I'm getting this strange behaviour affecting the dynamic attachment of a rod to a non-kinematic rigidbody, where the rod is generally quite stiff with 0 or low bend compliance but at the attachment point it sags or deflects away from the attached rigidbody in a very unnatural way, even with constraint orientation on. I tried adding another control point close to the end of the rod as mentioned in the attachments page of the manual, but this doesnt help and gives a very strange shape, given the rest of the rod is stiff. When the attached object is aset to non kinematic this doesnt happen. When I reduce the mass scale of the rod, the issue reduces, but this negitively affects my simulation, I need the rod to have some mass in relation to the end attachments. Do you know what causes this issue and how to fix it?
Kind regards,
Matt
p.s. I have added some images that show the situation, but cant show too much more on a public forum. If you need more context or information I can PM you further images or try to recreate in a new project. Hopefully these show enough.
(01-11-2025, 04:23 AM)matty337s Wrote: When the attached object is aset to non kinematic this doesnt happen. When I reduce the mass scale of the rod, the issue reduces, but this negitively affects my simulation,
This sounds like the rod is colliding against the object it is attached to. An rod cannot simultaneously be inside and outside some object so, as a result, the rod can't decide whether to keep intersecting the object (what the attachment "wants") or to project itself outside (what collisions "want") so it takes on the shape that minimizes error for both conditions, which is usually undesirable.
(01-11-2025, 04:23 AM)matty337s Wrote: I tried adding another control point close to the end of the rod as mentioned in the attachments page of the manual,
If you don't set the control points' collision filters properly this won't help at all. The solution (using collision filters to disable collisions between the attached end of the rod and the object) is also described in the manual, see "attachments inside colliders": https://obi.virtualmethodstudio.com/manu...ments.html
I'm getting this strange behaviour affecting the dynamic attachment of a rod to a non-kinematic rigidbody, where the rod is generally quite stiff with 0 or low bend compliance but at the attachment point it sags or deflects away from the attached rigidbody in a very unnatural way, even with constraint orientation on. I tried adding another control point close to the end of the rod as mentioned in the attachments page of the manual, but this doesnt help and gives a very strange shape, given the rest of the rod is stiff. When the attached object is aset to non kinematic this doesnt happen. When I reduce the mass scale of the rod, the issue reduces, but this negitively affects my simulation, I need the rod to have some mass in relation to the end attachments. Do you know what causes this issue and how to fix it?
Kind regards,
Matt
p.s. I have added some images that show the situation, but cant show too much more on a public forum. If you need more context or information I can PM you further images or try to recreate in a new project. Hopefully these show enough.
Hi, first of all remember that when your rod has collision enabled (categories) with rigidbody and end particles end up inside it's collider this might cause unstable behaviour.
In case it's correct, please check if your problem is not the same as described in this thread: https://obi.virtualmethodstudio.com/foru...-4532.html
Thanks for the replies. I think I understand how collisions work and I dont think that is the issue, but have posted here to double check.
My connectors on the end of the rod have collision catagory 1, and collide with everything except 0. My rod particles have collision catagory 0 and collide with everything except 1 (is it fine to have 0 (self) selected here?).
If I have a second, inboard, particle then with will have all collisions selected, but the issue is independant of this so I dont think this is relevant.
Qriva, I did actually have that same issue, and commented on that thread at the time (https://obi.virtualmethodstudio.com/foru...l#pid17239), but I think that is a completly seperate and unrelated issue. Although now come to think of it I do occasionly see that issue repeating despite the fix in that thread. Maybe it is related... Not sure.
04-11-2025, 09:44 AM (This post was last modified: 04-11-2025, 09:44 AM by chenji.)
(04-11-2025, 09:31 AM)matty337s Wrote: Thanks for the replies. I think I understand how collisions work and I dont think that is the issue, but have posted here to double check.
My connectors on the end of the rod have collision catagory 1, and collide with everything except 0. My rod particles have collision catagory 0 and collide with everything except 1 (is it fine to have 0 (self) selected here?).
If I have a second, inboard, particle then with will have all collisions selected, but the issue is independant of this so I dont think this is relevant.
Qriva, I did actually have that same issue, and commented on that thread at the time (https://obi.virtualmethodstudio.com/foru...l#pid17239), but I think that is a completly seperate and unrelated issue. Although now come to think of it I do occasionly see that issue repeating despite the fix in that thread. Maybe it is related... Not sure.
Thanks!
What's the resolution of the blueprint? Try to adjust the value of it and/or the value of thickness. You can enable the particle renderer to see how many particles are there.
(04-11-2025, 09:31 AM)matty337s Wrote: Thanks for the replies. I think I understand how collisions work and I dont think that is the issue, but have posted here to double check.
My connectors on the end of the rod have collision catagory 1, and collide with everything except 0. My rod particles have collision catagory 0 and collide with everything except 1 (is it fine to have 0 (self) selected here?).
If I have a second, inboard, particle then with will have all collisions selected, but the issue is independant of this so I dont think this is relevant.
Hi!
I believe the fastest way to diagnose this is by taking a look at the scene itself. Would it be possible for you to share it by sending it to support(at)virtualmethodstudio.com so that I can take a closer look?
