Steps between cameras in spectra of tile 75000

desihub / desispec

DESI spectral pipeline

BSD 3-Clause "New" or "Revised" License

35 stars 24 forks source link

Open julienguy opened 3 years ago

julienguy commented 3 years ago

Segev: Last weekend while doing VI for BGS Andes reinspection of tile 75000, I was struck by the large fraction of spectra with mismatches in the regions between the B, R, and Z cameras. For example, you can seem them easily in petal 9 https://data.desi.lbl.gov/desi/spectro/prospect/target-reinspection-andes/BGS_7050/specviewer_BGS_reinspection_70500_9.html in gallery view. There are quite a few discontinuities and sharp dips or spikes at the camera overlaps.

Screenshot from 2020-11-11 14-48-26

julienguy commented 3 years ago

I suspect it is due to a bad background subtraction in the blue CCD. We should see if the new bias + dark model improves this.

julienguy commented 3 years ago

Example plot

plot_spectra -i $DESI_SPECTRO_REDUX/andes/tiles/70500/20200303/coadd-6-70500-20200303.fits -t 35191232049185004