Dspline problem

[Billabong]

I have a project where I have a spiral path that leads down to a glass that the fluid from the path is suppose to fill. I have been messing with this Dspline for two days and have achieved nothing but complete frustration.

Ill get it coming down vertically, but then once it starts on the spiral, the fluid will either stop completely or it will pass through, but it will eventually break apart. I have taken the Axial CP from .005 to 25 and nothing seems to work. my CP radius has gone from 2 to 4. My main axial strength for the Dspline was at 1, but changed it to two, with no luck. I could really use some advice. Should I stick with the Dspline on this or would this be a job for Crowdflow(which is another very confusing Daemon), I watched the only tutorial there is on Crowdflow(didnt help a lot).

Ive uploaded a video so you can see what happening as well as a screengrab on the path. I added a link the higher res screengrab.

Thank you guys for any help you can give me.

www.billa1.com/downloads/dspline_path.jpg

http://www.billa1.com/downloads/dspline.mp4

[tsn]

When the fluid stops the radial forces are not strong enough to move the particles through the spline. The axial forces create an acceleration towards the spline and this normally makes the fluid thinner. These breaking effects you can see, often occur when the distance between the control points is very large. More points will help you to "transport" the fluid. If "Radius" is very large then it might happen that a control point affects the fluid in unwanted areas, because its influence intersects the spline/fluid.

A very good method for avoiding these breakups is to add a "Sheeter" daemon and fill the gaps. This way you're able to get a closed stream of fluid. I also recommend keeping the vortex force rather low, because they introduce a twist that might break the fluid as well. One problem with Dspline is that - despite of all actions - some parts of the fluid stop moving after a while.

So, maybe it's better for your project to use a "Spline" emitter instead, because here you'll have a constant particle stream with the "Edge" mode. Again, a "Sheeter" daemon can help to avoid breakups. If you want to stop particle emission at a certain time, add an expression to the "Speed" parameter, e.g. stop at frame 50: if(f>49,0.0,2.0) - the second value (2.0) is the original speed. Substitute it if necessary.

I hope this will help you to get what you're looking for

Cheers,
Thomas

[Billabong]

tsn,

I cant thank you enough for that detailed explanation. The Spline emitter never occurred to me. I am going to try and set that up tonight.

One question. Why would I need an expression to stop the emitter, could I not just keyframe the speed?

Once again thank yo, once I get something going on it, ill post back and let you know how it went.

-Brandon

[tsn]

You're welcome Brandon. I'm happy that you find my suggestions useful.

With regards the expression: of course it's possible to keyframe speed. this was just meant as an extra tip for changing parameters quickly.

If you need further help just drop us a line here.

[tsn]

Maybe you also want to check out these links:

https://vimeo.com/114870384
https://vimeo.com/107295411
https://www.youtube.com/watch?v=VLkMxeIwU-E

[Billabong]

Ive watched the last one, but haven't seen the first two, so i will definitely check them out.

Also what is your thoughts on the Wetwork plugin "Cursum"? We are still using 2013, so this would actually work if I can convince my boss to purchase it.

http://www.informatica-3dart.it/?p=398

Thank you
-Brandon

[tsn]

I've been using the Cursum Liners plugin some time ago and my experiences are ambiguous. The biggest pro is definitely that it creates very nice results. Its biggest con: it takes some time to get started and achieve a convincing simulation, because it's not so easy to create an initial setup. You have to define the fluid's initial shape by adding and positioning a large amount helper objects. In the end I've been using the spline emitter again...

[Billabong]

So after watching the other videos you posted, it seems I was overthinking this process and messing around with the individual Cp's too much. So I reset them all to their default value and increased the main radial force and decreased the main axial force and i got it working correctly. I then adjusted the individual cp's to get it perfected.

The only thing I don't like is how the fluid will twist at some points even with vortex turned completely off.

Thanks so much for the help tsn.

-Brandon

[Billabong]

well i fix one problem ,then i get another. At the end of the dspline, the fluid insists on going back through the path instead of going in the glass. Any ideas? I have changed the axial to high and low numbers on that CP but it doesnt help

[tsn]

Good to see that you've been able to move one step forward.

The behaviour you get at the last CP is normal, because this point attracts all particles. The other points do that as well, but here the particles are moved through to next point. What you can do to is to continue the spline until it reaches the glass's bottom.

Then we have to find a way to release these "terminating" particles to finally fill the glass. One method is to create a second emitter at the glass's bottom. It starts to create particles when the spline particles hit the glass and will fill the object. These spline particles, on the other hand, are deleted with a k Volume daemon (I hope you understand what I mean ).

Another way is to use a bounded filter daemon. As soon as the particles enter the filter's boundaries they're shifted to a container (this container has to be unlinked from the DSpline daemon!) and then they're no longer attracted and can fill the glass.

So, you're almost there and it only takes a few last steps to get this running.

The twist is difficult to remove, because it's a matter of how the forces are shaped. You could try counteracting "@ CP vortex" values: when the twist is clockwise, enter a counter-clockwise vortex value at the twist's closest control point. Maybe this will help a little...