Hello Mahmoud,

From the help writeup for a toroidal surface: “The radius of rotation is set on parameter column number 1. To model a cylinder lens which is flat in the X direction use zero, which OpticStudio interprets as infinite radius.”

In the attached file is an aspheric cylinder, flat along the x axis. If you want it flat along the y axis, you can specify a tilt about Z of 90 degrees in the surface properties.

The lens can be given a rectangular shape with rectangular apertures on the front and back surfaces:

Hey David,

thanks for the quick reply. I used Toroidal lenses to simulate our optical system which contains 3 microlens arrays in nonsequential mode and only the 3rd lensarray have power on x axis. The only diffrence from the other 2 lens arrays is that the third lensarray have a higher “coeff1 y^6” value and if i lower it (see attached image) the problem is solved.

How can i simulate the system without changing the values??

Thank you.

It’s hard to tell from the screen shot. Perhaps you could attach a ZAR file in a zip. 

I see you are using non-sequential mode. The Toroidal Lens works the same way as the Toroidal Surface in sequential. A ZAR file will help. But I also note in the image above that the y^6 coefficient has a value of 500 and is not marked as variable in optimization. How did the parameter come to have this value? What is the problem you are seeing and why do you believe the parameter needs to be this value?

Hey David,

the coefficients were calculated with the z ( r ) equation and our goal is to have paralell rays coming out the system but when the rays pass through the third lens they dont remain paralell.

Thank you.

Hi Mahmoud,

I notice something interesting in your design. The defined fields are X-Angle fields with the extreme fields at +/- 2 degrees. But the cylindrical lenses only have power in the Y direction.

The merit function attempts to minimize RMS Angular x+y centroid with non-zero weights in both x and y. Because the cylindrical lenses have no power over the x-angle fields, optimization cannot make rays from non-zero x-angle fields parallel the z axis.

Since optimization is with respect to the centroid, perhaps this does not matter  

I frankly find this design difficult to think about in the context of the usual sequential optimization tools. If it were me, I would layout the design with three lenses in pure sequential mode and use optimization to determine the prescription, and then use that prescription in a pure non-sequential design of the arrays.

Hello Mahmoud,

From the help writeup for a toroidal surface: “The radius of rotation is set on parameter column number 1. To model a cylinder lens which is flat in the X direction use zero, which OpticStudio interprets as infinite radius.”

In the attached file is an aspheric cylinder, flat along the x axis. If you want it flat along the y axis, you can specify a tilt about Z of 90 degrees in the surface properties.

The lens can be given a rectangular shape with rectangular apertures on the front and back surfaces:

 

Dear David,

May I ask how I can draw the same cylindrical lens with limiting X and Y apertures in non-sequential editor? I know I can draw this lens first in sequential mode and then convert it in non-sequential mode. But this approach is very difficult for me as I might want to add much more cylindrical/conical lenses at any time or any location.

Regards,

Jiang

Hi Jiang,

You can construct the cylindrical lens in non-sequential using the Toroidal Lens object the same way as it’s done in sequential mode. Instead of placing apertures on the front and back surfaces, in non-sequential the Toroidal Lens object has parameters Radial Height and X Half-Width which define a rectangular cross section.

I attach a ZAR containing such an object. The file was created by converting a sequential design to non-sequential, but you can examine the resulting non-sequential object and use it as an example for defining such lenses directly in non-sequential mode.