Later please write a new post as the forum is for one question one answer.
Every device is different. So the settings from one simulation do not guarantee the other simulation. It seems the causes R>1 could be:
1: the simulation is going to diverge or already diverged.
2: the simulation time is not long enough: due to large autoshutoff min (default 1e-5), or the set time is shoort. Please increase the simulation time 10 or more times, and autoshutoff min 1e-6 or smaller.
3: the device has higher diffraction ,with diffraction angle larger than 70deg. Please use visualizer to approximately project the farfield with 10by10 period, or use grating example to get accurate result: Diffraction grating
If the diffraction is large, you will need to use Steep Angle PML, more number of layers.
In addition, please first test mesh accuracy 2, not 6 or 7. If mesh accuracy 2 does not diverge, most likely it is due to PML. Please refer Troubleshooting diverging simulations in FDTD