Orbiting: Wrong physics simulation?

[Carraux]

Note: Initially a bug report but as the issue discussed isn't a bug this has been moved and turned into a discussion thread.

Johnno

Hi there,

I'm new here so forgive me, if this bug has already been posted (though I searched the bug forum, but found nothing).

The bug is quite obvious so I wonder why nobody(?) noticed it.

How it should be:

An object orbiting a celestial body with no further movement/rotation will always face the same side to the celestial body. This is caused by the centrifugal force (the gravitational pull is acting like a string between celestial body and the object. Try to attach a string to a ball and spin it around and you will see what I mean). Best example for this is Luna. The effect: If the object completes one circumnavigation, it turned around it own axis once/one full rotation.

How ist done now (wrong):

The object does not rotate at all. It stands still in the unversal coordinate system.

(pardon me if I mixed up spin/turn/rotation/circumnavigation, but English isn't my mother language)

I know that some mods (like MechJeb) depend on this behaviour, but...

Edited November 8, 2013 by Johnno
Added note

[vexx32]

Nope, not true. The Moon is tidally locked, yes, but not every object will be. A sufficiently lengthy object will behave as you described, but in order to see a noticeable effect, to the best of my knowledge, it would have to be immensely long in comparison to most things that people or Kerbals send up into orbit.

[Kasuha]

An object orbiting a celestial body with no further movement/rotation will always face the same side to the celestial body.

Erm... nope.

In reality, objects keep their rotational momentum. If we keep relativistics aspects aside, the rotational axis is fixed relative to "universe". So if the object in orbit does not rotate along axis perpendicular to its orbit with period exactly equal orbital period, or the orbit is not circular, the object will not keep facing the body it orbits.

In KSP, rotational momentum is not kept and objects keep their orientation relative to universe. That's second best approximation of the behavior.

Edit: XKCD

Edited November 8, 2013 by Kasuha

[Superfluous J]

The tidal locking (not centrifugal effect, that's a different effect) on Luna took millennia to occur. The ISS doesn't naturally do this, they give it a little rotation so Earth is always in the same place. Hubble, for example, would have quite a bit of trouble observing the universe if it rotated all the way around every 90 minutes!

[Damaske]

He is right with that all objects within KSP have there orientation relative to the universe and not to the object that they are in orbit around. I feel that he is right and that it NEEDS to be for the object that its in orbit around NOT the whole universe. Even if mods and other things need to be rewritten then so be it. I feel that the current system is dumb, BUT before you bash me. I realize this IS in development stages and things are done for a certain reason. Why they did this the way they did they must have a good reason or it would not be there. It may change it may stay the same.

Anywho, What good is it to ask others to test something in beta if nothing is questioned on how or why they did one part of the game/program. Feedback even if negative can be constructive if written right.

[Superfluous J]

He is right with that all objects within KSP have there orientation relative to the universe and not to the object that they are in orbit around. I feel that he is right and that it NEEDS to be for the object that its in orbit around NOT the whole universe. Even if mods and other things need to be rewritten then so be it. I feel that the current system is dumb, BUT before you bash me. I realize this IS in development stages and things are done for a certain reason. Why they did this the way they did they must have a good reason or it would not be there. It may change it may stay the same.

Anywho, What good is it to ask others to test something in beta if nothing is questioned on how or why they did one part of the game/program. Feedback even if negative can be constructive if written right.

I have no problem with the fact that the original poster found a bug with the game. My problem was the fact that there is actually no bug in the game in this case, and the original poster's assertions were instead what was wrong.

It is true that in Kerbal Space Program, our ships don't rotate to always face the same face at the planet. What is NOT true is that they would if they were real life space ships orbiting the Earth.

Now, one thing you CAN do in reality that you can NOT do in KSP is to give your space station a little bit of a rotation to make it APPEAR that it's naturally always oriented with the Earth. But that's all it is: An appearance. Twist the knob a different direction and your station will just as happily spin around in some other random way.

[Kasuha]

He is right with that all objects within KSP have there orientation relative to the universe and not to the object that they are in orbit around. I feel that he is right and that it NEEDS to be for the object that its in orbit around NOT the whole universe.

Have you ever seen this?

The ONLY bug in game is that ships don't keep their rotational momentum. But it's damn convenient.

[KerbMav]

The Moon is tidally locked because Earth's gravity slowed his rotation down to that point and the same is true for every moon/planet in orbit of a planet/star, it just takes a different amount of time.

[maltesh]

Nope, not true. The Moon is tidally locked, yes, but not every object will be. A sufficiently lengthy object will behave as you described, but in order to see a noticeable effect, to the best of my knowledge, it would have to be immensely long in comparison to most things that people or Kerbals send up into orbit.

A sufficiently-long object, left to orbit in reality would eventually wind up radially oriented due to tidal effect, yes.

