I did some more tests with my mask last night operating at 500mm (f/2.8) on my Epsilon.

I discovered that best overall focus (reported by ccdinspector) occurred when the lines did not quite intersect. This struck me as odd.

My only idea is that the camera is tilted very slightly and so to reach best overall focus over the chip I have to slightly defsous the center.

So I moved the star into all corners of the CCD but the diffraction pattern did not change enough for me to detect by eye.

I'm wondering if at f/2.8 the critical focus region is so small its going to be difficult to visually detect differences.

Has anyone focused a telescope faster than f/4 with this mask?