NOAO System Technology Center

KOSMOS
Kitt Peak Ohio State Multi-Object Spectrograph

KOSMOS and COSMOS News

Science and Top-Level Requirements
System Engineering and Design
Management
KOSMOS and COSMOS News
OSMOS Web Pages at OSU
Home Page

November 18, 2013
After a long hiatus, here is an update on KOSMOS and COSMOS.

Executive Summary

  • KOSMOS has been completed and used on the Mayall telescope (Octoer 2013)
  • It works really well
  • Long-slit mode has been exercised: wavelength coverage from 3800-9600Å at R~2200 (2 grisms)
  • Preliminaries for testing multi-slit mode have been completed
  • We are working on user documentation

But in the meantime, read on....

Instrument Overview

KOSMOS (and COSMOS) are adapted from the successful MDM instrument, OSMOS, built by Paul Martini and the instrumentation group at Ohio State. Paul is the PI on the two NOAO instruments, with Sean Points as instrument scientist and Jay Elias as project manager. Links to the OSMOS instrument design can be found here:

http://www.astronomy.ohio-state.edu/MDM/OSMOS/

In outline, the instrument optics contain:

  • A precision shutter, which allows precise exposures down to the sub-second level
  • A slit mask wheel, which can hold up to 5 long slits or multi-slit masks (plus an open position, required for imaging or acquisition). Slit masks can be changed during the night.
  • A disperser wheel, which can contain up to 5 dispersers (plus an open position for imaging/acquisition). The dispersers are grisms that operate in a fixed configuration that allows the same mechanical configuration to be used for all grisms and for imaging. Currently, there are only 2 grisms available (see below). If there should ever be more than 5 grisms available, grism changes would occur during the day only and would need to be scheduled in advance.
  • 2 filter wheels, which can hold a total of 10 filters plus the necessary open positions. At present, 5 of the positions are used for dedicated blocking filters and the other 5 can be used for imaging filters or specialized blocking filters.
  • A refractive camera
  • A CCD dewar. At present, this is a 2k x 4k e2v CCD, but we plan to make an LBNL CCD available some time next year. CCD dewar changes would occur during the day only and would need to be scheduled in advance.

Performance

The following figures show:

Fig 1 - System throughput using the blue and red grisms. This is slitless throughput measured from the top of the atmosphere, and therefore includes both the atmosphere and the telescope mirror coatings in addition to the spectrograph itself. This is a factor of 2 or more better than what is quoted for the RC Spectrograph. However, users are cautioned that KOSMOS will generally be used with a narrower slit than RC Spec, so "real" gains in performance will be somewhat less.

Fig 2 - Resolution with the blue grism and a hypothetical 1-arcsec slit. The current set of KOSMOS slits are 0.6, 0.9, 1.2, 1.5 and 3 arcsec, equivalent to integer numbers of pixels, so the resolution one can achieve in practice is either slightly higher or slightly lower. The optics do support use of the narrowest slit, although focus may not be entirely uniform across the field - look for updates on this topic. The variation in resolution is due to the combination of the VPH grating and dispersion in the prism material.

Fig 3 - Red grism resolution, as in Fig 2.

How KOSMOS Works

The user interface for KOSMOS is adapted from the "NOCS" software already in use for Mosaic and NEWFIRM. It is intended to reduce overheads at the telescope by allowing users to prepare observing scripts in advance (either before they come to the telescope, or at the telescope ahead of the observation). What follows is not a detailed description of use of the NOCS, but rather a top-level view of the observing process:

  1. Acquisition. Set up the guider, then take an image of the field containing your target(s). The full field of view is outlined below, and for MOS mode you will most likely want to image over the full field. For single-object spectroscopy a smaller region of interest can be used to speed up readout times (5 x 5 arcmin). You also take an image of the slit (or mask) you want to use, and use IRAF and the 2 images to determine the offset to put the object on the slit; for MOS mode there is also likely to be an adjustment to the rotator. Note - the initial rotator setting needs to be done at zenith, whether it's to get the right field rotation for MOS mode, or the parallactic angle in single-object mode. Small adjustments can be done without returning to zenith.

Fig 4 - KOSMOS imaging field of view. Once you've made the adjustment, you will likely want to check by repeating the sequence, and you may need to iterate if the adjustment was large enough.

Fig 5 - Stylized view of acquisition. The slit image would show light through the slit at the approximate position of the red line, and the object should be visible in the object image. If the object is centered on the slit, and not too faint, you should see it on the slit image when you check the centering. Note that you can acquire through a different filter than the blocking filter (or open position) used to obtain spectra.

