For coronagraph masks, add "bar_offset_y" for NIRCam. Add similar for the round occulters and MIRI as well.

[mperrin]

Issue by mperrin
Monday Feb 19, 2018 at 20:32 GMT
Originally opened as mperrin/webbpsf#191

---

As requested by @ariedel for use in Pandeia ETC coronagraph generation script

[mperrin]

Comment by ariedel
Monday Feb 19, 2018 at 20:33 GMT

---

Use case is that Pandeia requires the PSF to be centered on a pixel; this currently requires us to snap to pixel positions rather than the actual requested XY location.

[mperrin]

Comment by mperrin
Monday Feb 19, 2018 at 23:45 GMT

---

This turns out to be a real pain in the neck for MIRI, given the different rotated coordinate systems. Because of how the optics are implemented, the "natural" way to offset the FQPM is in V2V3, rotated by ~ 5 deg from the instrument detector X and Y (which is what Pandeia wants to work in). So even implementing a "simple" option to offset the focal plane mask by a given amount inevitable is going to involve some coordinate transforms. Not hard, just more work.

NIRCam's comparatively simple.

[mperrin]

Comment by mperrin
Friday Apr 27, 2018 at 22:38 GMT

---

@ariedel, remind me where we left things with the coronagraph PSF generation script. I have this open issue about adding an option for shifting the masks around, and added code back in February to implement it. I'm not sure if any more is needed now? If you're happy with the coronagraph PSF generation script then good and I won't worry about any more effort on this issue.

[ariedel]

To clarify:
As a developer, I want to be able to model a typical pointing error for coronagraphy modes. This 'typical pointing error' is roughly 5 milliarcseconds.

Right now, I can't do that because I need all of my PSFs to be centered on a pixel. That limits me to offsets no smaller than the pixel scale of the detector (or, if I take advantage of oversampling, pixel_scale/oversample). We currently simply don't have a pointing error PSF - and that's been accepted, even if it's less than optimal.
I could, in future, create a 1-(oversampled)-pixel offset PSF (~10 mas for NIRCam SW, ~20 mas for NIRCam LW, and ~33 mas for MIRI), which wouldn't quite be the same but might still convey useful information to users.

If there's another way to get this apart from offsetting the occulting elements (rather than the PSF location), I'm open to alternatives.