Corona688

NOT A BUG: Reaction wheels don't care where they're mounted

102 posts in this topic

On 1.09.2016 at 11:34 PM, The_Rocketeer said:

The craft will always rotate around it's CoM wherever any wheel of any size on any rotational axis is physically placed on the vessel,

One caveat:

The wheels, to provide any reasonable torque within any reasonable RPM, need to be somewhat massive - of mass that is at least somewhat comparable to the mass of the craft.

And that mass does contribute towards the craft's moment of inertia. Creating the counter-factual impression that point of application of torque matters.

Imagine two KSP constructions, where 'xxxxxxxxxxx' is a long, but very light bar, made

 

[reaction wheel][battery]-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx-[OKTO2]-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx-[battery][reaction wheel]

and

-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx-[reaction wheel][battery][OKTO2][battery][reaction wheel]-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx-

Which will be spun up easier? The latter obviously.

But is it because the torque was applied in different locations? No. It was because a mass was applied at point of application of the torque; changing the moment of inertia of the whole craft!

 

1 person likes this

Share this post


Link to post
Share on other sites
1 hour ago, Sharpy said:

The wheels, to provide any reasonable torque within any reasonable RPM, need to be somewhat massive - of mass that is at least somewhat comparable to the mass of the craft.

And that mass does contribute towards the craft's moment of inertia.

Well, you're certainly correct that the distribution of mass on the craft matters.  Having more mass concentrated nearer the CoM will result in a lower moment of inertia, making it easier to rotate the craft.

However, in practical terms, I've found that that doesn't matter much in KSP, mainly due to the reaction wheels being so overpowered.  The 1.25m reaction wheel has a mass of only 100 kg, but provides easily enough torque for a craft that's several tons.  If the mass of the reaction wheel is under 2% of the overall ship mass, you're not going to see any noticeable change in the craft's moment of inertia based on where the reaction wheel is.  Yes, there's a difference, but so tiny as to be unnoticeable.

1 person likes this

Share this post


Link to post
Share on other sites

@Sharpy

Hrrrmmmm...

If what you're saying is what I think you're saying, you're basically applying my point about the mass-distribution of the reaction-torque flywheel to the mass of the whole craft? In which case, yes I think that's true, but it isn't really a caveat to the quote, to which it isn't contextually relevant. It's speaking to how the same amount of kinetic energy is distributed around two differently-arranged systems rather than the transfer of energy from one system to another.

It is a neat point tho that may have been overlooked in discussion so far, so have a rep :)

Share this post


Link to post
Share on other sites
25 minutes ago, The_Rocketeer said:

In which case, yes I think that's true, but it isn't really a caveat to the quote, to which it isn't contextually relevant.

Eh.

At first, I wrote "with change of location of the flywheels, location of CoM changes."

Then I realized this only really matters, because same shape with different CoM will have a different moment of inertia, so I corrected to "with change of location of the flywheels, location of CoM and the moment of inertia changes."

Then I realized the wheels may be moved in such a way that CoM doesn't change at all, but moment of inertia does, so I removed the CoM part entirely. Which left my reply contextually irrelevant. Uh, sorry!

Share this post


Link to post
Share on other sites

I would have thought that one would need reaction wheels for each of the x, y, and z axis.

A reaction wheel is a type of gyro, right?

Share this post


Link to post
Share on other sites
23 minutes ago, Core said:

I would have thought that one would need reaction wheels for each of the x, y, and z axis.

A reaction wheel is a type of gyro, right?

Yes, but the reaction wheel parts in KSP represent an assembly with several individual reaction wheels inside.  So you don't need multiple reaction wheel parts on your ships, except to multiply the force. 

Share this post


Link to post
Share on other sites
38 minutes ago, Core said:

I would have thought that one would need reaction wheels for each of the x, y, and z axis.

A reaction wheel is a type of gyro, right?