That said, possibly because people rarely did build things long enough to make the effect noticeable, and even more rarely did they leave physics enabled for the hours to days it would take to happen, KSP stopped simulating differential gravity for the individual parts of a spacecraft back in 0.11. In KSP as it is now, all parts of a particular spacecraft behave as if feel the exact same gravitational attraction in the exact same direction, regardless of the size of the spacecraft. As a result KSP Vessels experience no tides, and cannot make use of gravity gradient stabilization by themselves.

I've done some basic experimenting with a 300m-long chain of KAS-attached craft that /seemed/ to attempt to orient itself radially after several hours in orbit, but the fact that any spacecraft chain set up so would come apart when put on rails makes the experiment of extremely limited utility in actual play. Especially since the tidal forces on that 300m craft in LKO wound up being pretty darned small.

[Carraux]

I do apologize for the misunderstandig that the cause of Lunas motion were either tidal or centrifugal. I took Luna only as an example for the kind of motion you can observe. Sorry for that.

Let's see:

If you shoot a grenade in a - say - 45° angle into the air, the grenade will follow a ballistic path and will hit the ground with its nose and not its butt, because it will turn along the flight path. This turn is accomplished by constantly adding the force of gravity to the flight vector. The grenade turns.

This turn is also present in orbital flight, it won't vanish. Of course is additional rotation possible, this is only a matter of adding vectors.

In KSP as it is now, all parts of a particular spacecraft behave as if feel the exact same gravitational attraction in the exact same direction, regardless of the size of the spacecraft. As a result KSP Vessels experience no tides, and cannot make use of gravity gradient stabilization by themselves.

Thank you very much, you explained it better than me.

If wee look at an orbiting space ship, the inner side (faced to the celestial body) has a lower orbit than the outer sides. Different orbits mean different speeds, the outer side of the ship is faster than the inner side. The size of the ship doesn't need to be big to get this effect, any size > 0 is sufficient.

KSP hasn't do much math to correct this. Only one addtional vector addition per orbiting object each frame.

Now I found this interesting thread here in the forum: http://forum.kerbalspaceprogram.com/threads/53016-Orbiting-Satellites-Orientation-Attitude-question/page2

The video there is really good.

At last: I do not discuss against the conservation of the angular momentum. I am only missinng one rotational component.

[Kasuha]

If you shoot a grenade in a - say - 45° angle into the air, the grenade will follow a ballistic path and will hit the ground with its nose and not its butt, because it will turn along the flight path. This turn is accomplished by constantly adding the force of gravity to the flight vector. The grenade turns.

The only thing on the grenade path caused undeniably by gravity is that it hits the ground again. Regarding which part of the grenade will touch the ground first, there are much bigger players in place, paticularly air drag and conservation of angular momentum law, making any gravity contribution to it negligible.

Gravity does not turn things. Tidal forces do, and space curvature induced by gravity does. But in case of normal gravity fields and normally sized objects these forces are so small that it's serious problem to measure them.

Arrows have feathers or stabilizers at their rear end because they need air drag - not gravity - to keep the head pointing forward. Similarly spears turn in air not thanks to gravity but thanks to careful balancing and air drag.

Edited November 8, 2013 by Kasuha

[Technical Ben]

I have no problem with the fact that the original poster found a bug with the game. My problem was the fact that there is actually no bug in the game in this case, and the original poster's assertions were instead what was wrong.

It is true that in Kerbal Space Program, our ships don't rotate to always face the same face at the planet. What is NOT true is that they would if they were real life space ships orbiting the Earth.

Now, one thing you CAN do in reality that you can NOT do in KSP is to give your space station a little bit of a rotation to make it APPEAR that it's naturally always oriented with the Earth. But that's all it is: An appearance. Twist the knob a different direction and your station will just as happily spin around in some other random way.

Interestingly you can add a little spin. It's just currently the game does not save spin states, so you need to add it again after every load. That's in agreement with you! Because a ship launched from erbin should not rotate of it's own accord as the OP (and those seemingly to agree) suggest it should. The game is working as intended, as follows intuition (if even possible) and as is "realistic" too.

PS, and again in agreement, AFAIK a ballistic trajectory covers the flight of the object, not it's rotation. As a hint, fire a ball out of a ballistic gun (with or without air). Colour one side of the ball one colour, the other side another. You'll notice it does not rotate. Alternatively take a balistic bullet to the moon. It will follow a ballistic trajectory, but the bullet will not rotate at all, and will land as it left the gun (facing "forwards", but as you orbit, that'll be relative anyhow depending on where you land).

Edited November 8, 2013 by Technical Ben

[Johnno]

After reading this thread and conferring The Moderator Team doesn't feel that it fits in the Bugs & Support subforum. It doesn't really fit in Suggestions and Development discussions either, so it will be moved to Science Labs.

Title will have a question mark added to it to promote discussion and a short note will be placed in the opening post to explain that the thread's been moved. OP feel free to rewrite parts of your opening post if necessary.

[Xeldrak]

An object orbiting a celestial body with no further movement/rotation will always face the same side to the celestial body. This is caused by the centrifugal force (the gravitational pull is acting like a string between celestial body and the object. Try to attach a string to a ball and spin it around and you will see what I mean). Best example for this is Luna. The effect: If the object completes one circumnavigation, it turned around it own axis once/one full rotation.

