Wierd 1 frame particle "explosion"

[marnix@rckmn.nl]

I've got a scene which is a dishwasher with a rotating arm.
To this are 6 emitters are linked to a part that cycles aroung it's own axis and around the middle of the dishwasher.

While simulating this I noticed that at frame 45 there is 1 frame of interruption of the flow.
It's almost as if an invisible collision happens with all 6 emitters at frame 45.




While trying to analyze the problem I couldn't figure it out.
Changing the sub-steps / collision objects / liquid resolution didn't matter.

So i tried to adjust the animation curve of the SD file and imported it again.
with the adjusted animation the same thing now happens at frame 37.

It turns out that when the original driving dummy is rotated at ~61/62 degrees the small 1 frame explosion occurs.

does this sound familiar to anyone ?

i've attached the realflow file where it will explode at frame 45.
only 2 of the six emitters are on at very low resolution.

[LuisMiguel]

Hey,

Selecting the Geo_Emitter, in the frame 45 the axis Y changes from 177 to -175 degrees for the Y axis, and for that reason the particles are changing their position in that frame due to you have parented to the emitter to the Geo_Emitter geometry. You can solve this in RealFlow clicking on the SD<>Curve button for the Geo_Emitter geometry, and then applying a euler filter with these lines in a batch script. Keep in mind you may need to apply the euler filter to all axis so you would need to change the "Rotation.Y" axis for "Rotation.X" or "Rotation.Z".

Code: Select all

 curve = Geo_Emitter.getParameterCurve("Rotation.Y")
curve.eulerFilter()

luisM.

[marnix@rckmn.nl]

Hi Luis,

Thanks for the quick response! wasn't aware of this possibility.
Does this always occur when importing SD files with rotations?

Could it be that it's enough to just convert SD > curve?
As when I do the problem seems to be solved (and there is a fluent curve visible)

I've never ran scripts before:
When I have it in the batch script window, i just press run whille the curve editor is open in the matching axis?

tnx!

[LuisMiguel]

Hey,

Does this always occur when importing SD files with rotations? No really, but it depends on the number of frames you are putting for that animation. For instance, we can animate a sphere rotating 720º in the 3d package, so we add a key in the frame 0 with 0º and another one in last frame with 720º. For the 3dprogram could be fine, but when exporting the interpolation is not pretty good and it may be wrong.

Just converting to SD<>Curve won't fix the problem, and for this reason you'll need to apply the euler() function.

Running scripts is very simple. In that case it is BatchScrip, so press F10 and copy and paste the code, and Run the script from the menu Script.

luisM.

[marnix@rckmn.nl]

Hi Luis,

I fear I do need some more assistance..
Below an image of my curve after importing a SD file created in 3ds max.

This is how the curve looks after converting SD<>curve.
But this is also how the curve looks after running your script lines.
The message window says it ran and is done, and I've changed the Y in your script to Z.

should this not be a smooth curve after running the script?

thanks,
Marnix

[LuisMiguel]

Hi Marnix,

I can't remember the script right now, but it would be easy by selecting all keys and converting them to smooth curve. I think it is creating a bezier curve for each key.

Let mek now.

luisM.

[marnix@rckmn.nl]

Hi Marnix,

I can't remember the script right now, but it would be easy by selecting all keys and converting them to smooth curve. I think it is creating a bezier curve for each key.

Let mek now.

luisM.

Hi Luis,

I indeed solves it with changing the bezier curve interpolation for all keys. Had never notices this before.
Would be cool if there were an SD-to-Curve optimize button in the future for these things

thanks,
marnix