I believe our LOD classes do not support Orthographic camera, only Perspective. In this sandbox you can see how everything just tends to move towards tile zoom 14. Sandbox: https://codesandbox.io/p/sandbox/geo-three-lod-2yfnq4 https://github.com/user-attachments/assets/c8f8f082-64d0-4906-904c-8b7d2351780c