Map & Tracking Station Wrong trajectory prediction through SOI changes for many bodies (affects planetary return missions)

[Draradech]

KSP2 0.1.1.0.21572   -   Windows 10 Home 21H2   -   Ryzen 9 3950X   -   RTX 2080 Ti (Driver 531.18)

 

Summary:
Predicted trajectories leaving the SOI of most bodies are incorrect. This affects the manuever plan trajectory as well as the trajectory displayed after burn. The trajectory flips to the correct one after leaving the SOI.

 

I noticed that the map view camera rotates by the same amount as the perceived wrong angle of the trajectory relative to the bodies affected when switching focus between the vessel and Kerbol (or the parent body for moons). The following are the approximate rotations relative to the parent body, these match my experiences for Kerbin, Mun, Minmus and Duna, as well as the screenshots of others I have seen in this thread:

Spoiler

Moho	-85
Eve	-15
Gilly	-90
Kerbin	0
Mun	0
Minmus	-115
Duna	-135
Ike	0
Dres	-10
Jool	-50
Laythe	0
Vall	0
Tylo	0
Bop	-35
Pol	-15
Eelo	50

(Note that I had the Lazy Orbit mod installed to get to all the bodies easily. The observed behaviour matches an unmodded installation for all bodies I have been to so far.)

 

Original report:

KSP2 0.1.0.0.20892   -   Windows 10 Home 21H2   -   Ryzen 9 3950X   -   RTX 2080 Ti (Driver 528.49)

(Save file link removed, this affects all craft).

History: This craft launched from Kerbin, landed on Duna, picked up a stranded Kerbal, and launched from Duna to the east. Now in 200km Duna Orbit.

Problem: As can be seen in the video, a maneuver node set up to leave Duna retrograde leads to a planned trajectory of a Kerbol orbit with an AP further out than Duna and PE at Duna (this is wrong). This will also be the case for the displayed trajectory after the burn, right until the craft exits Duna SOI. At Duna SOI exit the trajectory flips to the correct one, with a Kerbol AP at Duna and PE further in (see images - images from a different run than video).

 

Edited Thursday at 06:43 AM by Anth12
Added additional information, changed title

[LameLefty]

There are multiple threads on this bug since shortly after release. I was hoping it would be fixed in Patch 1 but it wasn’t. 

[Scarecrow71]

This morning I plotted a maneuver to get from Duna back to Kerbin.  You can see in the first 2 screen shots the plotted trajectory and the actual trajectory after the burn; I'm a bit off on the actual trajectory due to my fingers not working fast enough, but I am ok with this as I should be able to do a course-correction burn once I get into Kerbolar orbit.  Unfortunately, once I exit Duna's SOI, I end up on a course headed for Dres, and not for Kerbin.  Absolutely nothing happened between burning the trajectory and exiting the SOI, other than warping to exit the SOI.  I checked multiple times to make sure that I didn't warp too far (you know, getting too far around the trajectory to actually miss Kerbin or some such), and nope.  As soon as I exit Duna's SOI, the trajectory changes from a Kerbin intercept to trying to intercept Dres.  And while I want to see the rings, I do NOT want to do this right now.

[Scarecrow71]

And a video to show what is happening.  You will notice in the video that, when moving to the SOI exit, there is no dotted line between SOI and the orbit/trajectory until well after it should appear, and even then it's calculated way off.

 

[The Aziz]

No, upon your SOI exit it's calculated correctly based on your maneuver. It's the initial prediction that's wrong. And it's been there since day 1, reported multiple times.

[Scarecrow71]

5 minutes ago, The Aziz said:

No, upon your SOI exit it's calculated correctly based on your maneuver. It's the initial prediction that's wrong. And it's been there since day 1, reported multiple times.

Well, regardless of whether it's the initial burn or the actuality afterwards, it's wrong.  And, well, squeaky wheel gets the grease and all. 

[alohl669]

1 hour ago, The Aziz said:

No, upon your SOI exit it's calculated correctly based on your maneuver. It's the initial prediction that's wrong. And it's been there since day 1, reported multiple times.

Ok, I have this same issue with the initial prediction. I was exiting minmus and then I saw myself orbiting kerbol  I lost Jeb

[Starhawk]

Multiple overlapping bug reports merged.

