Skip to content

Particle Self Gravity

Particle Self Gravity creates mutual mass-based attraction or repulsion with a hierarchical GPU gravity solve.

Unlike the Gravity effect, which applies one constant vector, Particle Self Gravity calculates a field from the current positions and masses of the affected particles. Place Change Mass earlier in the effect order to change how strongly particles contribute to the field.

Setup

  1. Add Particle Self Gravity after the Emitters whose particles should contribute.
  2. Use Apply to to restrict the solve when only one population should participate.
  3. Start with Fast Quality while establishing scale and motion.
  4. Set Dimensions to 3D or the required world-space plane.
  5. Increase Strength gradually.
  6. Use Softening and Acceleration Limit to stabilize close encounters.
  7. Compare Balanced or High Quality only after the intended large-scale behavior is established.

Force and dimensions

  • Strength scales mutual attraction. Negative values reverse it into repulsion.
  • 3D evaluates full spatial separation.
  • 2D · XY, 2D · XZ, and 2D · YZ ignore separation on the omitted world axis when calculating the gravity field.

A 2D mode changes the interaction calculation but does not itself lock particle motion to that plane. Add Planar Constraint when particles must also remain on the plane.

Particle mass matters as source mass: a heavier particle contributes more strongly to the shared field. The resulting acceleration is not divided by the receiving particle's mass. Keep mass values and Strength within a range appropriate for the scene scale.

Close-range stability

Point-like gravity becomes extreme at very short distance. Two controls contain that behavior:

  • Softening smooths the force inside a short distance, preventing singularities and slingshot motion. Increase it when close pairs accelerate erratically. Excessive softening removes small-scale clustering.
  • Acceleration Limit caps acceleration from this effect. Zero removes the cap. Set a finite cap to contain acceleration as particles form dense clusters.

Softening changes the near-field shape; Acceleration Limit clamps the final magnitude. They are not interchangeable.

Quality

The solver groups distant particles hierarchically instead of evaluating every pair directly.

  • Fast uses a shallower hierarchy and broader far-field approximation for iteration.
  • Balanced resolves more local and distant structure at moderate cost.
  • High uses the deepest available hierarchy and stricter far-field approximation.

Quality affects field accuracy and GPU cost. It does not replace Substeps. Increase global Substeps when fast orbital or collapsing motion needs better time integration.

Lists and Nodes

In Lists, place Particle Self Gravity in the Effects stack and use Data Mapping on Strength when it should vary by particle data.

In Nodes, connect it in the particle stream and use a compatible Field on Strength. Branching can restrict the effect to a disjoint population, but all contributing and affected particles must remain in the intended stream. Rejoin populations before shared domain solvers and output.

Practical tuning

  • Collapsing cluster: use positive Strength, finite Softening, and an Acceleration Limit. Add collision or material behavior if the final cluster must resist overlap.
  • Repelling cloud: use negative Strength. The effect supplies repulsion but does not impose a boundary or maximum speed.
  • Planar orbital design: choose a 2D dimension mode and add Planar Constraint. Use Drag only when orbital energy should decay.
  • Mass hierarchy: assign different emitter masses or use Change Mass before Self Gravity.

Troubleshooting

  • Particles slingshot from close encounters: Increase Softening, lower Strength, set an Acceleration Limit, or increase Substeps.
  • Small structures disappear: Softening is too large or Quality is too low for the required detail.
  • Particles leave a selected 2D plane: Add Planar Constraint. Dimensions only changes the gravity calculation.
  • Motion is unexpectedly asymmetric: Check particle Mass and the effect's Apply to filter.
  • The solve is expensive: Use Fast Quality while iterating and restrict the affected population where the artistic setup allows it.