I'm trying to simulate a system that focuses a beam with lenses after it reflects off of a DMD. I have made a rudimentary model of a DMD using NSC, but if I try to use POP I get an error message saying that the entrance pupil can't be located. I noticed that most of the posts that come up about POP show setups that are purely sequential. This might be a naive question, but is it not possible to use POP when you have non-sequential components in your system? Or is it because I have my entire system in non-sequential mode, so it doesn't take any of the lenses to be the entrance pupil?
Using the regular ray tracing method in non-sequential mode doesn't show any of the expected diffraction from the mirrors, and I don't observe a Fourier transform where I would expect.
Thanks for your help!
Best answer by Mark.NicholsonView original
If you're using hybrid mode NSC, make sure the stop is before the NSC group and use an 'object-specific' aperture definition like entrance pupil diameter or float-by-stop. You'll probably also need to use a plane phase reference and have pretty high sampling. Check out the sample file Documents\Zemax\Samples\Physical Optics\Lenslet Array.ZMX for an example of a similar system.
There's a little-known capability in Non-sequential that might do what you need better. Check out the sample file Samples\Non-sequential\Coherence Interference and Diffraction\Interference Example 4- Diffraction Limited Imaging.ZMX. Note the use of the Huygens Mode flag on detector object 6. This adds the rays landing on the detector as a coherent sum of plane waves to give the production of a full PSF rather than a geometric point image:
This may be the easier way of getting what you want. I recommend raeding the documentation on the Huygens Mode of the detector
Quick follow up, check out https://my.zemax.com/en-US/Knowledge-Base/kb-article/?ka=KA-01402 for a full discussion of the POP and lenslet approach. The same thing should work for POP and DMD.