search

dkavolis / Ferram-Aerospace-Research

Aerodynamics model for Kerbal Space Program
Other
81 stars 32 forks source link

CoL marker is dodgy #104

Open BTAxis opened 3 years ago

BTAxis commented 3 years ago

I've been having issues with the CoL marker, which sometimes shows up in the wrong place, or not at all. I'm unsure what exactly the problem is, but it seems related to the undo function of the editor.

Here is one example, with a simple rocket made out of a Mk1 pod, three Fleas (for some reason) and four basic fins: image

If I delete the fins, and then restore them with undo: image

The CoL has moved to a diffferent place. If I remove the fins and put them back manually the CoL moves again, but not quite back to where it was in the first picture. Repeating the process moves the marker back to where it started in the first picture.

I've also seen other wonkiness. I've seen the CoL marker disappear upon undo, but show up again when a part is being placed (until the marker is reset using the toggle). I've also seen the marker getting completely stuck, failing to change on any change to the craft at all. A demonstration of that latter behavior, on a heavily modded KSP 1.10: image image

I've mostly tested this on KSP 1.9.1 and 1.10.1, with and without other mods installed, using CKAN to install mods.

azraelzephyrian commented 3 years ago

I've had similar problems (same KSP versions). Seems there's an intermittent voxelization problem and it's driving me absolutely nuts.... I have also had problems with the FAR aero analysis....sometimes the static analysis graphs read out completely incredible (as in: not credible) stuff. Sometimes the graphs refuse to change when I add new parts--e.g. I do a mach sweep of a MK1 pod, get a very high Cd (expected)...but when I add a nosecone and a fuel tank+engine+fins and try the sweep again, it just stays exactly the same.

When this occurs, if I go into the transonic design section and turn on the visual aid/augmented-reality-graph things, they just present as a single horizontal bar that extends past the hangar walls (for pressure coefficient and curvature cross-sectional area) or a single thin vertical bar (for plain cross-sectional area). If I delete the craft and try again, it just completely refuses to voxelize anything and static analysis fails. Kinda unfortunate, and I'm surprised to hear that my woes are probably not due to mod conflicts....