Predicted Periapsis

[Razor235]

I'm having a lot of trouble figuring out the proper PE altitude for an aerocapture on various planets. I know the actual altitudes, and how to get there. The problem is, the predicted altitude for the periapsis upon entering an SOI never matches with what actually ends up occuring. On my first attempt on Eve, I brought the predicted PE to 60 km, but the hyperbola ended up passing through the planet itself. No problem, I'd simply compensate for the planet's radius. This worked well enough for Eve, Jool, and Laythe. However, on a recent Duna mission, I attempted the same thing. Somehow, I ended up with a PE higher then what was predicted. Thinking I screwed up, I attempted to lower the PE from the same maneuver node. This ended up being even higher then before. I eventually had to resort to hyperediting the PE after an hour of frustration.

So, my question is, what the heck is the predicted PE centered around?

[rhoark]

I think there are just unpredictable hiccups when crossing SOI boundaries.

[rodion]

I think there are just unpredictable hiccups when crossing SOI boundaries.

This definitely tallies with my experience. Unfortunately, KSP can only resolve maneuvers and predictions to a certain degree of accuracy, especially when interplanetary distances are involved. The best thing to do is just make sure you check how the motion has changed as soon after you enter the body's SoI as possible, so you aren't deep in its gravity well and it only take a few m/s correction to make the necessary tweaks that are required.

Edited September 29, 2013 by rodion

[Kasuha]

These differences are caused by small arithmetic errors when swapping spheres of influence. Don't rely on just single burn, you need to apply corrections - at least right when you enter the SOI, but sometimes it's better to make a stop or two on the way. If you do it right you shouldn't need more than a few RCS bursts.

[arq]

As others suggested, if you drop timewarp near SOI boundaries this will help immensely.

[Razor235]

OK, will do. Thanks guys!