detail on move/mesh changes in 3.7.0?

Submitted by pfalkingham on Fri, 07/14/2017 - 19:53

Hi, where can I find more info on the following listed change for 3.7:

"Arno Mayrhofer (DCS ) modified the fix move/mesh command that now has reference frames (for the center of rotation) that depend on parent fixes"

Thanks.

[edit]
The reason I asked, is that my stacked rotations are now all being combined into a single rotation. That is, the reference frames all stay the same, when I've been using a rotating reference frame to essentially run my x y and z rotations separately.

Forums:

arnom's picture

arnom | Thu, 07/20/2017 - 08:13

Hi,

there was a bug in how linear and rotational movement were interacting with each other which required the rewrite.

Basically what I mean with reference frame only refers to the center of rotation. So the center of rotation of the second rotate should be dependent on the first rotation (NOT the axis). Maybe you can post a minimal working example to highlight the issue you are having.

Best,
Arno

DCS team member & LIGGGHTS(R) core developer

pfalkingham | Fri, 07/21/2017 - 15:29

Thanks. Here's some example files:

Liggghts files (STLs, plus I use a setup then run file, where the setup fills the volume with particles, so I can apply different motions without needing to regenerate the particles).

https://1drv.ms/u/s!Am5GkbZS_98XrPtIWuAOKujJuyOZaw

Here's a quick animation of what the motion should look like:

https://1drv.ms/u/s!Am5GkbZS_98XrPtIWuAOKujJuyOZaw

it's a 90 degrees rotation about world Y, then that is reversed while the cylinder rotates about its long axis by 90 degrees. finally it rotates 90 degrees about world Z. I've been fudging how I give that information to liggghts supplying 3 axis-angle rotations parented/childed under each other. This worked fine upto and include 3.6.0, but with 3.7.0 the parenting doesn't seem to take effect, and during the second part the long axis rotation frame doesn't seem to be parented to the y-axis rotation. The result is that the cylinder does not return to horizontal properly.

I know I need to update my workflow to try and get a single axis-angle rotation rather than parenting 3 but that's previously been difficult (possible now perhaps) - that may be all that's needed to fix it.

By the sounds of it, the reason it's been working at all is I've been exploiting a bug all this time?