Spin Stabilization Question

[Claw]

If I had to guess, I would say numerobis' comment about having a reference point is because the game forcebly sets the camera to the CoM. So if your frame of reference is the CoM, rotation around any given point will always appear as rotation about the CoM because of camera movement. The nav ball is incapable of anything besides rotation. It isn't going to show translational forces imposed on the CoM from rotation about an arbitrary point.

And it's pretty easy to repeat the experiment by hacking gravity. Then the ground provides a fixed frame of reference.

As for game logic, I have no idea how they are actually implimenting the math. Gyros and flywheels have all sorts of other wacky reactions to being turned and moved. I sort of like the simplified model of "apply torque here, watch things move."

Edited February 14, 2014 by Claw

[Kasuha]

The ship does rotate around its CoM even if the torque source is at its end. Even KSP physics is not that bad to do something different.

Camera wobbling when the ship bends and rotates is caused by something else: the camera's original reference point is not in the ship's CoM, it is in the ship's reference point (root part I guess) and if that part moves relative to ship's CoM during wobble, camera wobbles along with it.

[radonek]

…Camera wobbling when the ship bends and rotates is caused by something else: the camera's original reference point is not in the ship's CoM, it is in the ship's reference point (root part I guess) and if that part moves relative to ship's CoM during wobble, camera wobbles along with it.

I'm pretty sure camera is anchored to CoM, since it moves when pumping fuel around.

[Claw]

I'm pretty sure camera is anchored to CoM, since it moves when pumping fuel around.

I suppose since I may have started this tangent discussion, it is rather irrelevant where the camera anchors. If you don't have an external reference frame, the ship will always appear to rotate around the center of the camera view as long as the camera angle is fixed. It won't matter if it's centered on the CoM or another random part.

Also realize that if the parts of a ship are wobbling asymmetrically, the CoM would be moving too.

[Kasuha]

I'm pretty sure camera is anchored to CoM, since it moves when pumping fuel around.

I'm not sure to what exactly is the camera anchored but I'm pretty sure it is not CoM. The view is shifted towards CoM from that reference point but if that wobbles or twists, camera may get very far away from where CoM is.

Anyway - for all who don't believe I prepared a simple testing tool. Symmetric ship with reaction wheels all around, attached to a small reference ship. Get them to orbit, set up whichever reaction wheel you want to use, decouple, and watch what happens.

Sure enough, it rotates around CoM whichever reaction wheel you use.

Javascript is disabled. View full album

[Jank]

One consideration is that you use them because large massive ships are unwieldy to turn. So add control modules attached to those parts and staged with them. When you dump the mass, you won't need the extra control module weighing you down.

I have found however that for me, the big ASAS modules often break a stack due to lesser hull strength or something like that. When I put them in the middle of a big orange stack, or at the ends, they sometimes break the whole rocket in half, like a weak link that the mainsail just breaks. Sometimes directly upon firing up the engines, sometimes during ascent. It's gotten to the point that I kludge up girder-strut assemblies to keep them together.

[Kasuha]

I'm pretty sure camera is anchored to CoM, since it moves when pumping fuel around.

After just one experiment, I am certain camera is anchored to the root part and fixation to CoM is implemented as iteratively updated shift relative to that part.

Note: Recorded using hack gravity

Both "ships" are the same (okay the stack of reaction wheels is upside down but that doesn't affect the result).

Upper album - command pod is the root part of the ship

Lower album - upper orange tank is the root part of the ship

Watch what the camera does in either case

Javascript is disabled. View full album

Javascript is disabled. View full album

[Claw]

After just one experiment, I am certain camera is anchored to the root part and fixation to CoM is implemented as iteratively updated shift relative to that part.

Note: Recorded using hack gravity

Both "ships" are the same (okay the stack of reaction wheels is upside down but that doesn't affect the result).

Upper album - command pod is the root part of the ship

Lower album - upper orange tank is the root part of the ship

Watch what the camera does in either case

That is hysterical.

[Claw]

I have found however that for me, the big ASAS modules often break a stack due to lesser hull strength or something like that. When I put them in the middle of a big orange stack, or at the ends, they sometimes break the whole rocket in half, like a weak link that the mainsail just breaks. Sometimes directly upon firing up the engines, sometimes during ascent. It's gotten to the point that I kludge up girder-strut assemblies to keep them together.

Yes, large ASAS modules have lower crush strength. If you put them in the middle of a heavy rocket being pushed by a lot of thrust they break. That's why it's nice to be able to put them anywhere.

Of course, they could always increase the strength...

