Few 1.2 liquid questions

[LarsSonparsson]

So I have played with new liquids one night and it already looks really promising.

There is few things that manual doesnt explain.

1. Does the viscosity have kinda on/off switch? 0 is off and everything above
turns it on. I found that 0.01 changes the simulation totally.

2. What is the time base of the drying time when using wetting option. Frames? Seconds?

3. How could I get more stable solutions?
I'm simulating basic test scene, where I pour water to the box.
At the point when I start to get enough details, simulation gets really unstable.
Well haven't got it to simulate more than 40 frames and then it brakes.
What is the key to make it more stable? Conservation quality? SPF upper limit? Lowering the maximal step value?

I'll post some samples later when there is something to show.

[Ivaylo Katev]

1 - yes, really 0 is off and all above is on.
2 - seconds
3 - hard question. if the liquid is free flowing the stabilization is not too hard, but when is cached in a vessel the things change a lot.
there are two different types of instability - pressure driven and surface artifacts.
the pressure driven instability happens when the liquid is pushed with some force (the gravity in the most cases) and can't move toward the force. in this case the extra velocity given by the force should disappear during the conservation process, but there are some remaining movement that produces boiling-like behavior. of course, the high conservation settings can help here, however the right scene setup can be also very helpful. the basic idea is to avoid the deep waters, if you need a deep looking water, replace the areas not involved in any activity with solid geometry having the same material settings as the water. as starting point you can see the paints sample - it shows a liquid that gets calm after some initial activity. there is an important rule i found playing with liquids in equilibrium - you have to use the SPF limits to control the advection, not the step! in numbers the rule looks like this: 15*SPF*FrameRate*cellsize>=acceleration. this means - you can start from the settings of the paints sample (it satisfied the rule) and doubling the gravity, you need to double the SPF lower limit, doubling the resolution (linear) you need to double the SPF.
the second instability (the surface artifacts) has very typical appearance, it looks like stable or slow changing roughness over the surface. this instability is strongly related to the slow moving liquids, it never happens over the surface of a fountain or some other faster sim. the cure is also related to the lower SPF limit and there is also a check box "static surface correction" provided to help in such situations.

[LarsSonparsson]

Thank you!

I need to think about what you said to the 3rd one. Too much information!
Anyway, what I meant with the unstable solution is that the whole simulation explodes.
Liquid won't stay as a one mass, but it totally brakes to "particles" which doesn't seem to have any connection to each other anymore.
Hard to explain without right terminology.

Well, current simulation is going just fine, so maybe I'm starting to learn something.

[Ivaylo Katev]

Anyway, what I meant with the unstable solution is that the whole simulation explodes.

if you are using bidirectional conservation method this can be the reason.

[LarsSonparsson]

Nope, using smooth method.
But the reason for earlier result might be some other user error.
Now my current simulation is going just fine and looking good.

[Paul Oblomov]

For me it was stable when lower limit was 10 and conservation right about 40.