01 / Simulation accuracy

Your simulator says clear.
Your robot disagrees.

Collision checking runs during toolpath calculation — not after. The simulation is the execution model, not a separate approximation of it.

SIM // ENCY ROBOT [ synced simulation viewport ]

CELL // WALL-CAM [ real robot, same frame ]

"If you don't crash a robot at least once per project, you're not really doing the job."
SkyeFire · robot-forum

"The SprutCam simulation is totally fine, but practical application is not."
Mischa · milling thread

"In RoboDK on simulation the robot rotates the 4th axis by +. The real robot rotates by –."
serg_kub · robodk.com/forum

The cost of a sim-vs-reality gap

One crash isn't a debug session. It's the quarter.

$14k–$200k

Per scrap event in aerospace composite trim — layup labor plus autoclave time, sunk before the cut.

$22–50k / min

Automotive line-down, back-billed to whoever owned the cell when it crashed.

20–30%

Contract holdback you lose when a FAT crash happens with the customer in the room.

What changes with ENCY Robot

Tool, holder and machine collisions surface while CAM calculates — every move you commit is already verified. docs ↗

Mill operations simulated against true solids — stock state matches the part, not a voxel approximation. docs ↗

Drop in a robot; ENCY maps its axes and generates collision-free, singularity-avoiding trajectories — no week of setup.

Transitions in machine-space shortest path — the path the robot actually takes, not the one geometry suggests.

30-minute live walkthrough on your geometry — your fixture, your robot, your tooling. Engineer-to-engineer. No slides.