Using Centre of Mass, Centre of Drag and Centre of Thrust

[LukeTim]

I have yet to use these three new features in 0.17... and that is primarily because I don't actually know how to use them to help me in building a rocket. I don't know where my centres ought to be for a well designed rocket...

Does anybody know and can help us non-geniuses out by giving us a little guide?

[Letum]

Centre of mass and centre of thrust must be vertically in line with each other in a rocket.

In a plane centre of mass and centre of thrust must be horizontally in line with each other in a rocket and the centre of lift should be close to the centre of mass.

[um3k]

From what I've found, it works more or less like this:

Center of mass doesn't require any special positioning in most cases.

Center of lift (not drag) should be as close to the center of mass as possible. This is more important for a spaceplane, rockets are generally better off with the center of lift near the bottom.

Center of thrust should be positioned and possibly angled as such that the arrow points toward the center of mass. This reduces unnecessary torque on the craft, thereby making it easier to steer. In most cases, this will happen naturally. It's only necessary to do special things with the center of thrust when you're building complex, non-symmetrical vehicles.

[HarvesteR]

The center of lift indicator is a bit of an oversimplification of a more complex problem actually. It is calculated by assuming the ship is flying forward at a set speed, and querying all lift-producing parts to find their lift forces given that same velocity. The thing is, those forces will vary depending on the angle of the wing surfaces to the airflow, so the information you get from that single node doesn't really tell the whole story.

The best way to use the CoL indicator is to grab the entire ship by the command pod, and use the Shift+WSADQE keys to rotate it around. That way, you'll be able to see the CoL changing to indicate the position and direction of the net force acting on the ship, for any orientation. Think of it like holding a paper airplane against the wind at different angles, to feel how it reacts.

If your vessel has the CoL passing directly through the center of mass at some orientation or another, well done. Your ship is very stable. If not, there is another way to test for stability: If by pitching the ship up and down you find a point where the CoL changes direction, that means that there is an angle of attack where the lift forces balance out, and you are also stable.

When I was writing the code for these indicators, this "problem" became apparent. We realized here that the ideal way to visualize the aerodynamics of a vessel would be to not have a CoL indicator at the construction facilities, but to have a proper dedicated wind tunnel facility, where you load in a vessel to test against simulated airflow. It's not only a better solution for testing, it's also a lot more fun too.

Although as you can imagine, it's by no means a small feature. We decided to add the CoL indicator on the VAB and SPH now as they are, and keep the wind tunnel on our feature wishlist for now.

Cheers

[draeath]

As well, if you detach the lower stages, the center of mass will tell you where to put your RCS thrusters. They should be centered around it - eg two groups, one equally above the center of mass, and one below by the same distance. This is because the center of mass is the center of rotation.

Also note if you have not yet attached boosters/engines the Lift point will stay in the center of the VAB floor, even if you move the rocket. As soon as you stick some engines on there it will correct itself.

[RoboRay]

Center of lift (not drag) should be as close to the center of mass as possible.

Center of lift should be close to but must be slightly behind the center of mass, if KSP behaves like the real-world. This imparts positive stability, with lift-induced drag weather-vaning the aircraft to keep the nose out front.

Putting the CoL on the CoM creates neutral-stability, so you can literally spin the plane around all over the place, regardless of which direction you are trying to go. That's great for fighters, which do put the CoL very close to the CoM, but not so good for passenger or cargo craft, which keep them much further apart for a more stable ride.

Putting the CoL ahead of the CoM results in negative stability, where you will find it almost impossible to keep the nose pointed in the right direction.

Edited September 27, 2012 by RoboRay

[draeath]

Putting the CoL ahead of the CoM results in negative stability, where you will find it almost impossible to keep the nose pointed in the right direction.

This is how the F16 works. The flight control computers (there are two!) constantly combat the instability. To turn, it actually 'relaxes' constraints in a way that lets the plane spin itself in the correct direction. While it is less stable - hence the computer control - this means it can execute stupidly quick maneuvers as the plane "wants" to spin and point backwards.

[RoboRay]

Yes, like I said, fighter aircraft do tend to put the CoM and the CoL together, or even rarely reverse them like in the F-16. But we don't have those computers in KSP. If you want a stable craft here, move the CoM aft.

[Banbite]

Yes, like I said, fighter aircraft do tend to put the CoM and the CoL together, or even rarely reverse them like in the F-16. But we don't have those computers in KSP. If you want a stable craft here, move the CoM aft.

