unitree-g1-openarm-adapter.stl — units and orientation in the G1 wrist clamp
Hello, and thanks for publishing the G1 hardware modifications — very useful.
We are using STLs/unitree-g1-openarm-adapter.stl as a reference for the Unitree G1 wrist interface (we design add-on hands/adapters for the G1). Two questions:
- Units: the STL seems to be saved 100× smaller than millimetres (bounding box ≈ 0.28 × 0.15 × 0.49). Scaled by 100 it matches the Wuji G1 docking adapter's plug to within 0.03 mm, so we assume the export unit was off — can you confirm?
- Orientation in the wrist clamp: the photos in
media/show the adapter off the robot. Could you share a photo of it mounted in the G1 wrist clamp, or say which way the gripper faces relative to the palm side of the original G1 hand (and which face of the plug points up toward the clamp cover)?
Thank you!
Forenly AI (hi@forenly.ai)
Fixed the units, camera should be facing up when performing the task like so.
Thank you for the quick fix and the mounted photo — that settles the units, and we confirmed the new file is exactly the old one ×100.
To place the adapter numerically on the G1 model, two small follow-ups, using the axes of the corrected STL (mm):
the plug spans Y = −4.97 … 11.5 (cross-section 28 mm in X × ~49 mm in Z), the neck Y = 11.5 … 16.5 and the circular plate Y = 16.5 … 21.5.
- Up direction: in your photo (camera facing up, toward the oval screwed cover of the wrist link), which STL axis points up — +X, −X, +Z or −Z?
- Insertion depth: is the plug pushed fully into the clamp so that the step at Y = 11.5 sits flush against the end of the wrist? If not, roughly how many mm of the plug stay outside?
If you happen to have the G1 + gripper in a URDF/MJCF (e.g. for simulation), the transform from left_wrist_yaw_link to the adapter would answer both at once.
Thanks again!
Forenly AI