[Scarecrow71]

So, a bit more digging into this.  The initial orbit I had going there, if you look straight down on Duna's north pole, was moving clockwise around Duna.  So I reverted to a save where the probe was on Duna's surface and got back into low Duna orbit going counter-clockwise.  You know, to see if that changed things.  Nope.  Still borked.

[Razor235]

So it seems to consistently affect the bodies further out from their respective systems: Minmus, Duna, and Jool for sure.  Has anyone encountered it on Eeloo, Dres, or Jool's outer moons?  What about Eve and Moho?  Can't say I've bothered trying to go interplanetary while this bug is still around.

Edited March 21 by Razor235

[LameLefty]

I have not encountered errors great enough to notice when leaving the SOI of any planetary moons for the SOI of the parent body.

By contrast, I have encountered the issue repeatedly on transfer burn from inside one planetary body to another planetary body elsewhere in the system. Interestingly, the errors seem *much* greater when burning retrograde from an out-system body moving in-system. This past weekend I transferred from Kerbin to Duna, and from Kerbin to Jool. In both cases, it required well under 100 m/s in correction/refinement burns to setup a good encounter.  However, the converse was vastly different - in both cases, I needed substantial correction burns to create proper Kerbin encounters for the return, Jool requiring about 730 m/s, Duna needing closer to 300 m/s.

I also disagree with the notion that the error is the display of the trajectory within the SOI rather than being associated with the change of SOI itself.  In the case of my Jool return experiments, I have been returning within hours of the calculated optimal phase angle, and my plotted maneuver node burn ends up within 20-30 m/s of the optimal burn to setup the encounter back to Kerbin. After that burn, the trajectory looks absolutely perfect, a textbook KSP planetary transfer of the kind of I have been making for literally a decade.

AFTER the SOI change, however, the trajectory is completely borked, including an Apoapsis higher than the orbit of the planetary body I’m leaving, which is physically impossible for a retrograde return burn. To me, that seems very much like a mathematical error occurs during the transition between SOI’s, and the magnitude of the error increases the further out-system it occurs.

[Starhawk]

My experience with this bug indicates clearly that the bug is that the wrong trajectory is shown until you exit the SOI of the distant planet.

I first experienced it on my first return from Duna.  My standard procedure for this is to use an online tool to find the transfer window.  Warp to that date and set up a node.
I set up the node to leave Duna in the direction I know the ship needs to go to get a Hohmann back to Kerbin.  I adjust the position of the node on my Duna orbit (rather fiddly) so that my trajectory is leaving the SOI in exactly the opposite direction that Duna is travelling in its orbit.  This involves a bit of zooming out and zooming back in.  As I said, fiddly.  But it works.

So imagine my surprise when I zoom out much farther to see just how close I'm going to come to Kerbin and I see the trajectory taking me out much farther from Kerbol.  That is exactly the opposite of what I expected.  I could see this was a bug.  It seemed far more likely to me that the trajectory shown after leaving Duna's SOI was displayed wrong rather than actually being calculated wrong.  This turned out to be the case.

If you know which direction to burn and ignore what the map tells you will happen after leaving the SOI of the planet you're leaving it will all be good.  If you believe what map mode shows, it could be that you will not return from space today.

The map display appears to add my velocity on leaving Duna's SOI to the orbital velocity of Duna rather than subtracting it.  Perhaps a minus sign slipped?

Happy landings!

[Scarecrow71]

7 minutes ago, Starhawk said:

My experience with this bug indicates clearly that the bug is that the wrong trajectory is shown until you exit the SOI of the distant plane

I would believe this were it not for having done burns on both sides of Duna, traveling both clockwise and counterclockwise.  And in all cases, the trajectory changes as soon as I leave the SOI.

In the long run it doesn't matter apart from it being a bug and needing to be fixed.

[Starhawk]

7 minutes ago, Scarecrow71 said:

And in all cases, the trajectory changes as soon as I leave the SOI.

Yes.  The bug is that the wrong trajectory is shown for your path outside the SOI until you leave the SOI.  After exiting the SOI the correct trajectory is shown.

 

7 minutes ago, Scarecrow71 said:

I would believe this were it not for having done burns on both sides of Duna, traveling both clockwise and counterclockwise. 

Perhaps I can still convince you.  The direction of one's orbit around Duna does not make any difference (other than where you place the node).

If you are looking down on Duna's north pole and Duna's direction of travel around Kerbol is 'up' on your screen, then your predicted trajectory away from Duna orbit should be 'down' on your screen.  This is when viewing the map within Duna's SOI.

