Hi Eran, I might be the only one experiencing this problem, but I’m having some trouble envisioning what you’re trying to accomplish. Could you provide any sort of diagram or visual example?

You’re probably on the right track with the idea of using the multiple configuration editor (MCE), but I can’t yet give any more specific advice.

Hi Alissa,

The required magnification in yz is 42times than in xz, while spotsize can be 2 times larger at yz.

Hence I have no other idea but to use cylindrical shape lenses, or bi-conic.

While using cylindrical shape, I have the privilege to separate the system for 2 configurations, each one with different aperture location. But, If I would like to further decrease system size and number of surfaces this approach wouldn’t be valid anymore.

Hi Eran, if you need one system that has different performance metrics in X and Y, then you may not want to use the multi-config editor at all… the reality is that your system is not changing (like a scanning mirror or zoom lens), you just have non-rotationally symmetric requirements. 

For example, you mentioned using cylindrical lenses (yep) and in fact any surface shape that has different X and Y components will work for you in this situation.

A few other tips:

1. You’ll want to have a full grid of field points defined in +X, -X,  +Y, -Y and various points in between. This will ensure that all analyses are accounting for full field performance, not assuming any rotational symmetry.
2. When setting up your merit function, you can start with a wizard-built MF but then you’ll want to confirm that “assume axial symmetry” is disabled and eventually you’ll add your own operands by hand, like GENF or GENC as a substitute for RMS spot size. This is because the RMS will not give you different measurements for X v Y, but GENF and GENC can do this.
3. Your best bet for measuring & optimizing magnification is to use REAX, REAY targeted to specific locations on the image surface. Operands like PMAG won’t get you what you need due to the X/Y asymmetry.

Hi Alissa,

Thanks for the advices, 1 & 3 I knew, but No.2 is really helpful! I really encountered this issue with spot size.

Nevertheless, I still ask myself if to start from 2 independent systems wouldn’t be better as a starting point. If the lenses are pure cylinders the aperture can be split for two 1D apertures, one for XZ and the other for YZ.

Hi Alissa,

regarding No.2  - I replaced the optimization wizard default operands by the GENC for each field point with X & Y encircled energy criterions but seems the performance only got worse. Something I’m missing here ?

Hi Eran, you can try 2 things…

1. You want to make sure that you add on top of the wizard-built MF, don’t try to directly replace any lines. The wizard built MF has rows in a specific order and manipulating anything after the DMFS line can spell trouble. You can select the DMFS line and use your insert key to add new rows on top. Then you can play with the relative weights of the wizard-built portion and your own operands.
2. You can also try optimizing your system first with a wizard built MF just to get into the realm of reasonable systems, then remove that MF entirely and build your own with things like GENC. 

As for starting with two discrete systems, I guess I’m not yet convinced that it’ll help you achieve the outcome you need… but if you try it and find out that it does help, please update this thread!