a way to prevent "skipping" atmosphere during high-speed re-entry

[thereaverofdarkness]

Well, I am quite familiar with Ludicrous Speed, so I think I have to speak up for this one (Sorry. I just had to do that). Seriously, even at 75 km/s, you wouldn't entirely "skip" the atmosphere, unless you happen to be running KSP at one frame per second (at which point I would congratulate you for getting a ship that lags your computer down to 1 FPS up to 75 km/s). Even if you are running KSP at 30 FPS, 60 being the maximum on most screens, 75 km/s would mean that you would have at least 30 frames of deceleration if you entered Kerbin's atmosphere vertically. Less, of course, on Duna, but still, my point stands. To fully "skip" the atmosphere, you would have to go 30 times 75 km/s, or 2250 km/s, at 30 FPS.

I don't know what the rate of recalculation is, but I have noticed that craft going very fast will experience very different effects from one another. Sometimes the slower one impacts harder, and other times the impact doesn't even occur. The two Duna crash entries on my challenge (by MarvinKitFox and jackkymoon) are the exception and not the rule. They both were slowed around the same amount, but in watching lots of atmosphere entry videos and doing this stuff myself, I have learned that the same speeds on entry often do not result in the same g-forces. What I am certain of is that the physics bug out at these speeds.

[John FX]

I woudn`t call this a bug, more a programming trade off. More ticks would slow down the game. Maybe the best solution would be to have the game check for boundary crossing (space->atmo) in the near future and increase ticks on the fly for the boundary crossing and when in atmosphere.

[Vaporo]

I don't know what the rate of recalculation is, but I have noticed that craft going very fast will experience very different effects from one another. Sometimes the slower one impacts harder, and other times the impact doesn't even occur. The two Duna crash entries on my challenge (by MarvinKitFox and jackkymoon) are the exception and not the rule. They both were slowed around the same amount, but in watching lots of atmosphere entry videos and doing this stuff myself, I have learned that the same speeds on entry often do not result in the same g-forces. What I am certain of is that the physics bug out at these speeds.

I don't know that it's the physics bugging out, but the g-force recording mechanism. Sometimes I'll reenter and the g-force meter will hit maximum, but when I read the Flight Results window it says that I got two or three g's. Of course, that may be just because it just doesn't take FAR aerodynamic forces into account for that number. (I play with FAR installed). Also, you say that two craft entering at different (though both extremely high) speeds may have vastly different results. Do you mean two of the same craft or two different craft? Because if you mean two different craft with different aerodynamic profiles (even with stock aerodynamics), I'm not surprised if the slower one may impact harder after falling through an atmosphere.

Still, don't get me wrong, physics bug out at these speeds.

[thereaverofdarkness]

Also, you say that two craft entering at different (though both extremely high) speeds may have vastly different results. Do you mean two of the same craft or two different craft?

They often are different craft, but using the same aerodynamic settings. Most rockets have about the same drag, but I haven't ruled out such things as being the issue. I'm fairly certain there's a lot more to it though.

[impyre]

yeah, the real problem is that if you're going too fast you skip the atmosphere entirely between physics ticks... by the time the game realizes you're in atmosphere... you've already found the ground. In any case, destruction is imminent. If you wanted to enforce some atmospheric effect occurring at speeds high enough to "skip" the atmosphere, just have the game check to see if you're "about to enter atmosphere" and if so, destroy your craft at the appropriate time.

[rkman]

I think we can consider ourselves lucky that collision detection with a planet's surface still works at such a speed.

[elloopa]

I think we can consider ourselves lucky that collision detection with a planet's surface still works at such a speed.

Well, Scott Manley did prove that there are ways (cheating ways, mind you) to fling yourself at Jool and be flung back at a tenth the speed of light.

[stupid_chris]

Honestly, I'd rather not. As far as I know, the calculations are done once every physics frame (as it should be), doing it more often would only slow you down for no real purpose. Even though you're not asking to slam down into a planet at ludicrous speeds, it's what it implies. If you reach speeds high enough that this situation would be problematic and reenter any atmosphere fast, you would not survive if the drag was actually correctly calculated. I takeit as a buggy way of saying (you exploded on reentry).

Not seeing any good reason for it to be worth it.