Question about the physics of joints in simulations

[fairytalef00]

I wanted to analyze the 7 dof manipulator in the model browser.
What I'm curious about is that when I removed the script and simulated it, the manipulator's link did not rotate.

If I imagine, it was strange that Link was affected by gravity and had to rotate using the joint as the axis of rotation, but it just stayed still.

So I created a simple experiment to understand the principles that apply to joints.

https://1drv.ms/v/s!AuvU7BTQNjpWh_IyTUK ... Q?e=IRe0ho

Except for the joints and links to be verified, all other elements were removed as shown in the video.
I also copied this manipulator to create the exact same model. We will call this the test model.
In the test model, I removed the original joint and recreated a revolute joint whose position is equal to the axis of rotation.

When tested as shown in the video, the test model rotates as I imagined, but the preference model does not show any movement.

I exported the urdfs of the two models and compared the contents, and they were exactly the same.

Why does it look like this?
Are there more features I don't know about?

[coppelia]

Hello,

we cannot see the video you posted.

Keep in mind that a joint can be in kinematic or dynamic mode. And when in dynamic mode, it can be in one of following control mode: free, force, velocity, position, spring or custom.

Most manipulator joints are in dynamic mode, and in position control: so even if you remove the script controlling that robot, your joints should still hold their position.

Cheers

[fairytalef00]

Thank you for your reply.
I've fixed the video again so you can see it.

https://1drv.ms/v/s!AuvU7BTQNjpWh_IyKCQ ... g?e=cbQQ74


Of course, even if you didn't watch the video, I knew what the problem was thanks to your answer.

The original example, the 7 Dof manipulator, was set to 'dynamic mode - position target 0' so it was supposed to keep its original position.

It's a small thing, but I'm so glad I figured it out.

Many thanks.