floating particles
#1
Hi, I'm a beginner and spend a lot of time studying this (amazing) plugin. even so, I don't seem to fully know how to plan the build up for what I need to do - 
I have a closed container, in an irregular shape (not a box or sphere or nothing like that). in it there are a fixed amount of particles that act like balloons - float in space, but they are rigid - they don't have cloth quality. at the end they should look like spheres that float in water. they collide with themselves and with the inside of the container. after a while they are pushed into a side pipe. i use gravity so they float, but they stick to the upper side of the container. I also don't know how to make them enter that pipe. the effect I want is this - like a closed, transparent donut. filled with liquid and spheres (you don't really see its a liquid, it's so just for the spheres to look like they are floating) that when you push it, the spheres move away from the point of pressure. 
i'm not sure i'm clear. I'm attaching a diagram. two stages - 1: floating, 2: pushed through pipe. 
thank you!!
  Reply
#2
I don't see diagram you mentioned.

Sounds like a pretty easy job for PhysX.
Add PhysX shape to your particle balloons so they collide with each other, add PhysX Collision, and choose that container (switch from covex to mesh mode).
Select Tyflow Icon and disable gravity. Add force operator, and increase turbulence with small noise (like 0,03) so they slowly float. add Slow operator, so it compensate force increasing speed.

Now, for the pipe entering... I often use spherical inverted wind.
So, just create Tywind (spherical), give it a negative value, so it attracts the balloons, add another Force operator to Tyflow and pick that Tywind.
Now you can animate Tywind (both position in space, and strength), and very easily force your particle to go te way you want them.
  Reply
#3
thank you! I'm trying it right now, what it the difference between collision and physX Collision?
  Reply
#4
There's 2 solvers in Tyflow.
Particle Physics, and PhysX.
You use PhysX Collision when dealing with PhysX solver.
You use regular Collision for pretty much everything else.
  Reply
#5
(04-30-2021, 04:39 PM)d4rk3lf Wrote: There's 2 solvers in Tyflow.
Particle Physics, and PhysX.
You use PhysX Collision when dealing with PhysX solver.
You use regular Collision for pretty much everything else.

so because i use the PhysX shape, I need the PhysX collision?
  Reply
#6
I'm still in the first part of the simulation and having a few problems. I used the Birth Voxels in order to fill the container. added shape, PhysX Shape, PhysX Collision (mesh) and then add the force (turbulence set to very low) and slow. it is a closed container but the particles "run away" from it. I'm left with little particles, when I need them all, because I will have to push them in a tunnel. I also don't get the floating effect. The container is very narrow and at parts should only contain 1 particle in depth, and it is also a in a very small size. What am I doing wrong?
  Reply
#7
Can you share some screenshot so we better understand the scene?
Or anything that will help visually describing?
  Reply
#8
here are two screenshots. 1 is the particles in the container and one when the run away.
I tried turbulence but i couldn't get to find the right setting for it to look good. it looks like a wave and disappears.

oh. I think I have made progress! look: (first attachment). I used speed. i want to have like a humming bees effect. make the particles move around themselves. I need to make sure they wont enter one another but have some bouncing look.
The next stage needs to simulate a push from the back side and the particles moving into that tunnel. (second attachment).
their surrounding is liquid, but it might as well be solid (unless it could make my life easier...)


Attached Files Thumbnail(s)
           
  Reply
#9
Huh..
They are way more dense then I expected... not sure how you expect them to flow around, where they are basically touching between themselves.

It's hard to give precise advice (at least, for me), on this, unless I try scene myself.
But I'd try to push them towards that small entrance, and there I'd test them with some surface test operator, to bring them in new event, and in the new event, I'd make path follow, and they would follow a line (I previously created), in the tube.
  Reply
#10
I'm trying it right now, and of course it does not go well. I'll keep trying before asking for a save...
I'll reduce the amount of particles to see more float effect. thanks for the remark about that.
I work in a small scale. does it effect my workflow? should i enlarge the objects?
and
do I have to copy and paste the shape, physx shape and physx collision to each event?
  Reply
#11
PhysX Shape automatically transfers in next event (same is for the shape (geometry)), so no need for that.
PhysX Collision doesn't transfer, so you need a copy of it, in the next event.

If you want, save me a file (or at least that main collision object), with line path), and I can try tomorrow morning, if clients don't sit on my back at that time. Smile
Just save it for Max 2018, because it's the version I am using.
  Reply
#12
Wow, Thanks!!!!

I know those clients.. I hope the work is interesting though, at least Smile
What I did so far (and you'll see in the file if you'll get to it) is create a second event with find target at frame 50, and I linked the target to a dummy with path constraints.
Right now the particles get stuck and won't get into the tunnel. I'll try to change the friction.
When I tried to use the Push operator it went crazy so I'm looking the other way wishing it will go away. 

