Guidelines for placings reaction wheels in multisection stations

[LethalDose]

I'm really pretty confused about where to place SAS units in modular stations. On vessels, it's seems effective to have the reaction wheels near the CoM. However, activating reaction wheels near a stations' the CoM tends to cause potentially fatal oscillations (I shut the torque off before the station shreds itself) when I try to make even small corrections in station orientation.

Oddly, the reaction wheels near the ends of the station tend to be a crap shoot (some cause the oscilations, some don't)

So... What's up with this, and what guidelines do other players use to place reaction wheels on modular stations, especially in terms of expandability?

[Taki117]

The position of the reactions wheels has no effect on how effective they are.

[LethalDose]

The position of the reactions wheels has no effect on how effective they are.

Okay, fine, but that's far from the point of the original question, which is how can I place them to reduce oscillations in modular stations. This must have positional component because it is happening in my games.

[Taki117]

Okay, fine, but that's far from the point of the original question, which is how can I place them to reduce oscillations in modular stations. This must have positional component because it is happening in my games.

As far as placing them, I think it has to do with them being at odd angles with eachother. I have not noticed any scilations when they are placed in line or 90 deg from center. Without a pic of the station there is little I (Or anyone else for that matter) can do

[Geschosskopf]

As far as placing them, I think it has to do with them being at odd angles with eachother. I have not noticed any scilations when they are placed in line or 90 deg from center. Without a pic of the station there is little I (Or anyone else for that matter) can do

The only time I get oscillations when rotating large piles of docked modules (each of which has its own SAS unit) is when there's a significant off-center mass, like a lander docked on 1 side with no counterweight on the other side. This moves the axis of rotation out of line with the station's body and the thing starts wobbling during the rotation.

[LethalDose]

As far as placing them, I think it has to do with them being at odd angles with eachother. I have not noticed any scilations when they are placed in line or 90 deg from center. Without a pic of the station there is little I (Or anyone else for that matter) can do

I'll post images when I get home. It can't be placement of the reaction wheels in relation to other wheels because this is happening when only a single SAS module is active: Without multiple reaction active, the interaction of multiple SAS modules can't be responsible for he oscillation.

The only time I get oscillations when rotating large piles of docked modules (each of which has its own SAS unit) is when there's a significant off-center mass, like a lander docked on 1 side with no counterweight on the other side. This moves the axis of rotation out of line with the station's body and the thing starts wobbling during the rotation.

This makes more sense: the axis of rotation moves outside the station's body (or at least out of line with the SAS module) during maneuvers. However, the use of a counterweight doesn't seem to lend a stable solution as it creates two different situations, each conditionally unstable:

1. No counter-weight: The station is stable on SAS when the lander is docked, but becomes unstable when it is docked, OR
   
2. With the counter-weight: The station is stable on SAS while the lander is docked, but then becomes unstable when the lander is undocked as the added counter-weight becomes unbalanced without the lander.
   

My concerns are really less about maneuvering/rotating the station itself and more about dampening the movement caused when vessels dock with the station. Again, has anyone found any effective strategies to deal with this issue?

[Horn Brain]

It's kind of a hard problem. Right now, SAS works by trying to point the active "command from here" part at a set attitude (determined by when you turn on SAS, or when it leaves damping mode). If you ship is rigid, that's fine, because everything is rotating the same way. When your ship is wobbly, like a big stack of docked modules, it doesn't account for the extra inertia and momentum of the other parts properly, so the other parts act like big, floppy arms that drag your command vessel's attitude around in ways that the SAS doesn't anticipate. The SAS then starts correcting in the opposite direction, but sometimes this causes the floppy arms to gain energy because of resonance.

The best solution (it seems to me) would be to have SAS work individually from each part that applies torque to try to keep it pointing in the right direction. This would make any torque modules on the floppy arms work to damp out the rotation in the arms, and by tuning the SAS controller properly, you can ensure that the structural rigidity of the vessel will dominate in the end to keep everything pointed straight. This is not going to happen anytime soon, but in that case you would probably just want to distribute torque modules in proportion to the size and mass (rotational inertia) of the various parts of your spacecraft.

