Prismatic joint in spring damper mode gets stuck

Typically: "How do I... ", "How can I... " questions
Post Reply
mthor13
Posts: 13
Joined: 03 Jul 2017, 08:32

Prismatic joint in spring damper mode gets stuck

Post by mthor13 » 03 Dec 2018, 09:10

Hi everyone,

I have a problem with the prismatic joint in spring damper mode.
I am using it in the feet of a hexapod robot. When it presses down on the joint (i.e. stepping with the foot), then it is compliant as expected. But when it lifts the foot again then the joint is stuck and does not return to its initial position.

Extra:
I have set the maximum force of the joint to the maximum value.

Best Regards,
Mathias

coppelia
Site Admin
Posts: 6867
Joined: 14 Dec 2012, 00:25

Re: Prismatic joint in spring damper mode gets stuck

Post by coppelia » 03 Dec 2018, 10:00

Hello Mathias,

what engine are you using? Bullet? You can try to adjust the Stop ERP and/or the Stop CFM to avoid this problem.

Cheers

mthor13
Posts: 13
Joined: 03 Jul 2017, 08:32

Re: Prismatic joint in spring damper mode gets stuck

Post by mthor13 » 03 Dec 2018, 10:47

Hi Coppelia,
I am using Vortex, so that is unfortunately not an option.
I tried to play with the target velocity, which helps. However, it makes the joint very shaky and unstable.

Best,
Mathias

mthor13
Posts: 13
Joined: 03 Jul 2017, 08:32

Re: Prismatic joint in spring damper mode gets stuck

Post by mthor13 » 03 Dec 2018, 19:40

So here are my findings if anyone is facing a similar problem.

The normal way of modelling a spring in V-REP is to use a prismatic joint set to spring-damper mode. However, this method turned out to be unstable and imprecise. The spring can instead be modeled as a revolute joint with the z-axis perpendicular to the spring and x-axis in the same direction. In the vortex engine properties related to joints, joint relaxation along the x-axis can be enabled and the appropriate parameter for stiffness can be set. By directly using vortex's (the physics engines) implementation of a spring a much more realistic, reliable, and stable model of the spring can be achieved. That is at least my experience.

Good luck everyone!

Post Reply