Yes, because the shape of the object, and the camera angle using the provided stand, does not allow it to see enough of the object, and there are areas where scan information is missing.
If you do a second pass, by rotating the object to give the camera a better view, you have to be careful, not to move it too much, as it can only reference with respect to what it saw previously using its scanned data.
I thought it would have realised, the valve cutouts were a different shape, and maybe picked up on the "93.xx mm writing" on the dome, and aligned this, but I think looks for larger feature variations and aligns with this??

Mike