Typically, an 8-deg angle-polished fiber is mounted with its axis at a 4-deg angle to the optical axis.  Therefore, from the perspective of the coupling optics, it effectively looks like a non-angled fiber mounted with its axis along the optical axis.  This means for optimizing coupling optics or simply computing coupling efficiency, you can essentially neglect all tilts for the standard FC/APC configuration.

More detail can be found here:

 and here

Not sure if you are using the standard FC/APC configuration, but if not (so that only the fiber endface is being angled and the fiber axis is *not* tilted) then the coupling efficiency will definitely be reduced.  In this case, I think the approach you describe should be okay for determining this reduced efficiency.  To restore optimal efficiency, set the first coordinate break to 12 degrees and the second coordinate break to -8 degrees.

Regards,

Jeff

Thanks Jeff, that helps.

I am using a standard FC/APC configuration and the optical axis of the beam is tilted to the fiber. If I do not add any tilts before the image surface, do I need to assign it any material?

Not unless you are using polarization mode, in which case the image surface material choice will determine the Fresnel reflection loss from the fiber face.