You've pretty much got it - at least 1 wheel/disc per axis. (though I guess in theory you could use a single sphere for the same effect). You can also use sophisticated systems that balance a whole array of wheels (4+) to the same effect.

A reaction wheel is basically a flywheel that spins up fairly slowly using a motor, and then imparts torque to the parent structure through a mechanical brake. By putting the right amount of force into the wheel and taking it out in the right amounts, it's possible to make a vehicle that balances in an unnatural equilibrium, or that can turn to face any 3-dimensional direction.

1 hour ago, Sharpy said:

Uh, sorry!

Hey, not judging :wink:

But the critical point is the change in CoM as a result of moving the wheel assembly: if CoM and total mass stay the same (i.e. you can't just add a counterweight), nothing changes regardless of position. If either of those do change, then all the rest applies.

Oh god now I'm falling into the trap I'm trying to warn against! Let me adjust my thinking process and I'll try again.

Edit (much later :confused::D):

Ok, so the critical point is not the mass distribution of the craft around the reaction wheel, it's the mass distribution of the craft around it's own CoM. So, a craft shaped like a bike-wheel, with most of it's mass in the rim and only lightweight spokes, will turn as easily on that axial plane wherever the reaction wheels are placed.

However, compared with a craft of the same maximum dimensions, same reaction wheels and mass but with most of the mass concentrated around the 'hub-bearing' of the bike wheel and only a lightweight rim, the first craft will be harder to turn.

This is because the required average acceleration-per-mass of the craft is less in the second craft and therefore consumes less energy to achieve the same rotational velocity. What I mean is, most of the mass is in the middle, so most of it is turning in small circles, which have short circumferences, compared with the other craft where most of the mass is turning in big circles with big circumferences.

This is the distinction you're describing - the only reason moving the reaction wheels would affect this is because it could alter the mass distribution and therefore the average acceleration-per-mass.

I think this better answers @Diche Bach's question about why reaction wheels are often in small assemblies - it basically helps centralise the craft's mass distribution and thereby reduce power demands.

Edited by The_Rocketeer

Share this post


Link to post
Share on other sites
15 minutes ago, pandaman said:

Yes, but the reaction wheel parts in KSP represent an assembly with several individual reaction wheels inside.  So you don't need multiple reaction wheel parts on your ships, except to multiply the force. 

I had a feeling it was that. I was just uncertain after seeing the model design for the largest reaction wheel on the game, seeing how it was  hollow.

1 person likes this

Share this post


Link to post
Share on other sites
Just now, Core said:

I had a feeling it was that. I was just uncertain after seeing the model design for the largest reaction wheel on the game, seeing how it was  hollow.

Totally legitimate point. I never actually thought about it!

Share this post


Link to post
Share on other sites
58 minutes ago, Core said:

A reaction wheel is a type of gyro, right?

Not precisely. A gyro uses gyroscopic force and precession to control attitude, while a reaction wheel uses "straight" angular momentum conservation instead. The notable difference is that a gyro requires high rotational speed to be effective while most reaction wheel controllers attempt to keep wheel speeds as low as possible.

A simple explanation from a NASA person can be read here. Reaction wheels, momentum wheels and control moment gyroscopes are three distinct things, though they use similar concepts.

2 people like this

Share this post


Link to post
Share on other sites

Who would have thought that a modest, cartoony computer game could spark such a genuine interest, and raise this informative in-depth discussion on what's actually quite a 'sciency' subject.

Many physics teachers would love a 'tool' this effective at grabbing students attention.  It just goes to show that if you can make learning fun, more interest and more learning happens.

 

2 people like this

Share this post


Link to post
Share on other sites
1 hour ago, Core said:

I had a feeling it was that. I was just uncertain after seeing the model design for the largest reaction wheel on the game, seeing how it was  hollow.

Not hollow, though. Full of magic :wink: 

Edited by monstah
what the hell, wrong quote?

