the problem that the joint exceed the position scale

Typically: "How do I... ", "How can I... " questions
Post Reply
liyinghui
Posts: 11
Joined: 26 Apr 2022, 03:37

the problem that the joint exceed the position scale

Post by liyinghui »

I am working with a legged robot model. I set the joint scale from 0 to 50 deg, but when i started the simulation, I found that the joint exceed the given restrict(<0deg). It is true that the model will get the force from the ground because of the gravity, but why does that still happen even though I already set the constraints? And I also set the target velocity to be 0 and lock the motor as well. So I have no idea how to solve this problem. The problem is shown in the video:https://www.youtube.com/watch?v=OlyVFKJK_0Q

liyinghui
Posts: 11
Joined: 26 Apr 2022, 03:37

Re: the problem that the joint exceed the position scale

Post by liyinghui »

the phenomenon after a while in simulation https://youtu.be/DJ1EG6CwtRo. we can see that the revolute axis has changed which is not supposed.

coppelia
Site Admin
Posts: 10339
Joined: 14 Dec 2012, 00:25

Re: the problem that the joint exceed the position scale

Post by coppelia »

Hello,

your mechanism and problem is difficult to understand from the videos. Can you post the scene file?
Also, what physics engine are you using and does that problem appear with other physics engines too?

Cheers

liyinghui
Posts: 11
Joined: 26 Apr 2022, 03:37

Re: the problem that the joint exceed the position scale

Post by liyinghui »

Thanks for your reply!! I'm sorry that I didn't see your reply until yesterday. And given the privacy, I am not allowed to upload the whole model. but I simplify the model https://drive.google.com/file/d/1P_yagV ... sp=sharing and the problem is similar. The main problem is that the revolute axis turns to change which I don't want to see.
The problem is shown in the video.[youtube]https://youtu.be/FrL388lL0EI[/youtube]. I lock all the motors and I suppose the robot would stand still instead of falling with the revolute axis changing. It is true that the joint is under extreme pressure, but I don't expect the revolute axis to change. what can I do?

coppelia
Site Admin
Posts: 10339
Joined: 14 Dec 2012, 00:25

Re: the problem that the joint exceed the position scale

Post by coppelia »

Hello,

the problem is directly linked to the limitations of some physics engines. Namely, you are having way too much differences between the involved masses/inertias, i.e. they differ from a factor 300 or so. Make sure to carefully read the design considerations in the designing dynamics simulations section, maily design consideration 7 & 8.
Also, try to switch to the Newton engoine, and you will see that it handles the situation much better. Also, your lander feet are not part of the dynamic simulation (they are not respondable and not dynamic).
In next release, we'll have a new engine that should handle above problems much better

Cheers

liyinghui
Posts: 11
Joined: 26 Apr 2022, 03:37

Re: the problem that the joint exceed the position scale

Post by liyinghui »

Thanks for your reply! And I already tried the newton egine and it performs better. But it seems that not all the problem is solved. As you mentioned, i need to adjust the mass, while the question is that how should we process the universial joint because the link in this kind of joint is light. Is there any other suggesions? And what you mean next release, will that turn out soon?

coppelia
Site Admin
Posts: 10339
Joined: 14 Dec 2012, 00:25

Re: the problem that the joint exceed the position scale

Post by coppelia »

Normally one would balance masses and inertias across a whole kinematic chain, so e.g. instead of:

Code: Select all

1000kg --> 10kg --> 0.1kg

Code: Select all

900kg --> 90kg --> 10kg
The next release should be out in 2-3 months, but maybe you can get your hands on a beta at some point..

Cheers

Post Reply