Testing the Diffuser. it sits in the ND filter wheel – that is, in the collimated beam. Being situated there it removes any spatial information originating in the flat screen itself, the front objective and the first lens of the collimating system. Flat fields taken this way have a central brightening that looks spherical or parabolic in shape.
Perhaps, if that darkening could be understood and subtracted by some fitting procedure, we can use the diffuser to get good high spatial frequency information about the FF? It seems to remove the huge gradients that otherwise characterize the FFs we have been getting from sky, dome and lamp in the past. 
The difference between darkest corner and centre is about 10%. What is the origin of the circular feature, and what does its position off-center tell us?
We fit a second-order surface to the image above and subtract it. Add the mean back in, and normalize. The result is here:
While extremely similar in appearance to the other ‘superflats’ we find that the central red patch is about 1% above the rest.