  1. Observe. The grisms are highly repeatable, so it's practical to observe an object with both grisms in sequence. Depending on the wavelenth region, a blocking filter may be needed to block second-order light. Note - although there are only 2 grisms in fixed configurations, there is more choice of wavelenth coverage: you can choose to use a slit that is offset by ~2.5 arcmin in the dispersion direction (either way), which shifts the wavelength coverage by a little more than 10% in either direction. The throughput curves will peak in the middle of the new range. Curves similar to Fig 1 should be available eventually for the offset configurations.

Fig 6 - Spectra taken through blue and red grisms (center slit position). These are reduced spectra but give an idea of what you would see in the raw images. The figures only show the central ~100 arcsec, so night sky lines aren't as obvious as they would be in the full image.

  1. Calibrations. What calibrations are needed? The calibration system in the cassegrain rotator guider does not illuminate the full field of KOSMOS, although it provides some arc lamp illumination for almost the entire field. The continuum lamp is not uniform along the slit, and the "white spot" should be used. Depending on your program, arc exposures may or may not be necessary at each object position (there should almost always be a few night sky lines).
Current Status
  • The October observing run was primarily used to test the long slit observing mode, and obtain preliminary observations for multi-slit mode (scale check with astrometric field). Data were obtained for throughput measurements for the alternative slit positions (red/blue offsets described above).
  • The observations showed more flexure than we expected, which was subsequently tracked down and resolved.
  • The next run is in December, and it is intended to fultill several purposes:
    • Complete formal acceptance of the instrument
    • Re-measure flexure
    • Test multi-slit mode
    • Complete commissioning of long-slit mode
    • Verify a series of software improvements intended to make the instrument more "user friendly" or more efficient.
    • Test drive the draft user's manual; we hope to post a link to the manual soon after the observing run is over.
  • Additional time is available in January; we expect to post updated documentation on performance around the time of the call for proposals in late February, 2014.
  • Meanwhile, assembly of the COSMOS camera is complete, and full assembly and checkout of the instrument is awaiting completion of the CCD dewar. The optics have been verified in the lab. We will do a complete checkout of COSMOS in the lab in Tucson before shipping to Chile (nominally in January 2014, with the first T&E run early in semester 2014A). Ideally, we will have some preliminary results prior to the March proposal deadline, but regardless the instrument should be available in shared-risk mode in 2014B.
  • Availability of MOS mode for both instruments in 2014B is not (yet) guaranteed, although at present it seems likely. MOS mode would obviously be shared risk.

May 1, 2012
Although periodic updates have been provided through the NOAO Newsletter, this page has not been updated for a while.

As of today, both KOSMOS and COSMOS are complete except for the optics and some of the CCD dewars. The status of the missing items is as follows:

  • The collimators for both instruments are complete. The camera for KOSMOS is missing a triplet, which has failed low-temperature testing after cementing on two attempts. A revised approach to the cementing has been devised and will be implemented shortly. The COSMOS camera has similar issues, but we are deferring the work on COSMOS until the KOSMOS cementing succeeds.
  • Each instrument will have 2 CCDs: an e2v device with broad-band response, and an LBNL device with better red response. The KOSMOS e2v CCD is ready for integration with the rest of the instrument, but we are using the wait time for the optics to continue optimizing performance. Depending on schedule, we will either perform lab acceptance for COSMOS with this same CCD or with the COSMOS e2v CCD. Work on the LBNL CCDs has not yet started, and they may not be available when the instruments are first put in service.
Barring further problems with cementing the camera triplet, we expect to perform lab acceptance on KOSMOS over the summer, with commissioning at the Mayall to occur after the conclusion of summer shutdown, probably late September or early October. The details of COSMOS commissioning are related to availability of the f/8 focus on the Blanco; telescope acceptance should not occur until after the start of 2013.

August 16, 2011

This page will be updated periodically with news related to both instruments, including updates on commissioning, availability and other information of interest to prospective observers. For the moment, we include an image of the two instruments undergoing mechanical and electrical integration in Paul Martini’s lab at The Ohio State University.

The spectrographs are shown on their handling carts in the up-looking position, which is the same way they will mount in the cass cages of the 4-m telescopes using an adapter that is not shown. The dewars install on the bottom of the instruments. Various covers are not yet installed in these photos.

More information will be provided in the September issue of the NOAO Newsletter. Look as well for an update to this web page in about 2 weeks.