A possible impact on MTF of a complex beamsplitter coating has been investigated on a simple system comprising three lenses and two prisms on top of each other. Coordinate breaks and MIRROR surfaces were used to 'construct' the prisms in 2-D. The said coating has been applied to one of the MIRROR surfaces.
The system is polychromatic, on-axis (one field only) and 'polarisation' has been checked in the FFT MTF Setup Window.
Now a surprise: with the coating applied or not the values of MTF differ by up to approx. 3 precentage points in the medium frequency range in 'all fields' mode as compared to the 'field 1' mode. This is not dramatic but very probably shouldn't happen...
Any explanation?
Thank you, Dusan