Share this post


Link to post
Share on other sites

The only reason wheel placement does matter somewhat is because no practical craft is ever 100% rigid. So the reaction wheel will impart torque on the structure it's attached to, and this structure will flex. In the most extreme case, imagine a reaction wheel attached to a spacecraft on a piece of string, versus one attached on a concrete pole. :wink:

Even more true is this the case in KSP, where "craft flex" is more extreme than in real life, and it's the craft flex that will cause the 'control computer' (SAS) to chase after the craft's inertia. Because of the craft flexing, transmitting torque from one end of the craft to the other takes time, and if the sensor component (Jeb's inner ear or probe core) is attached to one end, while the actuator (an engine gimbal or big honkin' reaction wheel) is attached on the other with only a noodle of fuel tanks in between, you're gonna have a bad time.

A typical player's reaction is usually "My craft is wobbling and unstable, HELP, I need more control, so let's smack on more reaction wheels!", followed by cramming reaction wheels in all the wrong places, causing an even more extreme amount of torque being applied, followed by heavier oscillations, and eventually RUD.

The placement of reaction wheels does not matter for the amount of torque being applied on the craft, but it definitely matters in terms of control lag. Get those wheels as close to the 'Control from Here' bit as possible. :wink:

Share this post


Link to post
Share on other sites
5 minutes ago, Stoney3K said:

The only reason wheel placement does matter somewhat is because no practical craft is ever 100% rigid.

I don't know if you meant to, but your above statement dismisses the whole point that @Sharpy flagged up (top of this page). While I don't dispute your rigidity issue, it is not the only reason placement might matter.

Share this post


Link to post
Share on other sites
20 minutes ago, The_Rocketeer said:

I don't know if you meant to, but your above statement dismisses the whole point that @Sharpy flagged up (top of this page). While I don't dispute your rigidity issue, it is not the only reason placement might matter.

I had brought up that point too, but I actually like @Stoney3K's answer. Mine and @Sharpy's situation is kind of a silly one, @Stoney3K's happens very frequently in KSP :) 

1 person likes this

Share this post


Link to post
Share on other sites
31 minutes ago, Stoney3K said:

Get those wheels as close to the 'Control from Here' bit as possible.

Instructions unclear, disaster imminent:

screenshot3.png

It's better to distribute the wheels throughout the craft rather than concentrating them in one place, near the control point or not.

3 people like this

Share this post


Link to post
Share on other sites

Okay, so let me add to the confusion :)

Reaction Wheels, versus CMG. - Control Moment Gyroscopes,

Reaction wheels - simpler devices, fixed to the craft. One wheel per axis, each controls rotation in each axis, each needs to be desaturated separately. Gyroscopic effect is more or less deterrent to their action; they work directly on 3rd law of motion; wheel accelerates clockwise, ship accelerates counterclockwise.

CMG - a single wheel (rotor) mounted in a motorized gimbal can turn around full 360 degrees. The gyroscopic effect of the rotor is utilized as the rotor opposes being turned "sideways" providing a "grip" for the gimbal motor to push against;  pulling the whole craft with it. Much higher torque per volume, but also higher power usage, especially that desaturated reaction wheel is just stopped; a desaturated CMG has its rotor spinning at nominal, pretty high RPM.

Share this post


Link to post
Share on other sites
13 minutes ago, Sharpy said:

Much higher torque per volume, but also higher power usage

My understanding is that CMGs use less power rather than more.

Edit: Googled a bit, looks like RWs are more power efficient for low moment of inertia craft and more power efficient for large moment of inertia craft (source p.6). TIL.

Share this post


Link to post
Share on other sites
28 minutes ago, Red Iron Crown said:

My understanding is that CMGs use less power rather than more.

Edit: Googled a bit, looks like RWs are more power efficient for low moment of inertia craft and more power efficient for large moment of inertia craft (source p.6). TIL.