Nope - why should it turn? Gravitational pull does not work like a string...not at all - please do not use this metaphor again. It's like "We should nail horse-shoes to the tires of your car! Because cars are like horses - try to ride a horse without horseshoes regularly and you'll see what happens."

The object would only turn if the frontside would feel another gravitational pull than the backside. Wich it doesn't! And even if it would, the effect would be very hard to notice - a far cry from "allways face the same side to the celestial body".

If you shoot a grenade in a - say - 45° angle into the air, the grenade will follow a ballistic path and will hit the ground with its nose and not its butt, because it will turn along the flight path. This turn is accomplished by constantly adding the force of gravity to the flight vector. The grenade turns.

No! The grenade will hit the ground with it's nose because of aerodynamics. Why do you think genades are formed the way they are? Some of the even have fins to facilitate this! You add the gravity vector to every point of the grenade, therefore it does not turn (Same argument as above).

If wee look at an orbiting space ship, the inner side (faced to the celestial body) has a lower orbit than the outer sides. Different orbits mean different speeds, the outer side of the ship is faster than the inner side. The size of the ship doesn't need to be big to get this effect, any size > 0 is sufficient.

But you do realize, that lower orbits are faster than higher orbits, yes? Not the other way around....

Another example: The earth rotates more than once in a year! If your model was true Earth would allways has to face the same side to the sun. No days, no nights - one side a burning hell, the other one shrouded in eternal darkness.

Edited November 8, 2013 by Xeldrak

[Mr Shifty]

It is not a given that any satellite will be able to be stabilized by gravitational gradient torque. Specifically, the torque has to be high enough that it can overcome the rotational inertia of an object in orbit (and any lingering rotation the object has due it its deployment.) There are a set of criteria, based on the moments of inertia along the object's three rotational axes, that determine whether it can be captured into a stable (oscillating) orientation. You can find them listed here, on page 37. And there is a more in-depth discussion here.

[brdavis]

OK, I've just got to say as a physics professor who teaches this stuff for a living, and is writing next weeks lectures on rotation motion, torques, and angular momentum now…

This is fun. I should go get some popcorn.

--

Brian Davis

[Xeldrak]

OK, I've just got to say as a physics professor who teaches this stuff for a living, and is writing next weeks lectures on rotation motion, torques, and angular momentum now…

This is fun. I should go get some popcorn.

This is just cruel...

[sojourner]

Another example: The earth rotates more than once in a year! If your model was true Earth would allways has to face the same side to the sun. No days, no nights - one side a burning hell, the other one shrouded in eternal darkness.

Bad example, re:Mercury.

[SFJackBauer]

Orbiter (meaning the other space sim) modelled gravity gradients for in-orbit objects:

But I dont think it is a big issue for Kerbal, since in RL you would always counter this force to keep the craft (or space station) in a certain attitude with respect to the Sun.

[Xeldrak]

Bad example, re:Mercury.

Read again, Mercury is not tidally locked. Besides, I just have to come up with a single example to disprove his theory. Even if there was tidally locked planet, it wouldn't mean, that "An object orbiting a celestial body with no further movement/rotation will always face the same side to the celestial body. "

Edited November 8, 2013 by Xeldrak

[PakledHostage]

Bad example, re:Mercury.

Mercury, as is true of our moon and likely many of the known exoplanets, is tidally locked. Please see the above discussion about tidal locking as applied to spacecraft.

[Xeldrak]

Hello? Mercury is NOT tidally locked...

Try this link

Or this

[PakledHostage]

Hello? Mercury is NOT tidally locked...

Hello? Oh, hello grade 8, I thought we'd gone our separate ways?... What's that, I was wrong? Well I stand corrected. Sorry. Nice talking to you. Have a good day... [hangs up phone]

p.s. the point is the same whether it is tidally locked in a 1:1 resonance, or in some other resonance.

[Superfluous J]

Hello? Oh, hello grade 8, I thought we'd gone our separate ways?... What's that, I was wrong? Well I stand corrected. Sorry. Nice talking to you. Have a good day... [hangs up phone]

p.s. the point is the same whether it is tidally locked in a 1:1 resonance, or in some other resonance.

How is the point the same? Mercury has an odd rotation to go with its odd orbit, but it does NOT face one face to the Sun. Which is what the OP states will ALWAYS happen.

Anyway, comparing planets to space stations is like comparing hurricanes to toilets. Hint: They don't actually flush the other way in Australia.

[PakledHostage]

How is the point the same?

The point is the same because the planet or moon is rotating about its axis in a stable equilibrium. The stable equilibrium is driven by the dynamics of the system, in so far as an external torque exists that constantly restores the rotational speed of the orbiting body to a stable resonant frequency after any perturbation. In the absence of such a stabilizing torque, there is no stable equilibrium and no reason to expect the orbiting object to rotate about its axis once per orbit, three times for every two orbits, or at any other rotational rate.

Edited November 8, 2013 by PakledHostage
Added a pinch of Nitpickersbaneâ„¢