I am simulating a plane wave incident on a different number of dielectric spheres using HFSS vs. 10... I know HFSS has a built in sphere model... but I am finding that it is not very accurate... Does anyone out there have any suggestions on how to set the surface approximation (i.e. what are some good settings for the max surface deviation, max surface normal deviation, max aspect ratio)... It's hard to experiment with this because HFSS is so slow on my computer (2GHz, 2Gb or RAM, Windows XP OS)... Thanks everyone for your help!!!!