3. Slitamasks: Mask Design and Fabrication Software

3.1 Overview

DEIMOS gains a large multiplex advantage through the use of multiobject slitmasks, allowing spectra for typically 80 objects (per barrel) to be gathered at once. However, the masks must be designed and fabricated on a field-by-field basis, and thus constitute a changeable component of the spectrograph hardware, which must be "built" for each new set of targets. The design, manufacture, handling and storage of slitmasks is a major component of DEIMOS operations, some of which is controlled by the astronomer and some by the Keck support staff. The software for slitmask design and fabrication is given below. The operational steps (fabrication through installation in the spectrograph) are appended.


The following goals are included in the software design: The procedure for mask design and fabrication is based on that developed at UCSC for LRIS observers. These elements are included:

3.2 Figures

Fig. 3.1-1 -- Slitmask Design and Fabrication Flow (See also figure 9.2.6 for a Keck-oriented perspective)

Fig. 3.4-1 -- Optical Distortions shows the expected distortions at the CCD.

3.3 Nomenclature

Target list refers to the list of targets for which slits appear on a single mask, including alignment objects. For each object, the list includes ID, RA, Dec, magnitude, priority/code, optional slit PA, and optional slit lengths. (Priority/code is used to identify alignment objects or to assign a relative priority to program objects for use in automated selection algorithms.) A target list will also contain a line for the mask RA, Dec and PA.

Mask design refers to the selection of a final target list, and the specification of slitmask position and orientation on the sky.

Map file refers to a file which describes a slitmask. This file includes a physical mask layout (slit locations and PAs in slitmask coordinates; and object locations in three coordinate systems --- slitmask (in mm), CCD (in pixels) and celestial. (Slitmask coordinates refer to a physical coordinate system tied to the slitmask.) A single map file is associated with a slitmask and is used for mask fabrication, mask alignment at the telescope, automated spectral reductions, and archiving. The map file is a text file so that it may be edited by the user. Appended to this file are sections describing the intended targets, for inclusion in the archival database. (Current plans call for the locating-pin holes to be described in this file as well.) NB: we treat the map file here as a single entity, but in practice the information contained in the map file will be stored in the database and recalled as necessary.

The term astrometry will be used to refer to the mapping celestial coordinates onto the focal plane of the telescope (ie, at the slitmask) and onto the focal plane of the camera (ie, the detector). Astrometry involves the mapping of the celestial sphere onto a cylindrical slitmask surface, and is complicated by telescope distortions (up to the telescope focus) and collimator/camera distortions between the telescope focus and the detector.

3.4-3.5 Software Elements

Mask Design

The mask design program culls a specific target list from a larger list of potential targets, and selects coordinates for the field center and a position angle for the mask. It provides an approximate layout of the slits on the mask without concern for precise astrometric mapping (see the "mapping" next). This software is graphical and highly interative, allowing the user to adjust the position and orientation of the mask with respect to the targets, add/delete targets, etc. There will also be automated target selection options. The input and output target lists will have identical formats so that the program may be run iteratively, either to build up a final target list or to modify an existing target list. All inputs and outputs will be in celestial coordinates. A secondary list (of lower priority objects) may be input if desired. In addition, there will be practical checks (eg, telesope limit violations) and a display of the atmospheric dispersion effects relative to slit size/orientation.

Inputs include:

Output is a target list in the same format as the input list. The program will have the mask outline (including the positions of the locating pins) and guider field-of-view built in.

The exact list of options for this program may grow with time. Target selection can be extended to include, for example, selecting specific slit lengths based on magnitude, etc.

Mapping to Focal Plane

This process takes a target list and mask position/orientation, and produces a "map file" using precise astrometric mappings. In addition, slit lengths will be extended to use non-assigned regions of the mask, or shortened to avoid overlap. Slit length adjustments will be reported. The output map file will generally be sent to CARA for mask fabrication.

Inputs are:

Output is a "map file" which contains a "complete" description of the slitmask. The program will have the mask outline built in, and access to current maps of telescope and collimator/camera distortion.

Generation of Milling Instructions

This program has minimal inputs from the user (the map file, tool diameter). It generates the specific instructions (in AutoCad DXF format) required by the CNC milling device. Generally this program will be run by CARA personel, but it is provided as part of the general software for users who wish to have masks milled elsewhere.


Software will be required to actually drive the milling machine. Such programs are commercially available.

Software will be needed to precess coordinates and apply proper motion to the epoch of observation, outputting a format acceptable to the mask-design program.

