Sudden change in orbit display

[BluK]

Hello all,

I recently purchased the game and already had a lot of fun with it. I read many help posts as well so thank you for the help so far.

I'm experiencing a strange behaviour of the orbit displayed (conics I believe ?) when I'm changing from a body SOI to another. For example, assume I'm orbiting the Mun (yay !) and I want to go back to Kerbin. I'll play with the manoeuver nodes until I get a nice Kerbin orbit after exiting the Mun's sphere of influence.

I'll execute the manoeuver not too badly and then, right after I exit (end of the blue trajectory), the orbit will completely change and I don't understand why. But there's more. After a short time, the orbit will get back to what it was supposed to be.

Has anyone seen this before and could tell me what's happening ?

I can't say if the problem also exist with other bodies as I'm still at the very beginning. All my settings are default (I think), but I do have a few mods installed. If I didn't provide enough information, please ask for more.

[Snark]

Welcome to the forums!

If you mean that there's a brief "hiccup" right when you leave the SoI, just a second or two, when the displayed orbit looks wrong, before it snaps back to what it ought to be: that's a bug, it's been around since forever, and it appears to be a harmless cosmetic glitch as far as I can tell. They're completely redoing all the orbital-path display code in the upcoming 1.1 patch, so I wouldn't be surprised if it goes away then.

Or is your issue something else? Screenshots?

Edited December 31, 2015 by Snark

[MircoMars]

snark is right, that's normal behavior while changing SOIs. nothing to worry about. 

except when you dare to warp through a SOI-change, there's a good chance that the projected orbit changes drastically (though, 5x to 50x warp can be okay depending on speed)!

I use KAC with 0-5 seconds margin on every SOI change to avoid those hickups getting problematic.

[BluK]

Your description seems to match what I see.

Here are a few screenshots of the phenomenon, I left the clock in the upper left corner for you. The hiccup lasts about a minute in real time.

The third and fourth image show the wrong orbit, and the fifth is the back to normal situation.

Something interesting though, I used KAC and asked to pause 5 seconds prior to the SOI change. The hiccup happened before the KAC pop-up, and the orbit went back to normal a few seconds after.

So I just need to not touch anything during this minute and everything is fine, right ?

[Snark]

1 hour ago, BluK said:

Your description seems to match what I see.

Here are a few screenshots of the phenomenon, I left the clock in the upper left corner for you. The hiccup lasts about a minute in real time.

<pictures>

The third and fourth image show the wrong orbit, and the fifth is the back to normal situation.

Something interesting though, I used KAC and asked to pause 5 seconds prior to the SOI change. The hiccup happened before the KAC pop-up, and the orbit went back to normal a few seconds after.

I suspect that it's tied to real-time, not game-time.  How many actual seconds on your wall clock was it borked?  (One minute seems long to me; when I observe hiccups like this, it's usually only for a second or two.)

1 hour ago, BluK said:

So I just need to not touch anything during this minute and everything is fine, right ?

Yup.  Actually, since this just involves the display of orbital trajectories rather than your actual ship physics, you could do stuff during that time and it would be fine... it's just that you'd be effectively flying blind, since you wouldn't have an accurate map view to guide you.

So yeah, the answer is basically "don't worry about it."  It'll probably go away in 1.1 anyway.

[Geschosskopf]

57 minutes ago, BluK said:

Something interesting though, I used KAC and asked to pause 5 seconds prior to the SOI change. The hiccup happened before the KAC pop-up, and the orbit went back to normal a few seconds after.

So I just need to not touch anything during this minute and everything is fine, right ?

There are 2 path-changing things that happen when you're near an SOI boundary, one normal and expected behavior, the other a bug.

The normal thing is a large change in the direction of your blue path once you cross the boundary.  This is due to the game changing the frame of reference from one body to another.  This is why, in your 1st pic, your blue path is pointing back away from Mun while you're still in Mun's SOI, but once you cross into Kerbin's SOI your blue path is an ellipse pointed at Kerbin.  A side effect of the change in frame of reference is that the position of the ship has to change slightly to fit the new coordinate system, which produces a slight jerk causing the G meter to spike and slightly altering your trajectory.

The bug is that SOI boundaries are kinda "fuzzy", especially when warping.  The SOI itself is moving, tied to the body it's centered on, and so is your ship.  When you're warping, your ship and the SOI don't move consistently.  Your future path on the map assumes everything moving in synch but that's not what happens when you warp.  Thus, you can be going along thinking all is fine, no encounters with other bodies any time soon, and suddenly you're in a different SOI, because your ship and the nearby body got in different releative positions due to warping than shown by your future path.  This can happen when going to other planets but is mostly an issue when maneuvering around large moons, especially if there are several of them close together like at Jool.  Flying between Jool's moons can be very frustrating because of this but it also happens to a lesser extent at Kerbin/Mun and Duna/Ike.  Basically, any time you have reasonably large SOIs in close proximity, expect this sort of weirdness.

Part of this sort of weirdness isn't a bug per se but is a side-effect of KSP using a 2-body gravity system, and you can compensate for it but not the "fuzzy" SOIs due to warping.  Still, working around the 2-body issue will minimize your frustrations.  The 2-body issue goes like this:  Suppose you're at Laythe (Jool's inner moon) and you want to Vall (the next one out).  From a navigation POV, this is the same as going from Kerbin to Duna.  So you wait until Laythe is coming up behind Vall and then burn more or less tangent to Laythe's orbit so you arc out to meet Vall on the other side of Jool.  So, while you're still in Layth's SOI, you're going fast enough to have escape velocity and leave the SOI, meaning you're going faster than Laythe.  But then you're in Jool's SOI.  From Jool's POV, your now in a wider orbit than Laythe, which means you have to be going SLOWER than Laythe.  And the further you arc out towards Valls, the slower you go.  Thus, Laythe catches up with you from behind, and it's SOI is wide enough that if you haven't arced far enough outboard by then, you'll find yourself back in Laythe's SOI heading about 180^ from your previous course, despite having just left with escape velocity.

So,, to avoid such things, you have to NOT use the most efficient trajectory.  Instead of leaving the 1st body tangent to its orbit, you need to angle more towards the target body, so that you get far enough off the track of the 1st body that it's SOI will miss you as it catches you up.  But OTOH, you can use this issue to your advantage sometimes.  For instance, Kerbals can only get 2 stars going to Mun and Minmus, but ALMOST have 3.  You can get that 3rd star prior to sending them to other planets by briefly putting them in solar orbit.  The easiest way to do this is to burn tangent to Kerbin's orbit until you just BARELY leave it's SOI, then wait.  Shortly after leaving Kerbin's SOI, Kerbin will catch up with you and POOF, you're back in its SOI and can now land :).

[BluK]

17 hours ago, Snark said:

So yeah, the answer is basically "don't worry about it."  It'll probably go away in 1.1 anyway.

It's not too bothering anyway, I was just curious I'll see if it stills happens in 1.1. (And the hiccup did last about a minute in real time. I wasn't warping as I wanted to pay close attention to what was happening. But it's not that bad, don't worry).

17 hours ago, Geschosskopf said:

 The 2-body issue goes like this:  [...]

I had to read this twice but I got it This will probably be useful very soon. Also I knew KSP only simulated a 2-body problem but I never really thought about what it meant, I just assumed it would be easier. I think you just proved me wrong .

Anyway, thank you all for your answers !