Forums Technical Discussion Pop Drag Spin explodes the simulation

Pop Drag Spin explodes the simulation

   3288   5   0
User Avatar
paulprk
Member
7 posts
Joined: Feb. 2018
Offline
 
Hey everyone!

I have an RBD simulation with packed primitives. I'd like to use some drag force on the spin to reduce rotation of some of the smaller elements. If I understand correctly I'm supposed to use the POP version of the drag spin force. I tried plugging into different inputs of the rbd solver, but it does something really weird. The whole simulation just explodes or pieces disappear.
Am I just plugging it into the wrong place? How is it supposed to be used? Or is it just a bug?

Thanks for any help!
 Report
User Avatar
tamte
Member
8532 posts
Joined: July 2007
Online
 
I have the same experience, feels like a big to me
as an alternative you can just multiply down angular velocity like

v@w *= .98;

Or something like that, the same for v if drag explodes for you as well
Tomas Slancik
FX Supervisor
Method Studios, NY
 Report
User Avatar
jsmack
Member
7741 posts
Joined: Sept. 2011
Offline
 
tamte
I have the same experience, feels like a big to me
as an alternative you can just multiply down angular velocity like

v@w *= .98;

Or something like that, the same for v if drag explodes for you as well

What's the math for samplerate-invariant decay?
 Report
User Avatar
tamte
Member
8532 posts
Joined: July 2007
Online
 
jsmack
tamte
I have the same experience, feels like a big to me
as an alternative you can just multiply down angular velocity like

v@w *= .98;

Or something like that, the same for v if drag explodes for you as well

What's the math for samplerate-invariant decay?

v@w *= pow(0.98, @TimeInc);

Where ,98 factor becomes per second not timestep as previously so you'd have to decrease it a bit, but it's all about tweaking
As if you want proper drag it'd be proportional to squared relative w or v, etc
Tomas Slancik
FX Supervisor
Method Studios, NY
 Report
User Avatar
paulprk
Member
7 posts
Joined: Feb. 2018
Offline
 
tamte
I have the same experience, feels like a big to me
as an alternative you can just multiply down angular velocity like

v@w *= .98;

Or something like that, the same for v if drag explodes for you as well

Hey! Thanks for the answer. Maybe I'm wrong, but isn't pop spin drag doing something a little more elegant? I mean doesn't it take into consideration the weight and perhaps even the dimensions of the object?

Where should I be placing this wrangle? mix rbd solver and sop solver with multisolver and make thew wrangle inside the sop solver?
Edited by paulprk - Aug. 12, 2018 08:12:35
 Report
User Avatar
tamte
Member
8532 posts
Joined: July 2007
Online
 
paulprk
Maybe I'm wrong, but isn't pop spin drag doing something a little more elegant?
well, if it worked then yes

and while yes, you can work out math from a proper angular drag equation as discussed above, you can as well start simple

just do multisolver with Bullet and Geometry Wrangle:
float d = 0.5;
v@w *= pow(d, @TimeInc);
v@v *= pow(d, @TimeInc);

and see what value works for you, or vary the numbers for linear and angular, but again, nothing is stopping you from doing more involved math using quadratic drag force if you need, or as well submit a bug, the more the better
Tomas Slancik
FX Supervisor
Method Studios, NY
 Report
  • Quick Links