Hi Wubin,

Thank you for your post! It's possible in certain cases that the Automatically computed Clear Semi-Diameter is not accurate to accomodate all rays. Without seeing your file, it's a bit difficult for me to say, but I have seen similar cases happening when there are hard apertures placed on surfaces preceding the current surface. You can find a discussion of how OpticStudio computes the automatic clear semi-diameter and when it may not be accurate in the Help File at  The Setup Tab > System Group (the Setup Tab) > System Explorer > Aperture (System Explorer) > Fast Semi-Diameters.

I wonder if there is any hard apertures existing in your system and if you temperorily remove all hard apertures, would that make any differences in this case?

Best regards,

Hui

Even with Fast semi-diameters off, we still use marginal rays to estimate the size of the surface. Wubin, look at the marginal ray that touches the top of the Stop and follow it through the layout plot...it hits the image surface below the rest of the bundle! Such a severe caustic means that marginal rays cannot describe the size of the ray bundle.

Personally, I would put up with this until I had optimized the system such that such large caustics do not occur, but if you can't do that you'll have to set the semi-diameters by hand...there's no way to predict which ray to use to compute the semi-diameter for you.