[#9878] Wrong orbit prediction for close encounters

[cami]

I added this issue to the KSP tracker, please add comments there:

http://bugs.kerbalspaceprogram.com/issues/9878

Version: 1.1.2 Build 1104160 32-bit Steam on Linux 64-bit (also 1.1 and 1.1.1)

Description: there is a frequent (occurs most of the times) error in encounter-prediction display on the map view when a fly-by is very close, featuring a (real) trajectory just above or even intersecting the geosphere. The predicted encounter node is way too late, only briefly before escaping again. The resulting fly-by prediction is completely wrong. The prediction error does not affect the physical orbit, not even in warp mode. The prediction error also affects the flight plan display unless a maneuver node exists near the encounter (see example/workaround below).

Screenshot: solid gold = wrong, dashed purple = correct

Example: The screenshot illustrates this. The solid cyan trajectory is a transfer orbit from Low Kerbin Orbit to the Mün (periapsis 75 km, apoapsis 11 180 km). You can also see the encounter node directly at the apoapsis. The golden orbit shows the (wrong) Mün fly-by with the moon focused. It's obviously flawed, the encounter shown when the orbit is already leaving the Mün. The golden maneuver node was added directly before making the screenshot (without subsequent orbit modifications). Note that it's 0 dV. The predicted result is an immediate lunar encounter at the maneuver node (shown in gold, barely visible). The dotted purple line shows the (approximately correct) Mün fly-by resulting from this encounter (periapsis 12 km). During this (correct) fly-by, the Mün overtakes the spaceship (in the Kerbin reference frame), then is overtaken again by the spaceship as it is flung around the moon.

Workaround: Strangely, adding a zero-delta-V maneuver node on the orbit just before the real encounter shows the correct trajectory in the flight plan. To find the right spot, just add the maneuver node and move it a bit until you see the trajectory change. Note that if the maneuver is after the encounter, the predicted trajectory will be wrong, because the target's gravity is not taken into account before the maneuver node.

Files: https://filebin.net/ofk6u3i1fqkn0wxd/quicksave__154.sfs

Save files were created using this version of the game (not necessarily this build) and only used on this platform in "C" locale (C locale was set for steam and KSP only). The example features the vessel with pid 361d85618d174f7d8b76a9d68c226df0.

Specifications:

• KSP.x86 v1.1.2 Build 1104160 Steam
• OpenGL vendor string: NVIDIA Corporation
  OpenGL renderer string: GeForce GTX 275/PCIe/SSE2
  OpenGL core profile version string: 3.3.0 NVIDIA 340.96
  OpenGL core profile shading language version string: 3.30 NVIDIA via Cg compiler
  OpenGL core profile context flags: (none)
  OpenGL core profile profile mask: core profile
  OpenGL core profile extensions:
  OpenGL version string: 3.3.0 NVIDIA 340.96
  OpenGL shading language version string: 3.30 NVIDIA via Cg compiler
  OpenGL context flags: (none)
  OpenGL profile mask: (none)

• Linux 4.1.15-gentoo-r1 x86_64 Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz

Edited June 8, 2016 by cami
Issue tracker referral

[sal_vager]

Hi @cami, please report issues to the KSP issue tracker, they are not copied from the forums.

Thank you.

[cami]

Done. Sorry, my last report was still in alpha.

Please close the topic.

[Superfluous J]

Let me sneak in a "Thanks for the maneuver node trick" before this is closed.

I'm pretty sure this bug has been reported before (I think it may be part of a bevy of weird patched conics things) but that "trick" may not only help me keep my sanity, but find the underlying prediction problem as well.