fix: add fixed timestep accumulator to prevent energy drift in BouncingBall

[iamjaysingh]

Fixes #226

Root Cause

The BouncingBall simulation was using a variable timestep (real wall-clock
delta time) for physics integration. When frame rate varies — due to tab
switching, browser throttling, or display refresh inconsistencies — the
variable dt causes energy drift in the Euler integration, making the ball
lose or gain height on each bounce even at damping = 1.

Fix

Added a fixed timestep accumulator pattern to BouncingBall.jsx:

• Physics always steps at a consistent FIXED_DT = 1/120s
• Real frame time is accumulated and consumed in fixed increments
• Accumulator is capped at 0.1s to prevent spiral of death on tab refocus

Testing

• Damping = 1 → ball bounces at consistent height indefinitely ✅
• Damping = 0.8 → ball gradually loses height correctly ✅
• Damping = 0 → ball stops on first bounce ✅

[vercel]

@iamjaysingh is attempting to deploy a commit to the PhysicsHub's projects Team on Vercel.

A member of the Team first needs to authorize it.

[mattqdev]

Thank you for this fix, this applies also on the other simulations with the same issue?

[iamjaysingh]

Thank you for this fix, this applies also on the other simulations with the same issue?

Hi @mattqdev! Yes, the same fix can be applied to the other simulations
with the same issue (SimplePendulum, BallGravity).

The fix is the same pattern — adding a fixed timestep accumulator to each
simulation's draw loop. I can extend this PR to cover those simulations too
if you'd like, or submit separate PRs for each one. Whatever works best for
your review process!

[mattqdev]

I'd like to have one PR for fixing everything, do you think you can fix this issue in a centralized way (i mean, not changing every simulation file)?

[iamjaysingh]

I'd like to have one PR for fixing everything, do you think you can fix this issue in a centralized way (i mean, not changing every simulation file)?

Yes absolutely! I can implement the fixed timestep accumulator centrally
in Time.js so all simulations benefit automatically without touching
individual simulation files. I'll update the PR shortly!

[iamjaysingh]

I'd like to have one PR for fixing everything, do you think you can fix this issue in a centralized way (i mean, not changing every simulation file)?

Hi @mattqdev! I've updated the fix to be centralized in Time.js instead
of individual simulation files.

The fixed timestep accumulator (FIXED_DT = 1/120s) is now built directly
into computeDelta() — so every simulation that uses computeDelta()
automatically gets the fix without any changes to individual files.

Tested on all three affected simulations:

• BouncingBall (Damping = 1) → consistent bounce height ✅
• BallGravity (Damping = 1) → consistent bounce height ✅
• SimplePendulum (Damping = 0) → consistent swing amplitude ✅

Let me know if you'd like any changes!

[mattqdev]

Thank you! Is also Parabolic Motion fixed?

[iamjaysingh]

Thank you! Is also Parabolic Motion fixed?

Yes! ParabolicMotion also uses computeDelta() so it automatically
benefits from the centralized fix. Tested and working correctly.

[mattqdev]

Alright, perfect! Thank you!

[physicshub]

🎉 This PR is included in version 3.29.0 🎉

The release is available on GitHub release

Your semantic-release bot 📦🚀