Unexpected behavior when modeling ball screw mechanism using equality joint with `polycoef` #2563

valeriaskvo · 2025-04-09T13:40:17Z

valeriaskvo
Apr 9, 2025

Intro

Hi!

I am a PhD student at ITMO University, working on legged robots and using MuJoCo for my research on mechanical design.

My setup

I am using mujoco==3.2.7, python API. I use two operation systems Ubuntu 22 and Windows OS.

My question

I'm trying to implement a linear actuator model, specifically a ball screw mechanism that converts rotational motion into linear displacement. The goal is to control the actuator only through a rotational motor, and have the linear part move accordingly via the screw transmission ratio.

To achieve this, I used an <equality joint> with polycoef="0 10 0 0 0" to express a linear relationship between the revolute joint (input rotation) and the slide joint (linear carriage). I expected that rotating the motor by a given angle and stopping it would cause the carriage to move proportionally to that angle (based on the gear ratio), and then stop as well.

However, what I'm observing is that after I rotate the motor to a certain angle and stop, the carriage continues moving, even though there is no more motion on the motor. This seems inconsistent with the expected behavior of a rigid ball screw coupling.

Could you please clarify:

Am I using the polycoef formulation correctly?
Is this expected behavior due to numerical integration or solver tolerance?
Are there alternative (and more stable) ways to implement a ball-screw mechanism in MuJoCo where the slide joint is rigidly coupled to the revolute joint?

Thank you for your help and for developing such a powerful simulator!

Minimal model and/or code that explain my question

My model for linear actuator is:

<mujoco model="linear_actuator_sim">

  <option timestep="0.0001"/>
  <compiler angle="radian" />

  <worldbody>
    <body name="motor_base" euler="0 1.5707 0">
        <geom type="cylinder" size="0.0245 0.05" mass="0.4" rgba="0.8 0.8 0.9 0.5"/>
        <body name="motor_shaft" pos="0 0 0.05">
          <joint name="revolute_joint" damping="0.01" frictionloss="0.01" range="-50 0"/>
          <geom type="cylinder" pos="0 0 0.02" size="0.0049 0.02" mass="0.1" rgba="0.8 0.8 0.8 1" contype="0" conaffinity="0"/>
          <body name="motor_nut" pos="0 0 0.041">
            <joint name="slide_joint" type="slide" damping="0.01" frictionloss="6" range="-0.04 0" actuatorfrcrange="-600 600"/>
            <geom type="cylinder" size="0.02 0.001" rgba="0.8 0.8 0.8 1"/>
            <site name="force_sensor_site" type="cylinder" size="0.0049 0.001" rgba="1 0.65 0 1"/>
          </body>
        </body>
    </body>
  </worldbody>

  <actuator>
    <position name="motor" joint="revolute_joint" ctrlrange="-50 50"/>
  </actuator>

  <equality>
    <joint name="screw_to_nut" joint1="revolute_joint" joint2="slide_joint" polycoef="0 0.01 0 0 0"/>
  </equality>

  <sensor>
    <touch site="force_sensor_site" name="force_sensor" noise="10"/>
  </sensor>

</mujoco>

Confirmations

I searched the latest documentation thoroughly before posting.
I searched previous Issues and Discussions, I am certain this has not been raised before.

valeriaskvo · 2025-04-09T14:19:59Z

valeriaskvo
Apr 9, 2025
Author

I found a mistake in my model. I played with parameters a bit and find what the problem

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Unexpected behavior when modeling ball screw mechanism using equality joint with `polycoef` #2563

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{editor}}'s edit

{{editor}}'s edit

Uh oh!

Replies: 1 comment

Uh oh!

{{title}}

Uh oh!

Select a reply

Uh oh!

Unexpected behavior when modeling ball screw mechanism using equality joint with polycoef #2563

Uh oh!

Uh oh!

valeriaskvo Apr 9, 2025

Intro

My setup

My question

Minimal model and/or code that explain my question

Confirmations

Replies: 1 comment

Uh oh!

valeriaskvo Apr 9, 2025 Author

Unexpected behavior when modeling ball screw mechanism using equality joint with `polycoef` #2563

valeriaskvo
Apr 9, 2025

valeriaskvo
Apr 9, 2025
Author