With some cleverness you could get Mechjeb work that thing out for you, might even be a nice challenge for modders to make a working fly-by-wire plugin that works the same way as the computers in an F-16 (and every other modern Jet Fighters for that matter).

But the only thing I miss from the current CoM system is that it only shows it's position when all fuel tanks are full, you have to eyeball how it moves when they start to get depleted. Would be nice to have a slider which you can move to see how the CoM moves with the fuel consumtion.

[draeath]

The Advanced SAS can do that. Between the built-in gyros, RCS (if you turn RCS on, ASAS can control it), and control surfaces such as wings...

I bind "hold SAS" to my joystick trigger. Push in the direction you want to maneuver, and hold SAS (when it's toggled on, this will temporarily disable it). The ship should "spin" or "fall" into the turn, as the SAS/RCS should from you should nudge it into the right direction.

Worth a try!

Edited September 27, 2012 by draeath

[quarthinos]

Would be nice to have a slider which you can move to see how the CoM moves with the fuel consumtion.

Or even just for stage N...

Edited September 27, 2012 by quarthinos

[VincentLaw]

Center of lift should be close to but must be slightly behind the center of mass, if KSP behaves like the real-world. This imparts positive stability, with lift-induced drag weather-vaning the aircraft to keep the nose out front.

This is correct. Basically the further aft your center of lift is, the more stable your airplane is. The problem with this is that as your center of lift moves back, your maneuverability decreases and you may not have enough control authority to trim for level flight anymore. This can be solved by increasing the size of your elevators or giving them a longer moment arm. Basically it does not matter if your center of lift is not on the center of mass to fly level, it just matters that the pitching moment at the center of mass is zero, and you want the pitching moments of disturbed flight to return the airplane to its original orientation. This would work even if your airplane had negligible drag because the lift (and therefore moment) of your wings and stabilizers is a function of angle of attack.

Also, since fuel usually drains towards the back of the airplane in KSP, it is not a good idea to place the center of lift directly on the center of mass during construction, because your airplane may be stable at takeoff, and then after a while of flying become unstable and crash into the ground. You want your center of lift to be just behind the furthest aft possible center of mass.

Edited September 27, 2012 by VincentLaw

[thorfinn]

The Advanced SAS can do that. Between the built-in gyros, RCS (if you turn RCS on, ASAS can control it), and control surfaces such as wings...

I bind "hold SAS" to my joystick trigger. Push in the direction you want to maneuver, and hold SAS (when it's toggled on, this will temporarily disable it). The ship should "spin" or "fall" into the turn, as the SAS/RCS should from you should nudge it into the right direction.

Yeah, I'm quite amazed by ASAS really. It managed to fly a (nerve-wracking!) orbital insertion for a fairly aero-unstable rocket boosted spaceplane:

This thing will fly IF you always keep the angle of attack below about 3 degrees, at all times. You let it grow larger, and it suddenly departs. With lots of small corrections and a very careful gravity turn, you can loft it into orbit.

[Fuzzy Dunlop]

The important thing about centre of lift is that it moves.

This plane here (the X-2) is perfectly stable. It flies until the fuel runs dry without any control imput or ASAS. In fact neither of those would help becuase it has no control surfaces.

If we look in the spaceplane hanger we see that the Centre of Lift is in front of the Centre of Gravity. The tail is angled downwards and generates no lift. This makes the plane pitch up.

Let's rotate the plane in the hanger and see what happens to the Centre of lift.

Now the tail generates some lift. The Centre of lift moves backward behind the Centre of Gravity. This makes the plane pitch down, and brings us back to the start. In flight the plane will bob up and down till it finds a happy medium - this is the essence of stability.

[air805ronin]

This is how the F16 works. The flight control computers (there are two!) constantly combat the instability. To turn, it actually 'relaxes' constraints in a way that lets the plane spin itself in the correct direction. While it is less stable - hence the computer control - this means it can execute stupidly quick maneuvers as the plane "wants" to spin and point backwards.

Aside: The F-16 with a complete systems failure turns into a rather expensive lawn dart. Computers and engines. It has happened. Without computers a pilot can try to fight it, but without the engines.... Coincidentally exactly like my only two spaceplane attempts, except those have engines on.

[fireblade274]

sO how about that wind tunnel?

[gmpd2000]

2 and 3/4 years old.

That's actually impressive.

[sal_vager]

Closing.