On The Game are so overpowered, it doest matter. But in real life, it does

The reason for that has more to do with the distribution of torque through the entire construction than anything else. That's true in KSP as much as in real life. If you put your torque wheels at the end of a long construction, part of that torque will be absorbed by bending that construction, inducing nasty vibrations (that don't dampen out so quickly in space as they do on earth).

[Leftotian]

33 minutes ago, The Optimist said:

Is that a savage roast I smell?

Yeah, pretty sure he roasted himself by mixing up momen of inertia and moment of force.

[Boris-Barboris]

13 minutes ago, Kerbart said:

part of that torque will be absorbed by bending that construction

None will be. Wobbling wastes kinetic energy, but not total angular momentum of closed system.

Edited July 23, 2016 by Boris-Barboris

[RocketSquid]

19 minutes ago, linuxgurugamer said:

The Reaction Wheels On The Game are so overpowered, it doest matter. But in real life, it does

Actually, they aren't, at least in terms of torque or mass. They do, however, have a very high energy efficiency.

[Snark]

7 hours ago, tjt said:

Do Reaction Wheels work better if they're placed further from the CoM in the same way RCS does?

No, it doesn't matter.  Both IRL and in the game, their placement is irrelevant.  They will have exactly the same effect whether they're at the center or out at one end.

7 hours ago, Greenfire32 said:

it was found that there's no difference in performance regarding placement.

That is correct.

6 hours ago, bewing said:

But from an IRL physics perspective, they should work best at the CoM.

No, that is incorrect.  From an IRL physics perspective, it doesn't matter where they're placed.

5 hours ago, Leftotian said:

Should they? There should be no difference, where they are, torque doesn't care.

You are correct, sir, there is no difference.  All that matters is, 1. how much torque, and 2. the moment of inertia of the rotating thing.

Both of those numbers will stay exactly the same regardless of where the reaction wheel is placed.  Therefore no difference in effect.

5 hours ago, bewing said:

Imagine you are holding a steel I-beam out at the tip. How hard is it to turn the thing? Now you are holding it at the balance point. Is it easier to turn? Yes. Yes it is. The moment of inertia drops by half, I think (without doing the calculation).

That's correct, but irrelevant.  Because you have changed the center of rotation and therefore the moment of inertia of the I-beam:  it has a different moment of inertia rotating about the center than it does at one end.

That is emphatically not the case for a spaceship floating in zero gravity.  Reaction torque will always rotate it around its CoM, regardless of where the reaction wheel is placed.  It has to, otherwise you'd violate conservation of linear momentum.  Even if the reaction wheel is stuck way out at one end of the ship, the ship will still rotate around its center.

The reason why your physical intuition is tripping you up is that you're picturing earthbound physics, where things are typically anchored to something-- e.g. you have a fixed axle.

5 hours ago, Leftotian said:

From my two years of theoretical mechanics in university i remember, that there should be no difference. To be fair, it was more than a decade ago and i don't remember much, but just imagining forces applied to the stuff it doesn't look like there will be any difference. 

And I'll back that up with my degree in physics that says the same thing. 

4 hours ago, bewing said:

Jeez. Moment of Inertia of a Rod: http://hyperphysics.phy-astr.gsu.edu/hbase/mi2.html

I was off by a factor of two -- the moment of inertia is a QUARTER when you are applying the torque at the center of mass. So your theoretical mechanics didn't sink in too far.

The moment of inertia only changes if the center of rotation changes.  It doesn't matter where the torque is.

If you've got a fixed axle, and the axle is where the torque is applied, then you would be right.  But that's not the case here.

For example:  Let's say I have a long rod.  I grasp it in the middle, and then waggle my wrist back and forth to make it rotate.  This takes a certain amount of effort.  Now let's say I grasp the rod at one end, and waggle it that way, while holding my hand stationary.  It's a lot harder to waggle it.  That's because the moment of inertia has gone way up, because the center of rotation moved to one end.  True, the torque moved to the end, too, but that's not why the moment changed.

To demonstrate this:  Let's say I go ahead and move my grip to one end of the rod, same as in the second case... but now, instead of holding my hand rigidly stationary as I rotate the rod, I allow my hand to move in a semicircle, so that as I waggle the rod, my hand moves and the center of the rod remains stationary.  Guess what?  It just got a lot easier to waggle it again!  As easy as it was when I was gripping it in the center, in fact.  (Well, not quite as easy, because now I'm moving the mass of my arm, too, but you get the idea.)

Moral of the story:  Beware "common sense" and "physical intuition" when it's based on earthbound observations, if you're trying to make judgments about situations that you don't have personal experience with.  Our common sense can lead us astray.  Stick to the physics and you can't go wrong. 

[bewing]

You are all failing to grasp the concept of a lever arm, and the basic concept of a reaction wheel. An IRL reaction wheel applies torques around its own centerpoint. Not through the CoM. It is its own fulcrum. 

6 hours ago, Kerbart said:

You're talking about the moment of inertia, which is the equivalent of mass for rotation. The formula for angular acceleration is τ=I×α (similar to F=m×a). If the position of where torque is applied matters, then why does the formula for angular acceleration not take that into account? Torque is force×arm length; for the force to be applied the position matters. But if you apply torque (without any net forces), the place where it gets applied really doesn't matter.

