a circular orbit has no apoapsis/periapsis!

[rainerd66]

Dear Squad!

As a matter of fact, a circular orbit does not have an apoapsis or periapsis, because all points of the circle have the same distance to the planet!

This is even more true for computer simulations with limited precision.

The math used to yield such points in a circular orbit will definitely create problems that can be identified by jumping points due to the finite precision of computer calculations!

Many mods need these points for their function nevertheless.

So, why don't you replace apoapsis and periapsis with some useful values if apoapsis - periapsis < 2 * precision?

You could avoid these flickering and mods are happy.

Just an idea!

Edited October 19, 2014 by rainerd66

[Paolino]

circular orbits does not exist

[vexx32]

Yeah, actually achieving a circular orbit is quite difficult, if not bordering on impossible. The flickering is, however, a bit annoying. I would personally replace the Ap/Pe points with Ap/Pe regions, where all points on the orbit with the same lowest/highest altitude would be highlighted on the orbit path, perhaps.

[Godot]

As Paolino said ... I don´t think you will be able to complately nullify the difference between APO and PERI... there will always be a small difference, even if it is just a few mm

[Commander Jebidiah]

Yeah, actually achieving a circular orbit is quite difficult, if not bordering on impossible. The flickering is, however, a bit annoying. I would personally replace the Ap/Pe points with Ap/Pe regions, where all points on the orbit with the same lowest/highest altitude would be highlighted on the orbit path, perhaps.

I beleive this is what the orbit line in .13.3 had, the Ap half of the orbit was green and the Pe half was blue.

[Geneborg]

Well, even if it is difficult achieving perfect cirular orbit, i get close often enough that the Ap/Pe Indicators start to jump widely, so it is indeed an issue.

[Starwhip]

I beleive this is what the orbit line in .13.3 had, the Ap half of the orbit was green and the Pe half was blue.

Exactly. Why did they get rid of it?

I always use the flickering to tell me when I'm close to the right orbit.

[Red Iron Crown]

The flickering AP/Pe markers in near circular orbits is a bit annoying. Not game breaking or anything, but annoying. Hiding the markers when the orbit is close enough to circular is a good idea, would remove that distraction.

[Technical Ben]

Exactly. Why did they get rid of it?

I always use the flickering to tell me when I'm close to the right orbit.

http://xkcd.com/1172/

[r_rolo1]

Exactly. Why did they get rid of it?

I always use the flickering to tell me when I'm close to the right orbit.

IIRC they took that out because it interfered with the manouver nodes ( or something similar )

[rainerd66]

circular orbits does not exist

In reality it is difficult, but in a computer simulation with finite precision it is possible, because you only need an orbit near enough to circular!

And it is so easy to solve...

EDIT: In a computer simulation like this, the nearest circle is always inside a torus with the thickness of the precision.

Edited October 19, 2014 by rainerd66

[Andrew Hansen]

Obviously, the apoapsis and periapsis simply need to freeze their positions when a certain degree of accuracy is achieved. You don't want to suddenly lose your information of how high the orbit is, but you also don't want flickering, so freezing the orbital nodes makes sense to me.

[swiftgates24]

IIRC they took that out because it interfered with the manouver nodes ( or something similar )

The colors were removed before maneuver nodes. They were removed in .15 when patched conics came out.

[magnemoe]

In reality it is difficult, but in a computer simulation with finite precision it is possible, because you only need an orbit near enough to circular!

And it is so easy to solve...

Orbits is measured in meters, rotating your craft will change this unless your root node is at center of mass.

And no you don't want circular orbit, they are a mess to work with near circular ones.

[rainerd66]

Yeah, actually achieving a circular orbit is quite difficult, if not bordering on impossible. The flickering is, however, a bit annoying...

It is not only annoying, but mods that use their values get under stress

[Tery215]

So...

How do you display the altitude in the map mode?

[sal_vager]

So...

How do you display the altitude in the map mode?

Mouse over the node, or click it a few times to "pin" the readout

Oh from your craft? Mouse over that!

[problemecium]

You can also click the "i" button on the right to show your altitude.

[NotExactlyRocketDesign]

In order to avoid breaking a lot of stuff, it seems like the game would still need an Ap value and Pe value, i.e., storing the data for circular (or near-circular) orbits in a different way from elliptical orbits would mean that any game function or mod that tries to read orbital data would have to account for two different formats. Even hyperbolic or suborbital trajectories have both an Ap value and a Pe value, so it seems like this is probably a deep part of the game's code.

However, they could eliminate the flickering by changing the way that the orbit is displayed. Eliminating the Ap/Pe markers for near-circular orbits would be undesirable, because without mods, that's one of the main ways of determining your altitude (yes, you could still read off your altitude at the top, but it's not as precise, and I personally would find it frustrating to lose that source of information). Instead, I think you could keep the markers, but do the following:

1. As soon as Ap - Pe < epsilon, consider the orbit circular. The orbit will still have an Ap value and a Pe value, but the display will be different.

2. Instead of displaying the Ap and Pe markers, display two "Circ" markers (or whatever). These appear wherever the Ap and Pe were when the orbit circularized, and they do not move until Ap - Pe > epsilon. So the "true" Ap and Pe values might be swinging around wildly, but the displayed Circ markers are frozen in place.

This has the advantages of (a) preserving the way the orbital data is stored, ( continuing to display altitude markers for circular orbits, and © eliminating flickering.

[rdwulfe]

Even Orbiter has some glitchyness if you get REALLY close to a 0.0 Eccentricity, it'll twitch the AP and PE a little bit, if I remember right. Or it might've been an older version, but still.

This is all due to floating point, and how "innacurate" it is in Unity.. which is to say, still a bit more accurate than you'd want to do by hand. It's a symptom of the updates of the orbit, based on ship orientation and such. *shrug*

[Whirligig Girl]

You can't have a circular orbit in real life. The Moon or the uneven pull of the Earth's gravity, or a planet in the Andromeda galaxy will pull your ship off course from a circular orbit. If you ask me, the satisfaction you get from the flickering of a circular orbit in KSP is way more important than the annoyance it brings visually. Sort of like NaN for inclination.

[Whirligig Girl]

The colors were removed before maneuver nodes. They were removed in .15 when patched conics came out.

They were removed in 0.15 when Patched Orbit Splines came out. "Patched Conics" refers to the underlining mechanical system that KSP uses for gravitation, not the method that is used for predicting/connecting the lines together.

[BlackholeZ]

The flickering is because the minutest changes on the craft can greatly affect the location of the apo/peri. Next time you're in a circular orbit turn off RCS and SAS and it will calm down a lot. Beyond that it's the fidelity of the location (x,y,z) combined with the velocity vector that causes the jumping.