Software will be provided to generate a graphical illustration of the slitmask using the map file. Such illustrations will include target locations (science targets, alignment and guide stars) as well as slits. Both single-page and truescale formats will be available.

Astrometry software will be provided, using the distortion maps and known astrometric reference stars (if any) to provide celestial coordinates of objects on DEIMOS direct images. While errors in the absolute coordinates may be rather large, the relative coordinates will be sufficiently accurate to allow the construction of target lists from DEIMOS direct images alone.

Calibrations and Internal Data Sets

The physical outline of the slitmask is required for both the mask design and mapping programs. In addition, the outline and relevant position of the guider field is required for the mask design program, so that suitable guide stars may be assured to fall in good locations for the guider.

Distortion maps for the telescope (at the mask surface) and the spectrograph (mask--to--detector) will be required by the mapping program, and perhaps the mask design program. Telescope distortion, ie, mapping the celestial sphere onto the mask, will be derived analytically, with empirical corrections obtained from astrometry when the instrument is placed on the telescope.

Spectrograph distortions can be obtained (at any time) by imaging a special "grid-of-holes" mask. These distortions are needed to predict positions on the CCD array for the purpose of slitmask alignment.

3.6 Existing Software and Tools

Software for mask design, mapping and mask generation have all been developed at UCSC for LRIS. It should be relatively easy to adapt this software for DEIMOS.

3.7 Other Resources Required

Commercial software for controlling the milling machine will be required.

3.8 Dependencies on Other Components

The mapping process requires access to current distortion maps in the database.

3.9 Outstanding Issues

Nothing major. We need to settle on the exact nature of "slits", how to describe them and how they will be machined. For example, can we assume all slits are parallelograms, or should we specifically allow for arcs and circles (which can present significant difficulties at the data-analysis stage)? Also, how the slit-widths are entered into the description is TBD. There are no technical difficulties here, however.

3.10 Miscellaneous

Steps to Mill a Slitmask and Enter in the Database/Library

All blank slitmask stock will be labelled with a bar code (at UCO/Lick); the barcode and a description of the stock (eg, thickness, material, surface finish, etc.) are entered into an inventory table in the database. Normal operation: observer sends map files to Keck.
  1. Map file information, describing a mask design, is entered in database.
  2. Mask design is selected by mill operator.
  3. Stock is mounted and milled.
  4. After milling, slitmask is inspected, checked against an illustration, barcode-scanned, and a quality (eg. good/reject) is assigned. If the milling was unsuccessful, the previous step is repeated. Acceptable quality means that a mask design in the database is now identified with a physical slitmask.
  5. Before each run, the Instrument Specialist calls up the requested mask designs from the database and retrieves the slitmasks for loading in the spectrograph. At this point, reconstituted copies of the map files are placed in the observer's account.

Outside fabrication: Some users may want to manufacture the slitmasks elsewhere. In this case, bar-coded stock is sent to user, and the user arrives at Keck with map files and milled slitmasks. The steps are identical to normal operations except that no milling takes place:

  1. Map file information is entered in database.
  2. The mask design is selected from the database. The slitmask barcode is scanned, and the mask is verified against the illustration (quality "foreign" is assigned). The mask design in the database is now identified with a physical slitmask.
  3. For the run, the Instrument Specialist calls up the requested mask designs from the database and retrieves the slitmasks for loading in the spectrograph. At this point, reconstituted copies of the map files are placed in the observer's account.

Slitmask Handling

A typical DEIMOS observing run may have up to 100 slitmasks or more. There must be a temporary storage facility for the masks, designed to insure against damaging the masks, and for which the bar code labels are easily accessible. Some sort of carousel has been suggested. [Long-term storage is TBD.]

There must also be a table large enough to lay out 20 slitmask frames so that the slitmasks can be mounted.

There must be a cart with sufficient storage space for 20 mounted masks, for transporting the masks to and from the telescope. The cart should actually be designed with extra storage slots, so that old masks can be unloaded and new ones loaded, while preserving some kind of physical ordering.

There must be 4 sets of 10 frames (plus extras) -- these should be color-coded, to make it easy to distinguish barrel 1 vs. barrel 2, and old vs. new masks.

The standard mask handling procedure is this:

3.11 Astrometric Considerations:

See DEIMOS image WCS issues and FITS keywords
Andrew C. Phillips / Lick Observatory
Last modified: 12 Mar 96