I had my teachers at college, when I was studying mechanical engineering, explaining this to me. I'd like to see what you can back up your claims with.

It does take it into account, in the term for the moment of inertia. That is the point. In that link I gave -- look at the integral that calculates the moment of inertia. Look at the "R"s in it. That R is measured from the point where the torque is applied. A rocket that is being turned via a reaction wheel does not rotate through its CoM, unless the reaction wheel is at the CoM.

Edited July 23, 2016 by bewing

[Boris-Barboris]

1 hour ago, bewing said:

A rocket that is being turned via a reaction wheel does not rotate through its CoM, unless the reaction wheel is at the CoM.

It does. Don't mix torque application point and rotation axis, they don't need to be the same.

[OhioBob]

Regarding placement of torque, think of it this way...

Suppose we have a long truss structure in which we are using thrusters to apply the torque.  Straddling the center of mass we have two 100 N thrusters pointed in opposite directions and placed 1-meter from the CoM.  Let's say that both thrusters will produce a counterclockwise (positive) torque.  Let's computer the total torque,

T = 100 * 1 + 100 * 1 = 200 Nm

Let's say that we now move the thrusters out to the end of the truss. The thrusters are still 2 meters apart and pointed in opposite directions, but now both are on the same side of the CoM.  Let's say that the farthest is 6 meters from the CoM and the other is 4 meters away.  The farthest one will still produce a positive torque, but since the closer one has moved from one side of the CoM to the other, it will now produce a negative torque.  The vehicle will still rotate around the center of mass, so we compute the torque around that point,

T = 100 * 6 - 100 * 4 = 200 Nm

As you can see, we still get the same amount of torque at the CoM regardless of where we place the pair of thrusters (provided the thrusters maintain the same relationship to each other).  We can slide the pair of thrusters along the truss from one end to the other and they will always rotate the ship around the CoM with a torque of 200 Nm.

The same thing is true with reaction wheels - their placement makes no difference.

[Snark]

9 hours ago, bewing said:

You are all failing to grasp the concept of a lever arm, and the basic concept of a reaction wheel. An IRL reaction wheel applies torques around its own centerpoint. Not through the CoM. It is its own fulcrum.

No, we grasp it just fine, you're just mistaken about how this works.

You're absolutely correct that a reaction wheel produces torque around its own centerpoint.

But it will not magically rotate a structure around itself.  A free-floating mass that has no external forces will always rotate around its CoM.  It's physically impossible to make it move its CoM without either expending reaction mass (e.g. with a rocket) or by applying an external force (e.g. gravity).

If you have a long beam floating in space, and you put a reaction wheel at one end, the reaction wheel exerts torque around its center, yes, but the beam rotates around its CoM nonetheless.

To help illustrate this:  Let's imagine that you were right, and that it did rotate around where the torque is instead of around the CoM.  Let's say I have a 1-meter long beam, that has a small reaction wheel mounted at each end that can be turned on and off.  Initially the rod is floating in space such that wheel A is at the x=0m position, and wheel B is at the x=1m position.  The CoM of the whole shebang is at x=0.5m position.  Clear enough?

Now let's say we run reaction wheel B.  In this hypothetical universe where physics works the way you're claiming, this would rotate the beam around wheel B, which is at x=1m.  Let's say we run it until the beam has rotated 180 degrees and then stop.  Now the beam extends from wheel B at x=1m, to wheel A at x=2m.

Now we run wheel A and pivot around that, flipping the rod again.  Then we switch back to B.  Then A.  By doing that, we can move the rod as far as we like.

In other words:  We had an object in free-floating space, at x=0, with velocity 0... and simply by spinning some wheels and expending no reaction mass, we moved it to a different location.

That's physically impossible.  It violates conservation of linear momentum.  The center of mass of a closed system (i.e. not experiencing any external forces) must move at constant velocity.  Therefore, a closed system cannot alter the position of its center of mass.

Some additional references, if you need 'em:

https://www.quora.com/Why-does-a-rigid-body-rotate-about-its-center-of-mass-when-external-forces-stop

http://physics.stackexchange.com/questions/151374/why-does-an-unhinged-body-rotate-about-its-centre-of-mass

 

9 hours ago, bewing said:

That is the point. In that link I gave -- look at the integral that calculates the moment of inertia. Look at the "R"s in it. That R is measured from the point where the torque is applied.

No, you're mistaken.  The R in the calculation of moment of inertia is always measured from the rotation point.

Again, the thing that keeps tripping you up is that you're trying to use physical intuition instead of physics.  Let's look at that link you gave, shall we?

"A mass m is placed on a rod of length r and negligible mass, and constrained to rotate about a fixed axis."

Notice anything about that sentence?  Notice any similarity to something I said a while ago? 

11 hours ago, Snark said:

The reason why your physical intuition is tripping you up is that you're picturing earthbound physics, where things are typically anchored to something-- e.g. you have a fixed axle.

It's exactly this.

[Red Iron Crown]

8 hours ago, bewing said:

You are all failing to grasp the concept of a lever arm, and the basic concept of a reaction wheel.

You are failing to grasp that lever arms don't apply to applied torque, they only apply to applied force when speaking of rotation. A given torque applied to an object will induce the same rotation no matter where the torque is applied. 

You should quit while you're not too far behind here, honestly. The physics-major types correcting you in this thread are right.