Gemini GMOS

Overview

This file summarizes several instrument specific settings that are related to the Gemini/GMOS spectrograph.

Nod and Shuffle

For the time being, we have enabled reductions of data taken in Nod+Shuffle mode by simply replicating the calibrations. That is, we do not subtract the nodded images but reduce it as if it were a separate slit.

For this mode, you need to specify the gemini_gmos_north_ham_ns spectrograph; i.e., specify this spectrograph when running pypeit_setup and ensure your PypeIt Reduction File shows:

[rdx]
    spectrograph = gemini_gmos_north_ham_ns

Long Slit

1. Somewhat too frequently when using the longslit, the “3” slits are not all identified in the bluest detector. By default, PypeIt now reduces Gemini/GMOS data by mosaicing the detectors, so this may only now be an issue if you reduce the detectors separately.

To mitigate this, we recommend adding to this to your PypeIt file:

[calibrations]
    [[slitedges]]
        det_min_spec_length=0.1
        fit_min_spec_length=0.1
        edge_thresh=3.

One can also do a manual check with (beware using --debug may show a lot of plots):

pypeit_trace_edges -f my_pypeit_file.pypeit --debug --show
pypeit_chk_edges Calibrations/Edges_file.fits.gz

The code may fault and say there were no valid traces. This happens for some long-slit data where the slit edges are, in fact, beyond the edges of the detector. It returns an error:

TypeError: unsupported format string passed to NoneType.__format__

The solution is adding this to PypeIt file:

[calibrations]
    [[slitedges]]
            bound_detector = True

If bound_detector is True, the code will artificially add left and right edges that bound the detector.

Warning

Beware for faint objects! In this case, the object tracing crutch will be a straight line, likely not following the true object trace.

3. In some cases, the slits were detected, but then rejected due to a failure of wavelength calibration. The appearance of this issue is very similar to that of item 1 above, you will see a lack of slits in skiing, the difference is that, there will be such information on slit detection in log files: Not enough useful IDs.

In this case, you should check your wavelength solution, and try to adjust the wavelength parameters. This issue may be solved now that by reducing the detectors as a mosaic by default.

Wavelength Solution

Faint Lamps

The CuAr lamps are pretty faint in the blue which lead to some “unique” challenges. At present we have lowered the default tracethresh parameter to 10, i.e.:

[calibrations]
    [[tilts]]
        tracethresh = 10.  # Deals with faint CuAr lines

It is possible you will want to increase this, but unlikely.

FWHM

We also have a report (issue #1467) that the default value of the parameter fwhm_fromline=True can sometimes lead to poor wavelength calibration. If your RMS is a factor of 2-3 too high, consider setting:

[calibrations]
    [[wavelengths]]
        fwhm_fromlines = False

MultiSlit

Mask Definition

PypeIt can now take advantage of the mask definition file generated when one designs a GMOS mask. To do so, one needs to provide two additional files and specify them with the PypeIt Reduction File:

The mask definition file, aka ODF file
An aligment image (taken with the spectra)

The mask definition file must be the output generated from GMMPRS and in FITS format. We do not support ASCII mask files currently.

For the alignment image, ensure that the alignment stars in the image are centered in the mask’s square alignment slits. i.e. choose the final image in the sequence of alignment images from the observations.

The modifications to the PypeIt Reduction File will look like:

[calibrations]
    [[slitedges]]
        maskdesign_filename = GS2022BQ137-05_ODF.fits,S20221128S0038.fits
        use_maskdesign = True
[reduce]
    [[slitmask]]
        extract_missing_objs = True
        assign_obj = True

The two files provided must be located either:

in the path(s) of the raw files provided in the Data Block,
the current working data, and/or
be named with the full path.