So I have a checkerboard shaped array whose shape is of the form:

1 1 1 1 0 0 0 0
1 1 1 1 0 0 0 0
1 1 1 1 0 0 0 0
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1

The surface of 1s and 0s sits atop a 1.4mm dielectric medium that is backed by a PEC, where the 1s represent very thin, square copper plates-- and the 0s-- none.

I am fully able to design this structure; however, when it comes time to analyze it I must be doing something wrong.

Let me explain:
So I setup a Floquet Port a short distance above the surface of 1s and 0s (with that same dimensions as my dielectric substrate) and assign to it master/slave boundary conditions that are seen in HFSS manuals and on the internet (no errors, everything looks good). My resonant frequency is 37GHz, and I can sweep from 12-62GHz with ease. However, when it comes time to insert a far field calculation, things become tricky for me.

I can setup a far field calculation and obtain a RCS, but the problem presents itself in the solution! The RCS represents what one would expect from a PEC and dielectric alone, without the 1s. We should expect that, in a graph of RCS vs. Theta (incident angle off of z axis, going from -90 deg to 90 deg) that, at a 45 deg polarization, a significant decrease in RCS would be observed at normal incidence (0 deg). But like I said, there is no such observable. I have also tried to setup a unit cell analysis of this pattern to no avail. My question really is: how should I go about setting up an analysis for this type of pattern? I cannot find any resources specifically explaining how to model these checkerboard patterns in HFSS or CST. I can only find information for arrays of 1s only (both finite and infinite), but this seems to not be useful for my intentions. I have attached an image of the array.