Thank you again so so much. you are gold!


Attached Files
.rar   Forum_01.rar (Size: 2 MB / Downloads: 190)
  Reply
#13
Ok, here you go.
Find target won't work here (because, basically, all your particles would search one point), but as I said previously, path follow is the answer.
Notice that I increased (under Tyflow icon)  time steps to 1/2, and bumped out these 3 simulation settings under PhysX.
This is because, particles was way to fast to calculate, so they were exiting the collision mesh. If you make them even faster, and they start to go out, decrease time step to 1/4 or 1/8.

You can control their speed inside "pipe" with Path follow velocity parameter.
And don't forget that you can control overall speed of whole Tyflow system if you use retimer option in Tyflow icon.

Feel free to ask anything unclear. Smile


Attached Files
.rar   Forum_02.rar (Size: 2.01 MB / Downloads: 217)
  Reply
#14
Thank you!!!!!
I saw you changed the collision object (the one with the shell), and if the shell is kept, the particles get stack on entrance point. they work beautifully with wider tunnel. is there away around that?
And
in order to get the particle float in the main surface for awhile before they get tested and go inside the pipe, should I create another event, first with test for frames (as time) and than to the sphere trigger and entering the tunnel?
  Reply
#15
(05-02-2021, 07:30 AM)mic3mic4 Wrote: I saw you changed the collision object (the one with the shell), and if the shell is kept, the particles get stack on entrance point. they work beautifully with wider tunnel. is there away around that?
..........
in order to get the particle float in the main surface for awhile before they get tested and go inside the pipe, should I create another event, first with test for frames (as time) and than to the sphere trigger and entering the tunnel?

1) Yeah, you can change it back to shell. You only need to worry that when they are born, they are not intersecting with collision object. So, in this case, I guess you make smaller their birth (voxel) object... the object you used for birthing voxels... 
Just try to born them, so they don't intersect with collision object. 
Now, if the entrance area is too small for particles to be sucked in, they might stuck, like in real life... Smile that's why I used push modifier.. or you can make particles smaller...  

2) If you want them to float a little bit, you can just go to Surface test operator, and up on the option, you will see timing rollout. Open that, and change it from "continuos" to, let's say frame... 
So, if you want them to start from frame 200, type 200 for start (so that operator start working at 200 frame), and whatever number for end (1000 or more), and you can even variate them, so if you type 20 for variation, they will start from 180 to 220 randomly.  
You can do that timing for all other operators as well.. probably for that inverted wind (that is attracting them to the entrance), and that speed operator with icon that push them on left a little bit, but don't do it for force with noise, so they bounce all the time.
  Reply
#16
Spent the better part of the day changing the shape of the big chamber and the tunnel. I'm using your events and it look so much better! I can't thank you enough. I had the problems you predicted and followed what you said. I did have problem again of the particles getting stuck - just like in real life, so I changed the friction to 0 and so far looks good. I'm not rendering just yet, still many things to do before I set things up. I believe more questions will come up soon enough..
How are you with pheonixFD? Wink
  Reply
#17
I've used Phoenix for some water sims, but for the smoke, fire, explosion stuff, most of the time I am using FumeFX.
Although, I am not huge fan of fluid sims, because I am bored to wait all the calculations, and it annoys me that when you bump up the res, you can get pretty different results.
  Reply
#18
yeah, for now I found another way to do what I need. the particles suppose to float in a liquid, but the liquid does not really function as liquid. its more like an animated solid the one end of it move along that path. (in coloration with that particles being pushed inside the tunnel. i used a slice spline)
I have changed the surface test timing to a later frame, and they do wait (beautifully) before the enter the tunnel/ but now I'm not sure what i should change so the particles inside the big chamber will wait also before the get closer to the pipe's entrance. Is it the TyWind force or something else / more?
  Reply
#19
Yeah.. do the same timing thing with both Ty Wind and speed.

I actually mentioned it before Smile
  Reply
#20
Right! I didn't read till the end. god I'm tired. sorry...
  Reply
#21
Just sent my client a sketch, we'll see what they will say I need to change and then I'll come running to you Smile...
I have a question - if I take all the objects in the scene, including everything related to the tyflow elements and rotate it (so i wouldnt have to rotate the camera around the product) - will it make everything crazy? will I need to adjust tyflow parameters?
  Reply
#22
Not sure, but it shouldn't be an issue since you disabled gravity.
Just save your file as a backup... turn off Tyflow.. grab all objects except camera... rotate them... enable Tyflow, and see what's going on.
If the rotation pivots going crazy while you rotate the scene, you can quickly make a group of all selected object... rotate it... then ungroup...
  Reply
#23
I was afraid to try that... I'll get to it later on and let you know
  Reply
#24
It doesn't seem to work well.  i need to move the camera only..
  Reply


Forum Jump: