Orn_err is always zero

orn = torch.tensor([0,0,0,1], device=self.device).repeat(self.num_envs, 1)
orn_err = self.orientation_error(orn, orn)

In the control_ik function, the orn_err is set to be the error of two identical quaternions (orn), which the outcome is always zero. Is there any reason to ignore the orientation error for the inverse kinematics calculation? 