Sandbox and Research
The switch in the title bar changes how the scene is run and drawn. A scene remembers its mode. The mode is a setup, not a set of features: Pro switches between the two whenever it likes; Free runs in Sandbox (where the switch is a button that explains Pro). The time bar, graphs, tables and exports work in both modes (graphs, tables and exports are Pro).
| Sandbox | Research | |
|---|---|---|
| Time step | 1/60 s | 1/240 s, fixed |
| Drawing | colourful, for building | textbook style: thin lines, hatched supports, labels in italics |
| Dragging | drag and fling bodies (right-drag flings) | drag to place; the run is recorded |
| Resting bodies | fall asleep to save time | always simulated |
| Time bar | yes | yes |
| Graphs, tables, sweeps, exports (Pro) | yes | yes, in textbook style |
What Research guarantees
In Research the engine works in SI units and advances with a fixed step, so a run can be repeated exactly and compared with theory. The accuracy claim covers circles, boxes and polygons, distance, spring, pin, slider and weld joints, supports at any angle, static, kinetic and rolling friction, ideal cables (with or without mass along them), gears, belts and axle motors. Each of these has a test against an exact result: see Validation.
Two things are approximations and say so:
- The Rope tool makes a chain of small links: it sags, swings and drapes, but it is a model, not an exact cable. Use the Cable for exact tensions.
- Air resistance from a body's shape uses a drag coefficient you can change; the linear and quadratic models (F = −b v, F = −k v²) are exact.
Changing the step
Properties of the scene (nothing selected) → Simulation shows the time step, substeps and solver iterations. The defaults are the validated ones; more substeps make stiff mechanisms more accurate at the cost of speed.