[Clockwork_werewolf]

Yes, large ASAS modules have lower crush strength. If you put them in the middle of a heavy rocket being pushed by a lot of thrust they break. That's why it's nice to be able to put them anywhere.

Of course, they could always increase the strength...

This is exactly why this data is important.

And it's pretty easy to repeat the experiment by hacking gravity. Then the ground provides a fixed frame of reference.

I actually tried this first, but the air resistance on the object turning greatly affected the end turning speed of the object. After most of the turn it was moving at a snails pace. Also I was not interested in whether it translated the object only the the wheels placement on turning speed.

[Claw]

I actually tried this first, but the air resistance on the object turning greatly affected the end turning speed of the object. After most of the turn it was moving at a snails pace. Also I was not interested in whether it translated the object only the the wheels placement on turning speed.

Ahh. Yeah, if you're going to time the turn then space would be better. The need for a fixed reference frame was to address the "rotating around the CoM or around the SAS module" question.

[lajoswinkler]

That is hysterical.

That's what she said.

[DrD]

I think that in a freefall, frictionless environment it would rotate around the center of the gyro. A reaction wheel as has been stated simply spins and thus the craft moves in the other direction (when the wheel accelerates I think.) Another type used on larger craft called the control moment gyro is a bit different and may help to illustrate my case. In this type the gyro spins and stays in place, while a motor actually rotates craft around the fixed gyro. Kind of like having an anchor point.

In both cases I think the craft would rotate around the wheel with the axis of the wheel the center of the system. In a freefall frictionless environment there is nothing "holding" the craft which can move any way it wants.

Imagine again a pencil, you attach some kind of spinner to the middle, turn it on. Pencil spins around the middle. Put the same spinner on the end and the pencil will spin around like a helicopter rotor with only 1 blade.

Again I just don't see what force would compel the stack to be anchored at its center of mass rather than having said center of mass rotate around the axis of the gyro/wheel.

EDIT: OK I thought more and did a 180. I made the mistake of assuming the pivot (gyro) end is fixed but if course it's not any more than CM is fixed. So while the initial movement would be a pivot as I described the far end's momentum would cause it to "pull" the pivot point inwards, as there would be no centripetal force to make it go in a circle. Hence both the gyro at one end and the other non-gyro end would rotate around a central pivot, likely the CM.

As an analogy, imagine a spinning 2 bladed helicopter rotor with big, heavy blades. One breaks off. Picture how that would look. The blade wouldn't smoothly move around the previously central pivot as it did before but the whole thing would kind of shimmy around a different pivot. Wheew.

Edited February 20, 2014 by sjwt

[Claw]

Imagine again a pencil, you attach some kind of spinner to the middle, turn it on. Pencil spins around the middle. Put the same spinner on the end and the pencil will spin around like a helicopter rotor with only 1 blade.

Again I just don't see what force would compel the stack to be anchored at its center of mass rather than having said center of mass rotate around the axis of the gyro/wheel.

EDIT: OK I thought more and did a 180. I made the mistake of assuming the pivot (gyro) end is fixed but if course it's not any more than CM is fixed. So while the initial movement would be a pivot as I described the far end's momentum would cause it to "pull" the pivot point inwards, as there would be no centripetal force to make it go in a circle. Hence both the gyro at one end and the other non-gyro end would rotate around a central pivot, likely the CM.

As an analogy, imagine a spinning 2 bladed helicopter rotor with big, heavy blades. One breaks off. Picture how that would look. The blade wouldn't smoothly move around the previously central pivot as it did before but the whole thing would kind of shimmy around a different pivot. Wheew.

Your pencil analogy would work if your spinner device was massless and allowed to rotate freely with the pencil. Unfortunately it is not as I think you subsequently pointed out. You also can't use a helicopter as an example either, because the rotation point is fixed and when you lose a blade, the CoM of the blades is no longer coincident with the point of rotation. Application of torque is different than forcing the center of rotation on an arbitrary point. If you look back a couple pages in this thread you might see my example of torque on a rigid beam. Moments only impart vectored force if there is a fixed point which is not coincident with the center of rotation, which as you pointed out, the space craft isn't fixed when it's in space.

[DrD]

I agree. Another example I thought of:

Imagine you're spinning a rope with a weight on it. Imagine you can spin in place (or maybe you're twirling your arm over your head.)

As the mass moves around circle, you can feel it pulling on you. Specifically you are applying centripetal force to keep object moving in a circle rather than flying off straight.

Now imagine you're standing on ice. I would guess that you would now move around such that you and the ball were spinning around the combined CM.

Man it's been along time since I've thought of this stuff.

[dreamsofspace]

Actually IRL it doesn't matter either, it's a free torque