I am using a Unitree G1 and have noticed a significant difference in motor temperature behavior depending on the controller.
When I operate the robot using the default manual controller, the “High Temperature” warning rarely appears. However, when I control the same robot using the Gear-Sonic controller, the warning appears after a relatively short time, even when the robot is simply standing still without performing any motions.
I have heard that other users may have experienced a similar issue.
Observed behavior
- Default/manual controller: The robot can stand or operate for a long time without a high-temperature warning.
- Gear-Sonic controller: The motor temperature rises quickly, and a high-temperature warning appears even while the robot is standing still.
- The same robot and hardware configuration are used in both cases.
Questions
- What could cause this difference between the default controller and the Gear-Sonic controller?
- Does Gear-Sonic use different joint stiffness, damping, PD gains, target poses, or feedforward torque values from the default controller?
- Is it possible that the controller continuously generates excessive holding torque because of:
- an unsuitable standing posture,
- incorrect joint offsets,
- overly high stiffness gains,
- oscillation around the target position,
- or conflicts with the low-level controller?
- Are there any recommended gain settings or standing configurations for preventing motor overheating?
- Which logs or parameters should I check to diagnose this issue?
- Is this a known issue, and are there any existing fixes or recommended patches?
Any advice on how to identify the source of the additional motor load would be greatly appreciated.
I am using a Unitree G1 and have noticed a significant difference in motor temperature behavior depending on the controller.
When I operate the robot using the default manual controller, the “High Temperature” warning rarely appears. However, when I control the same robot using the Gear-Sonic controller, the warning appears after a relatively short time, even when the robot is simply standing still without performing any motions.
I have heard that other users may have experienced a similar issue.
Observed behavior
Questions
Any advice on how to identify the source of the additional motor load would be greatly appreciated.