Drawer and fork

Articulations

Pull open a drawer, then put a fork in its tray.

Play0.00 s

¼×½×1×2×

Loop

The steel preset (the default), recordedon an NVIDIA RTX PRO 6000 Blackwell.

Run it

cd examples
# the default preset (steel), then the others
uv run shelf play drawer_fork
# change solver settings live
uv run shelf tune drawer_fork

Source: examples/shelf/07_articulations/drawer_fork/ · ported from steel-webgpu assets/scenes/drawer_fork (hero_montage export) @ ad03e49

Presets

Preset	Solver	Status	Checked
steeldefault	steel	works	2026-11-10, newton f6fc82a8b, NVIDIA RTX PRO 6000 Blackwell Server Edition

Solvers with no row are untested on this example. The full settings are inpresets.toml.

Code

example.py

examples/shelf/07_articulations/drawer_fork/example.py

# SPDX-FileCopyrightText: Copyright (c) 2026 Skild AI and the Steel contributors
# SPDX-License-Identifier: Apache-2.0
# Direct source: steel-webgpu; see presets.toml and THIRD_PARTY.md.
# Modified for Steel Shelf: preset-driven execution and shared runner integration.

###########################################################################
# Example Drawer and fork
#
# The arm pulls an unpowered drawer open by its handle, then picks up a fork and places it in the drawer's tray.
#
# Rebuilt from steel-webgpu's exported scene (assets/scenes/drawer_fork: scene.json
# and scene.bin, copied next to this file); see common/exported_scene.py. The
# visual meshes draw when a steel-webgpu checkout sits next to the shelf.
#
# Command: python -m shelf run drawer_fork
#
###########################################################################

import newton.examples
from common.exported_scene import ExportedSceneExample


class Example(ExportedSceneExample):
    file = __file__


if __name__ == "__main__":
    parser = Example.create_parser()
    viewer, args = newton.examples.init(parser)
    newton.examples.run(Example(viewer, args), args)

example.py on GitHub

presets.toml

examples/shelf/07_articulations/drawer_fork/presets.toml

example = "drawer_fork"
source = "steel-webgpu assets/scenes/drawer_fork (hero_montage export) @ ad03e49"
default = "steel"

[check]
# Arguments `python -m shelf check` passes in test mode.
args = []

[gallery]
short = "Pull open a drawer, then put a fork in its tray."
title = "Drawer and fork"
frame = 420

[xpbd]
status = "untested"

[vbd]
status = "untested"

[mujoco]
status = "untested"

[featherstone]
status = "untested"

[semi_implicit]
status = "untested"

[feather_pgs]
status = "untested"

[steel]
status = "works"
# 4 substeps x 12 iterations, immediate response: 4.8 ms a frame on an RTX PRO 6000 (the port's 4 x 8 colored:
# 6.4 ms). One robot and ~150 contacts a step, so the serial sweep beats the colored one. The drawer opens to its
# 0.48 m stop and the fork lands in its tray within 0.4 mm of a 16 x 32 run; penetration 0.65 mm peak, 0.05 mm mean.
# 4 x 8 lets the fork slip 8 mm in the grip, 3 and 2 substeps reach 2.7-3.5 mm peaks, body-pair contact reduction
# shifts the loose cutlery 3.6 mm. Sleeping on: the cutlery and drawer sleep, and without it a frame costs 60% more.
# Joint limits (the drawer's stop, the fingers), built within 0.1 of a limit; no joint has a finite velocity limit.
sim = { fps = 60, substeps = 4 }
solver = { solve_mode = "matrix_free", iterations = 12, articulated_contact_response = "immediate", enable_joint_limits = true, joint_limit_activation_gap = 0.1, enable_joint_velocity_limits = false, enable_sleeping = true, free_body_constraint_max = 8192, articulation_constraint_max = 1024 }
checked = "2026-11-10, newton f6fc82a8b, NVIDIA RTX PRO 6000 Blackwell Server Edition"

[kamino]
status = "untested"

[implicit_mpm]
status = "untested"

[style3d]
status = "untested"

presets.toml on GitHub

← Controller operational space hybrid force motionAll examplesRagdoll →