In the current situation, I think that we can do some rigid body dynamics to determine the best strategy to place torque modules. I'm working on it a little bit, but it is pretty complicated, and what's worse is we're not really guaranteed that we can help the wobbles by placing the torque modules properly. The wobbles may be an inherent consequence of the simplistic control scheme in use.

[MockKnizzle]

You wanna minimize floppy bits however possible. Use Seniors, multidocking, or KAS struts to stiffen things up. That's really the only solution.

[Vanamonde]

Stations don't really need SAS. Their mass resists disturbances, and unless you sit and watch it long enough for small disturbances to build up, any rotations will be canceled the next time you leave the station and come back again. Putting SAS on stations is a case of the solution being worse than the problem.

[maarten2208]

have you tried turning all of the torque off (in the capsules as well) and only turn one on? it will turn slowly but you won't have as much wobble.

[Poynting]

You have run into the "trying to rotate a length of chain using a single link" problem. Obviously thats an over exageration but it is the essence of your problem.

Stiffening the structure is your best bet, but if you dont want to add ugly structural parts, stick to the suggested solution of avoiding the use of SAS and using the fewest reaction wheels possible. I also recomend damping rotation yourself as the game just isnt smart enough yet to do it for you in large bendy structures. The mass of the ship should resist rotations if you are gentle with it. Or you can just use time warp...

Also Horn Brain: That sounds like pretty serious ACS work, I imagine the only real spacecraft to use something like that is the ISS.

[Kasuha]

To minimize wobbling and flexing, you need to distribute your reaction wheels evenly throughout the mass of the station. Each "branch" should have torque proportional to its mass. If the "branch" is long, subdivide it to parts and give proportional torque to each.

Below is image of my modular ship. Notice there is one reaction wheel going with each jumbo tank. This arrangement substantially reduces flexing when I reorient it in space.

Edited February 27, 2014 by Kasuha

[Claw]

To be clear, placement of the SAS modules does not affect how well they do their job as rotation. In this manner, placement does not matter.

The center of rotation is always around the Center of Mass. This is why some people experience problems when docking a vehicle.

As others have stated, ship wobble is because the station is not an ideal "rigid body" structure. Kasuha's example is a good place to start the trial and error process of dealing with a flexible body.

Unfortunately there isn't a magic spot to always put SAS that fixes this. The ship can still flex and wobble around whether the SAS modules are near the CoM or out at the ends. Even distribution generally works best because no single place on the station experiences a concentrated bending load, but that isn't always a sure fire fix. Making the structure more rigid helps too. To some extent you'll simply have to use trial and error.

[arise257]

It's not so much the SAS, I've done some pretty crazy dockings. I've found that 3 things really matter a lot. Making similarly-weighted modules, adding RCS to every module, and turning on RCS when you reorient. Whenever I don't use RCS, I get the symptoms you describe. For what it's worth, I'm able to move a space station at 1/10 thrust into and out of orbit, with RCS on. Without it, the modules over-flex and break apart.

[Kerba Fett]

I usually leave them off entirely. Stations are supposed to be stationary. Why would you need to rotate it? So a ship can dock easier? Just let it drift past the station and approach from the other side?

[Kasuha]

Stations are supposed to be stationary.

Stationary you say?

When the Soyuz module is about to return back to Earth, the whole ISS may rotate 90 degrees to release it in the right direction for reentry.

[Kerba Fett]

Stationary you say?

When the Soyuz module is about to return back to Earth, the whole ISS may rotate 90 degrees to release it in the right direction for reentry.

I didn't realize that. It sounds very accommodating of them. Perhaps the pilot isn't very good with RCS and they're worried he'll run into the station if he has to steer around it.

In any case, I don't see a need to rotate my stations in KSP since I can just fly around the back and dock.

[Superfluous J]

I didn't realize that. It sounds very accommodating of them. Perhaps the pilot isn't very good with RCS and they're worried he'll run into the station if he has to steer around it.

In any case, I don't see a need to rotate my stations in KSP since I can just fly around the back and dock.

Funny, I don't need to do all that maneuvering because I can just rotate my station