Hi Yang,

I’ve had issues like this in the past, also with coordinate breaks, but I can’t find an example anymore. So, I’m curious to hear what people have to say about your file.

There’s one thing I can suggest though. If you look at your file after conversion, you will see that OpticStudio (OS) has converted the slanted face to a Grid Sag Surface. Now, it could be that if you manage to get rid of the error during conversion, then the slanted face gets converted to a different object, but I doubt it. So, assuming OS would still convert your slanted face to a Grid Sag Surface, this is quite inefficient in terms of raytracing. For such a well determined object as your lens, you will be much better off re-designing it ussing the rules described in this article.

For example, you could create a Cylinder Volume as the base shape for your lens, and then use a Rectangular Volume to remove (or kind of “etch”) a part of the Cylinder Volume to create the slanted face. If you look at my screenshot below, try to imagine what would be left if the magenta volume was removed from the green cylinder:

Using a Boolean Native, we can do exactly that:

And raytracing through a boolean native will be much faster than a grid sag. As a matter of fact, you don’t even need the boolean native if you are interested in raytracing speed (I let you read the aforementioned article for more details).

I have also attached this file to my reply for your reference.

I hope this helps, and take care,

David