 

32 minutes ago, Scarecrow71 said:

In the long run it doesn't matter

I must say that I think it does.  This is a particularly nasty bug since it will completely mislead the player about the result of doing an engine burn.  Fixing this is critical to allowing new players to understand orbital mechanics.

Anyway, the upshot of all of this is that if you make your burn in the correct direction it will give you the expected result.  It just doesn't look like that in map view until you exit the planet SOI.

Happy landings!

[Scarecrow71]

27 minutes ago, Starhawk said:

Anyway, the upshot of all of this is that if you make your burn in the correct direction it will give you the expected result.  It just doesn't look like that in map view until you exit the planet SOI.

Take a look at my video above.  I did make the burn in the correct direction, and it still pushed me to Eeloo after leaving Duna's SOI.

[LameLefty]

38 minutes ago, Starhawk said:

Yes.  The bug is that the wrong trajectory is shown for your path outside the SOI until you leave the SOI.  After exiting the SOI the correct trajectory is shown

I disagree - please see my photos and read my posts about this. I made my Jool return burns on nearly the “perfect” optimum phase angle and with nearly the same dV as suggested by the long-standing available KSP transfer window calculators. In every case, the results of the burn show a nearly perfect Kerbin return transfer, exactly as expected based on a decade playing KSP. 

As soon as I leave the Jool SOI, the trajectory morphs to something physically impossible, including an Apoapsis greater than the orbit of Jool, which indicates an mathematical error occurs when translating the relative velocity as one passes out of the planetary SOI. 

[Starhawk]

37 minutes ago, Scarecrow71 said:

Take a look at my video above.  I did make the burn in the correct direction, and it still pushed me to Eeloo after leaving Duna's SOI.

I looked at the video.  It clearly shows that your trajectory is leaving Duna in the exact opposite direction from that needed to head back to Kerbin.

Here's a screen grab I took from your vid.

In this image Duna is moving in its orbit from left to right.  If you want to go back to Kerbin your trajectory must be going to the left.  Not to the right as shown.  Your node is exactly on the wrong side of your Duna orbit.

Happy landings!

[Scarecrow71]

Ah, so you are correct then.  I have to burn on the dark side of Duna (as per my video), increasing the Ap on the light side of the planet.  I am burning in the wrong direction.  Unfortunately, the game shows my burn that way as going higher than Duna (away from Kerbin) as opposed to going towards Kerbin, so that's confusing.  I'll have to do a burn to get out of Duna's SOI and then a correction burn to get to Kerbin once I'm in Kerbolar orbit.

Thanks for the help.  It's still a bug, but at least I've got something to go on to at least try!

[Starhawk]

1 minute ago, Scarecrow71 said:

Unfortunately, the game shows my burn that way as going higher than Duna (away from Kerbin) as opposed to going towards Kerbin, so that's confusing.

And that is exactly the bug.  The fact that it's so confusing is what makes it a particularly nasty bug in my view.

Happy landings!

[ChromeDome]

Whenever I try to escape Duna, making a maneuver node so that you eject in the retrograde direction would make it say that your orbit would go HIGHER, and vice versa as if Duna itself was in a retrograde orbit, which doesn't make any sense. I have only encountered this issue with Duna

[Scarecrow71]

Related to the following.  Which should probably be stickied.

 

[The Aziz]

12 minutes ago, Scarecrow71 said:

Which should probably be stickied

Hmm, agreed. @Starhawkperhaps? Issue as common as decaying orbits. I know peoples rarely read what's in front of them but worth a shot I guess.

Edited March 22 by The Aziz

[Starhawk]

Another bug report has been merged into this one.

[Ambious]

So I was trying to get back from Duna to Kerbin, and the maneuver nodes didn't make sense. The only way I could get back to Kerbin according to those was to go prograde relative to Duna's orbit, which is of course wrong.
This is how it looked:

After I accelerated as the node suggested, it still looked like it would actually work:

But as soon as I left Duna's SOI, the orbits corrected:

(I instead made an educated guess on how much to accelerate in the correct direction (retrograde compared to the planet's orbit) and managed to get back to Kerbin, but this is obviously not expected behavior)

[Scarecrow71]

This has been reported and several topics merged into the following thread:

I will ask, again, that the mods PLEASE sticky this thread as it is now 3 pages down.  Far easier for people to find it stickied.

[Starhawk]

Overlapping bug reports merged yet again.