Author Topic: dices floating in space : weird behavior (Read 6256 times)

rosiernain · « **on:** May 18, 2013, 02:47:42 AM »

I'm new to universe sandbox ; I open the simulation of dices floating in space ; I don't understand : the dices come close ok, they collide , and then ... they stay close , and still , which is not what you can expect from basic mechanics.. (I expect they should come back to the initial position with the velocity coming smoothly to 0 , and then come towards eachother again and so on )
is there a bug with this specific simulation ??
how can I trust the simulations of universe sandbox ??

vh · « **Reply #1 on:** May 18, 2013, 07:56:11 AM »

hit shift-I a few times until the elasticity is 1. by default i think it's like 0.3 or something because planets usually aren't elastic i guess

also i suggest using the rk4 method because it conserves energy better iirc

--
on another note if the dice collide at angles it's interesting to see how the movement gets converted into angular momentum and the dice bounce lower if they're spinning

rosiernain · « **Reply #2 on:** May 18, 2013, 10:30:21 AM »

hi thank you for the reply ; I understand .
but how do you do that : to change the elasticity and to use rk4 method ; I don't find the buttons...
thank for your help

vh · « **Reply #3 on:** May 18, 2013, 10:46:53 AM »

you can change elasticity by pressing shift and I at the same time to cycle through the available settings

rk4 is on the left panel right under the simulation options and next to accuracy mode

vh · « **Reply #4 on:** May 18, 2013, 10:59:33 AM »

upon further experimentation it seems that even that doesn't completely solve the problem because the dice gain energy due to inaccuracies and fly apart.

in addition there's a bug where if you place two objects and bounce them, they'll suddenly move slower or somethin

swift_powaa · « **Reply #5 on:** May 18, 2013, 11:05:01 AM »

Hi Guys, I am new to this-kinda. How do i download textures to apply to planets? Thanks!

rosiernain · « **Reply #6 on:** May 18, 2013, 11:18:43 AM »

yes , exactly, it's still very weird : the energy and the angular momentum are not conserved at all..

rosiernain · « **Reply #7 on:** May 18, 2013, 12:19:41 PM »

this bug is strange to me because the collision seems to happend in a nice way (the velocities and angular momentum look plausible right after the collision ) and then things go wrong : suddenly the dices slow down very quick and stop rotate.
a very severe bug for one of the standart simulation..
hum...

rosiernain · « **Reply #8 on:** May 18, 2013, 12:23:08 PM »

what to do ? who is in charge of this software ?

vh · « **Reply #9 on:** May 18, 2013, 01:14:26 PM »

dan dixon

rosiernain · « **Reply #10 on:** May 18, 2013, 01:33:00 PM »

ok , thanks; I'll probably write to him.

I try another simulation : "dices swarm" , and I realise that the simulation is totally unrealistic regarding to gravitation : if you put dices in space , without initial velocity they will never form a swarm ..

So these are not bugs , but just a totally fancy program.

And I'm very puzzled :
I see on wikipedia that dan dixon worked years on this project and was assisted by an astronomer Eric Hilton.

what is this joke

tuto99 · « **Reply #11 on:** May 18, 2013, 01:40:27 PM »

According to the law of universal gravitation, anything with mass attracts each other. So if two dice were put in space with 0 velocity, they would eventually touch each other. And try to not be blunt with your words.

rosiernain · « **Reply #12 on:** May 18, 2013, 02:10:56 PM »

ok . I apologize for the acrimony.

I was just chocked to see those dices behaving absolutly against the law of mechanics unless
you assume that they are magnetized or something like that.

rosiernain · « **Reply #13 on:** May 18, 2013, 02:37:59 PM »

I made more experiments with the dices floating in space , changing the elasticity .

with elasticity 1 or 0.9 , after boucing the dices come back at a distance much greater than the one they started and so the potential energy (at the moment where velocity is zero) has increased , wich is fancy.

with elasticity less or equal than 0.8 they come back at a smaller distance than the one they started with (wich is coherent ) but then they don't move anymore , as if the gravitational force has vanished.

tuto99 · « **Reply #14 on:** May 18, 2013, 06:44:12 PM »

I am guessing you are using the Dice Swarm simulation.
It seems that with the high elasticity the dice seem to really bounce off of each other, and once they coalesce a few go wack and fly out of the swarm. That is probably because with so many dice colliding each other there is so much energy with that level of elasticity.
But other than that, I don't see what's wrong with the mechanics of the sim under 0.8 elasticity.

vh · « **Reply #15 on:** May 18, 2013, 08:18:27 PM »

^
i've noticed that they'll suddenly slow down for no reason

the collisions are probably not implemented perfectly since universe sandbox is primarily a gravity simulator; energy is gained or lost from collisions because it's not the most accurate calculation i guess

it probably won't be fixed until version 3 comes out

rosiernain · « **Reply #16 on:** May 19, 2013, 03:58:44 AM »

this is a reply to tuto99:
what is wrong with the simulation (dices floating in space , elasticity=0.8 modified by using "shift-I") is that the dices stop apart from from each other after the first collision , and they stay there without moving anymore (at least it is what happends on my computer ) , as if suddendly there is no more gravitational attraction between them.
They should collide again and again with shorter moves each time (because of the elasticity<1 ) .

I'm surprised that you don't see anything wrong with that.

tuto99 · « **Reply #17 on:** May 20, 2013, 03:14:32 PM »

Quote from: rosiernain on May 19, 2013, 03:58:44 AM

this is a reply to tuto99:
what is wrong with the simulation (dices floating in space , elasticity=0.8 modified by using "shift-I") is that the dices stop apart from from each other after the first collision , and they stay there without moving anymore (at least it is what happends on my computer ) , as if suddendly there is no more gravitational attraction between them.
They should collide again and again with shorter moves each time (because of the elasticity<1 ) .

I'm surprised that you don't see anything wrong with that.

I realized something when I played a simulation of two dice colliding. As soon as they hit, they stuck together. I think that's more of a glitch than screwy physics. Yeah, Dan never finished Ubox 2, so I wouldn't expect everything to be completed. Still a good program though.

vh · « **Reply #18 on:** May 20, 2013, 03:32:30 PM »

they only stick when the elasticity is below 0.5 or something

tuto99 · « **Reply #19 on:** May 20, 2013, 03:34:52 PM »

Quote from: vh on May 20, 2013, 03:32:30 PM

they only stick when the elasticity is below 0.5 or something

It's a glitch. They actually start to bounce off and then they just stay stuck there.

Dan Dixon · « **Reply #20 on:** May 22, 2013, 05:25:08 PM »

Yeah... if you increase the time step a bit... the dice will come together properly. Not really sure why that's happening and never got around to fixing it.