Hello, 

The beamsplitter is equivalent to a plane parallel plate. So, when you move it axially, you don’t change the position of the image. Zemax is right and thankfully so ! 

Regards, 

Thierry Lépine. 

Hi ​@Thierry.Lepine ,

Thanks for the reply! I get what you mean and I think you're right in the general case.

In my system though, the image is at infinity. I'm working afocally with collimated rays in image space and evaluating the wavefront error against a flat reference. So what I care about is not where the image forms but the vergence of the rays at the sensor.

The problem I'm trying to solve is more of a mechanical one: if during assembly the beam splitter ends up sitting 20 mm higher in Y, the physical distance between surface 2 (the paraxial lens) and the beam splitter drops by 20 mm as well, from 239.617 mm down to 219.617 mm. That change in distance does affect the ray vergence at the image plane and introduces a defocus, which you can actually see in the 3D layout when you manually reduce the thickness of surface 2 by the same amount.

So what I'm really after is a way to link the Y decenter of surface 4 to the thickness of surface 2, so that when I tolerance the decenter the thickness updates automatically.

Best regards,

Jesús

Hi Jesus, 

I though you were talking about the positions of the intermediate images. 

That said, I still don't quite understand your problem. When you decenter the beamsplitter, the optical paths of the two beams within the beamsplitter don't change and the two intermediate images don't move. This because the beamsplitter behave as a plane parallel plate. So I don't see the problem. 

Thierry Lépine.