keyed particle system, unexpected behaviour for direction of particles?

Support Basics & Interface
1

Hello, I’ve been playing with keyed particles this weekend. However, the particles do not respond the way I expect them to respond, and I am unable to figure out why they do as they do.

Situation (see attach): two planes, placed on top of each other. Both planes emit 10 particles.
Plane.2 on top: Newtonian physics particle system; Emitter; initial velocity: 10 (by Normal); no global effects; no rotations.
Plane.1 on bottom: Keyed PSys (to Plane.2); Emitter; initial velocity: 0 (for all); “First” keyed is enabled.

What happens: the Plane.1-keyed particles do not move directly to their Plane.2-particle partners, but for the first 15 frames move to the opposite direction (the -/- Normal direction). They slow down, then, go upwards to their particle partner.

This is influenced by the Plane.2 PSys: with higher initial velocity set for this PSys, the harder the keyed PSys ‘fall’ down before going after their particle partner. Also, it is irrespective of ‘Newtonian’ physics, same behaviour with ‘None’ or ‘boids’ physics.

I expected the keyed particles to match the Plane.2 particles behaviour in all respects, thus without the ‘move to opposite direction first’. Any clues on what goes on/what I fail to notice?

Attachments

unclear_direction.blend (163 KB)

2

OK some follow-up thoughts on my question. Perhaps it’s helpful to someone…or, it is wrong alltogether, then do tell me :slight_smile:

As far as I can tell, the ‘unexpected’ behavior is not ‘wrong’. A path is calculated between the two particle systems using an integrator; a mathematical function that calculates the path that keyed particle #1 has to follow to get to its target particle #2. Of course, both particles move dynamically, and in 3D space, and have different initial speed/rotation/mass etc.

Apparently, in the above situation, the integrator function finds the best fit for the keyed particle system to first go down, then up. This, explains that altering the Normal speed to 10 gives more effect: it is part of the formula to calculate the final positions of the path. Altering positions or rotation of Plane.1/2 etc.: also influence the input values to the integrator calculations.

So, I’ll stop looking into this for now, and be pragmatic: just tweak and tune until I get the desired results of my particle system, without figuring out details…