Guide to Creating a Configure Input File
For a relatively uniform target distribution and 500-800 targets, the time required for a configuration is ~20mins on a single CPU machine. For a similar field, with 1,500 targets, or for a centrally-condensed field such as a galaxy cluster, the configure time could be 1-2hours! Configure versions 7.9 onwards are multi-threaded and so multi-CPU machines will run much faster, reducing this problem. However, the user will generally be required to perform some kind of sparse pre-sampling of the dataset in order to have the ~500-800 targets per 2 degree field to run Configure in reasonable times. Possible options for pre-sampling are described in the Guide to complex configurations.
Configure reads in ascii .fld files. An example .fld file is available to illustrate the basic format example.fld. In constructing your own .fld files the following considerations need to be made:
Science targets: No more than 800 targets and these should cover a relatively small range in target magnitude (less than 3 mags is the standard constraint, but talk to your support astronomer if you require more detail here).
Calibration sources: If required, these should be set to Priority 9 in the .fld file with the priority of all science targets shuffled to lower levels so that the calibrators are always allocated.
Sky fibre positions: You will need 20-30 sky fibres in the observation, so 50-100 possible sky positions should be enough.Eyeball the sky fibre positions to check they are actually blank regions.
Guide stars (Fiducials): These are crucial to the success of your observations so pay careful attention here.
AAOmega has 8 guide bundles available. You should aim to use at least 4-6 and preferable all 8. This will require 20-30 candidate guide stars distributed across the field plate to prevent guide star selection compromising science fibre placement.
In dark sky conditions, guide stars should be 14-14.5mag (B,V or R mags on the Vega or AB system). Up to 12th magnitude stars may be needed during bright-of-moon, (and check the location of the moon relative to your fields).
Guide stars MUST be on the same astrometric system as your targets.
Guide star spatial distribution must match that of the stars.
The range of guide star magnitudes should be small, preferably less than a 0.5 mag spread. A greater spread will cause problems due to the dynamic range of the camera used to guide.
GUIDE STAR WARNINGS:
Simply selecting some bright guide stars from SIMBAD or GSC is NOT going to work, your astrometric solution MUST be the same for the guide stars AND the targets, and good to 0.3arcsec or better. This is a requirement for AAOmega observations.
UCAC-2 and 2MASS sources have proved successful in recent years, although the USNO survey seems to be somewhat inconsistent (probably due to plate boundary effects).
SDSS is an obvious source of guide stars. However, all stars need to be eyeballed as SDSS has funny artifacts at the magnitudes required here. Marginally saturated stars, which do not suffer obvious defects on examination, have been found to still give excellent results with AAOmega (the SDSS astrometic data for these objects actually comes from smaller edge CCDs so the stars do not actually saturate in their astrometric reference frame).
Eyeball your guide stars. Reject galaxies, reject binaries, reject objects with junk magnitudes. Stars should NOT be used blindly (guide globular clusters are next to useless and stars should not have spiral arms).
The target and guide star astrometry MUST be on the same system. Simply using two different catalogues that independently claim to be J2000 will result in poor acquisition and low target sensitivity.
We have had some success recently in including a small number (1-2 objects per configuration) of standard star calibrators in each AAOmega field. These must be chosen to be faint, to avoid contaminating science spectra. Drawing the calibrators from the recent sample of White Dwarfs and Hot Sub-Dwarfs of Eisenstein et al. ApJS, 2006, 167, 40 from SDSS has worked well. Absolute flux calibration is not possible with a fibre system such as 2dF/AAOmega, due to the unquantifiable aperture losses in any given observation, but including a standard star in each field plate observation can improve the quality of internal spectral calibration, and monitor data quality during a run. All caveats relating to astrometric accuracy apply to calibrator data as well as science and guide data.
An interesting paper on the effects of poor astrometry on Signal-to-Noise is Newman P.R. 2002 PASP 114 918
In the field header of the .fld file, specify a set of wavelengths of the form:
where n is an integer in the range 1-9 and wavelength is a wavelength specification in Angstroms.
For each program object, P, you may now append a wavelength selector of the form P_wn where n is in the range 1 to 9.
Once this is configured the wavelength specified in the 2dF positioner GUI by the support astronomer becomes the default wavelength (used when a specific wavelength is not specified and for Sky objects). All fibres with specific wavelengths will be positioned optimally for that wavelength and the observed .fits files have a "WLEN" field in the binary table associated with each object to indicate the actual wavelength the object was configured for.
Additionally see Chapters 4 and 5 of the 2dF manual available from the manuals page for instructions on 2dF astrometry and how to choose fiducial stars.
The telescope's Positioner GUI also handles atmospheric refraction effects when working out the positions of fibres on the field plate - including the effects caused by different observation wavelengths. Normally a single wavelength is chosen for all fibres and is applied by the support astronomer as described in this summary of the observational wavelength issue. However, it is possible that you may prefer to have fibres configured for different wavelengths. It is now possible to specify up to 9 different wavelengths in the .fld file (also shown in the example.fld file):
The SA algorithm is very good at allocating targets based on the 9 possible priority levels (9 is highest priority). However, the user should exercise some restraint when using the available levels. Using all of the available priorities to derive a complex priority selection function will almost always yield very limited returns at the expense of usability.
For most programs the number of targets in a given .fld file must be restricted (as described above) in order to allow the configuration process to be completed in an appropriate amount of time (~20mins). A field that is stacked with a large number of low priority targets will take a long time to configure. If these targets are indeed low priority then the user should consider carefully whether their inclusion in the .fld file is worth the overhead in configuration time they will incur.
Sarah Brough (email@example.com)