Hey, Fran, thanks for making me research this!
It's seems, I'd say, counter-intuitive the way the deflectors work.
But …
Based on the tests I just did, I'd say that, as far as "working as expected", if there were no 3D geometric options for the deflector tag, I'd say they
do kind of work as expected, that is,
A "half space" deflector (which perhaps should be called something like "limited planar") only deflects those emitted objects that "make contact" with its (by default, x and z) 2D geometry;
The "unlimited" deflector ("unlimited planar" probably a better label) is unlimited in its (by default, z) deflection. Rotating either of these (in x or z) will change the deflection as one would expect light (reflections) bouncing off a mirror, or billiard balls off the rail (no English!) - that is, angle of incidence = angle of reflection.
Now, we can get into the weirdness of the 3D geometric deflectors.
Let's just take the cone, since you singled that one out.
I found that the cone's "middle" plane acts much like the "half space" deflector, in that when emitted objects hit its
circular plane, they deflect (
or don't, seemingly based on something I'll get to in a sec), and if they do not "hit" that plane, then they fall right on through (I'm basing this on all tests done with objects emitted downward and deflectors by default pointing upward).
Fine.
So it begs the question, why is it shaped like a cone, then? Well, for one thing, you get that circular plane instead of the rectangular in the "half space" …
So I did a test: I scaled the cone deflector in the y to zero so that it was completely flat - and guess what? No particles deflect; they all pass through. If I scale it up a little, I don't see deflections until I get to around .5, regardless of how many particles are emitted. Interesting. And as I increase the y scale of the deflector, it deflects more and more.
So there does seem to be some logic to it, and I can see how these can be used in certain situations.
However, if I'm doing something that actually needs objects to bounce off other objects, I'm going to use 3D geometry and not use emitters/deflector - I'm going to use dynamic/rigid rigid body tags and Dynamics.
I hope this makes sense - I definitely have a better understanding of using emitters and deflectors since you posed the question, so thanks again. I enjoy learning stuff.
-CG
I'm currently struggling with a related issue.
I'm fine with using deflectors to repel my emitted particles, but it appears only the "half space" deflector works properly. All the others make no sense to me - perhaps anyone has some input regarding that?
For instance a cone deflector (inverted or not) will not bounce particles off its cone shape (which is what I need for my current project). Instead they enter the cone and are deflected if they hit the infinitely small center of it (meaning most particles just fall through), and even then they don't deflect according to the geometry of the cone, but seemingly randomly.
Am I misunderstanding how these deflector types are intended to work, or are they buggy?
Right now my only work around to have particles bounce of a spaceship, I need to either produce them without the particle emitter, or use half space deflectors as polygons to remodel the shape of the ship's nose...
Thanks for any insights you can provide!