Hi Alex,
Thanks for your file: The tear should be the lower value, and the reminder 100%. But the effect of missing polygons is untouched—I can see it. There was obviously a change, since the simulation itself is different in R17 to R18. So I checked even R12 as reference.
I’m not certain what other to say than, please contact the support. I really don’t know. If I dial the numbers down, the gap is gone, but it doesn’t break/tear anymore. To ask for dialing this in, animated, is perhaps too much to ask for or too “whatever”.
The problem is normally that the Weight values become multiplied with the value for the tear, so if the value is zero, then the “springs” inside the cloth are instantly broken, hence the tear, why the polygons dissapear is not clear to me and not in my power to fix. Support.
Cloth is a generator and this is calculated after generators normally, so there is a priority jump, and that might be the key that we can dial it to stay, then no tear, or no gap and tear. Not your concern nor your challenge, jsut to share a little bit the background.
Alternative—pre-version:
I have here something else, and please explore it, as my next file will build up on it.
It does not use a Weight map for the Tear, but for the Belt. The progress of the Belt-Weight is calculated with a falloff in XPresso. How far that goes with waving Fabric, I can’t tell. This was my Sunday, exploring this and I think that should be enough for this post. Perhaps you don’t like it, so I would like to see if it makes sense before I go further. Note that I just used all points to belt it to the Platonic object. Then I lower the belt with the weight map and the falloff. See the clip or the file.
One minute Clip (can be downloaded for better viewing experience):
https://www.amazon.com/clouddrive/share/3Hi7aB7pK4VFbggUYHraI0qNhGS8SaNdDJoj7de0Ojm?ref_=cd_ph_share_link_copy
File:
https://www.amazon.com/clouddrive/share/V3O2a2roDBKR5mYgnbcqjeB1ROIz8adkXR6gAvGW8tW?ref_=cd_ph_share_link_copy
Finally, the workaround:
Here I have prepared a set up like yours, but pre-fabricated the tear and used a belt to stitch it back, until the Weight map tells the belt differently.
The Falloff (XPresso) is set to source, and here I used the Sphere (the one that is set to Collide) and the points of the Sphere are the Falloff sources, each point!
So, when the cloth comes closer, the Falloff starts to change the Weight map and the belt function weakens.
File is here:
https://www.amazon.com/clouddrive/share/3WjHUjUDTZdRuJWesvLnfXKWfqJEeAu59YlJnBea2C?ref_=cd_ph_share_link_copy
My best wishes with the Project and the support.