Open Lisias opened 3 years ago
This is a test of stability of landed vessels wihout KSP Recall
I reproduced a setup I had used in the past: two kerbals, two crafts.
And then a 3rd big one on the scene, to stress the physics engine.
About 10 minutes later, the Kerbals and crafts are still on their places:
But the big craft has drifted a lot...
This suggests me that;
IMHO a proper fix would be clamping the torques on every part on the Physics engine itself, instead of zeroing them on the top level. But... I'm not the KSP developer, and I don't have all the information - so I can't criticize the solution . :)
edit: here, a KSP Team member explains the reasoning for this behaviour. Something on the PhysX engine demands a brute force solution to prevent drifting - but Squad forgot to apply it on the torque too!!!
Now I installed KSP Recall to check if it enhances, worse or just dry run on the same setup.
First, I replace the big craft using VesselMover (it's working fine, by the way)
Then I leaved the game alone for 10 minutes.
And then...
Well, the big craft drifted again, but WAY LESS than the stock solution on the Yaw axis....
Veredict: KSP-Recall Driftless is not needed on KSP 1.11, but it's still useful on borderline situations.
I will not prevent it from installing itself on KSP 1.11.
On the other hand, since Drfitless can still be useful on KSP 1.11, I think it is a good idea to further investigate if there's a way to still act when the physics engine is overloaded
Post edit: Stock KSP is heavily yawing the craft.
On Driftless, the yawing is way less severe, but the craft drifted downhill (or down airstrip). This suggests Driftless would make a good use of a "anchor".
Squad Team provided an pretty good explanation about this here.
So... I changed my mind, I will reopen this and see how Recall could help.
Interestingly, the Stock anti-drift works most of the time - but not always.
SOME DRIFT IS STILL THERE
I just launched this monster, leaved it parked the whole morning and came back to see what I got.
The position appears to be allright, but the thing is still drifting the heading...
The heavy drifting reported on KSP 1.12 is due a less than ideal decision on the default values for some new code that, effectively, solved the problem.
So I'm still unsure if this should be, indeed, tackled down on Recall. I will leave this open in the mean time, just to prevent this to be forgotten.
As the tittle says
EDIT:
KSP 1.11 needs an specialised version of Driftless, that would act only on Yaw, as the drift itself is fixed on it.
KSP 1.12 too.