Hello, I am encountering some issues while using XRT software to simulate a focusing Schwarzschild objective lens.

This is a classic two-mirror reflective structure with concentric convex and concave spherical mirrors, which is theoretically excellent at eliminating third-order aberrations. In my simulation, I used the ToroidalMirrorclass, setting R and r to the same value to approximate a spherical mirror. However, there are apparent problems:

1. After setting up the optical path strictly according to the theoretical structure, the rays fail to achieve an ideal focus. In contrast, I can achieve perfect focus in Zemax (as clearly shown in the attached comparison images between Zemax and XRT).

2. When I rotate either spherical mirror around the main optical axis, the blur spot shape on the image plane changes. This is puzzling because, in theory, I have set up a centrally symmetric spherical mirror. This makes me question the suitability of using the ToroidalMirrorclass to represent a true sphere.

I have attached figures showing the spot change after rotation. I would like to ask for your advice on the potential reasons for these issues and whether there might be a better way to simulate this optical structure in XRT.

Thank you for your time and insights. I look forward to your reply!

****