I’m working on a system where the player builds a complex assembly by attaching multiple parts together. I’m using physics constraints to connect actors, but I’ve noticed instability in the physics simulation. The goal isn’t high-precision physics, but rather to keep the assembly together as a cohesive unit and to handle basic collisions with terrain and other actors.
So far, I’ve mitigated some instability by disabling collision responses between attached actors using MoveIgnoreActor, and I have the following settings for the physics constraints:
- Collisions between the constrained actors are disabled.
- Projections are enabled.
- All linear and angular movements are locked at 0.
The benefit of physics constraints is their ease of use, particularly for allowing the player to deconstruct the assembly. However, I’m wondering if I’m over-relying on physics constraints, as maintaining stable physics with around 200 actors connected using constraints may be unrealistic?
Would the physics solver be able to handle this many actors with the above settings? Or am I better off switching to an approach where a new actor dynamically adds/removes static mesh components when building/deconstructing the assembly (even though managing the physics and state changes for this approach is a pain)?