PSF generated through on_the_fly True/False different

[spacetelescope]

PSF generated through on_the_fly True/False different

[mperrin]

Things we just checked that didn't explain this:

• WebbPSF version (ETC was using 0.6rc2 3 months ago and there haven't been major changes to 0.6 now)
• wavelength (I thought that maybe it was the ETC using precomputed PSFs at 10.60 and 10.835 microns, but that doesn't seem to make enough of a difference in some quick tests just now)
• sampling (I worried that since the ETC is using pixel/3 sampling it's an odd number instead of even so the FQPM cancellation is not as exact. but again this makes a small difference but nothing so abrupt as above.

[mperrin]

Aaaah, it's obvious if you read the PSF headers closely enough!

CORONMSK= 'FQPM1140' / Image plane mask

The "FQPM1065" PSFs were somehow generated with the 11.40 micron mask.

[mperrin]

Hi @ariedel - you'll want to see this one. @abhijithrajan has just found a problem with the existing PSF library for the MIRI FQPM1065 mode. Good thing you just told me yesterday you're working on updating the PSF generation script. Here's something to add to the to-do list for that!

[ariedel]

Ha, I caught that mistake when going though the PSF generation for a different reason. I'll file a ticket and get that fixed.