Stability issues with small parts (watch simulation)

[Mariah A.P.]

Hi everyone,

I’m trying to simulate the functioning of a wristwatch, but I’m running into instability issues that seem related to scaling.

Since the components are from a watch, they have very small masses and inertia values, which I think may be contributing to the problem.

My geometry is defined in millimeters and masses in grams, and torque/stiffness were calculated in N·mm. To match the scale, I converted these to μN·mm, but this results in very large numerical values. When I run the simulation like this, it becomes unstable and the parts behave unrealistically, almost as if they are floating or flying apart.

If I reduce the force and torque values significantly, the simulation runs smoothly, but then the units are no longer physically consistent, which defeats the purpose.

I’m using the NSC system with a timestep of 0.005 s, and I’m wondering whether this could be a solver-related issue, or if I’m approaching the unit scaling incorrectly.

Has anyone dealt with something similar or can suggest best practices for handling very small-scale mechanical systems like this?

Thanks in advance!