Idle desaturated RW is just stopped, zero power. You can't beat it. A stopped CMG is useless as there's no gyroscopic effect; the rotor must be kept spinning (against friction). But when actively exerting torque, CMG will take less power than RW for the same mass of rotor/flywheel and exerted torque. Plus CMG due to its motorized gimbal is large and heavy itself; not something you can stuff in a 1U cubesat.

Share this post


Link to post
Share on other sites

I just want to make sure that I understand this all properly. I think I do, but I'll keep at it if I'm wrong.

In-game reaction wheels have a fixed torque (τ = F\cdot \!\,d). Thus, regardless of the distance from the CoM (the placement of the reaction wheel), the torque on the vessel is constant, and therefore placement is irrelevant.

In contrast, a system that provides a force, rather than a torque (e.g. RCS) gives a variable torque, in proportion to the distance from the CoM. Thus, the torque provided by a force (RCS) is NOT constant, and thus placement DOES matter, given τ = F\cdot \!\,d.

For example, a vessel with the distance from the reaction wheels to the CoM of 2 m, and a reaction wheel torque of 10 Nm, will be equivalent to a system in which a thruster is mounted the same distance from the CoM with a force of 5 N.
However, if the distance from the reaction wheel (or thruster) is increased to 5 m, the situation changes. For the reaction wheel, torque is constant, and will be 10 Nm, regardless of this different placement. However, for the same thruster (F = 5 N), the torque will instead be 25 Nm. Therefore placement DOES matter for a force-based system, such as RCS, but DOES NOT matter for a torque-based system, such as reaction wheels.

Again, if there is something incorrect in here, then your assistance is much appreciated. :)

Share this post


Link to post
Share on other sites

@CommanderSmith That is all correct. Placement of RCS thrusters very much matters, they impart more torque the further they are from the CoM due to the longer lever arm.

1 person likes this

Share this post


Link to post
Share on other sites

Plus minus position of mass (dead weight) of the wheels influencing CoM and moment of inertia of the craft as a whole *trollface*

Share this post


Link to post
Share on other sites
21 hours ago, Red Iron Crown said:

Instructions unclear, disaster imminent:

screenshot3.png

It's better to distribute the wheels throughout the craft rather than concentrating them in one place, near the control point or not.

Maybe splitting a few hairs here but:

You're using a Vector, which has massive control authority even compared to the biggest reaction wheel. So what will happen is the Vector will gimbal one way and the reaction wheels will turn to compensate because the vessel starts moving around, which will cause oscillations and the familiar noodle-flop. Both the engine end and the reaction wheel end will try to turn the craft, which only results in bending it.

Try building the same craft with both a Reliant and a Swivel to see what the different results are.

(OK, we're drifting a bit off topic with this, may be a good idea to continue in a new thread for it.)

Share this post


Link to post
Share on other sites

I realize this is a bit of a necro, but i found this thread by googling reaction wheels vs monopropellant  (and i get that i may be mistaken in what monopro is for...). simply put, i have a ituitive grasp of physics and this topic has what looks like important info that escapes me (intuitive, but im clueless at the formulae). Can someone send me a "for dummies" explanation or link please?

1 person likes this

Share this post


Link to post
Share on other sites

Welcome to the forums!

Reaction wheels will only rotate (or stop rotating) the ship. They use electricity only.

Mono is fuel for little engines (RCS) that allow you to translate AND rotate. No electricity needed. RCS is used to do fine movements such as docking ships together.

  • You can turn RCS on and off by pressing R. The green light to the left of the navball will come on.
  • The normal WSAD keys will rotate the ship, just like with the reaction wheels.
  • IJKL will shift. H and N to go forwards and backwards.

http://wiki.kerbalspaceprogram.com/wiki/Reaction_Control_System

http://wiki.kerbalspaceprogram.com/wiki/Reaction_wheel

Does that answer your questions?

Feel free to ask more.

2 people like this

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now