Recommendations for planning observations with AF2
Observing with AF2 requires careful advance planning, and on this page
we offer advice about three critical issues:
In particular, do not underestimate how long it will take to produce the
configuration files. Start to prepare the files well in advance of the
run (e.g. a month), contact your support astronomer for advice on how
to plan the observing run and send her/him the configuration files as
soon as you can for checking them in the AF2 observing system.
wise observer would also arrange to arrive at the telescope at least
one day in advance of their observing as configurations may have to be
re-done at short notice due to unforeseen circumstances (e.g. the loss
of a fibre).
Astrometry / choice of fiducial stars
- The astrometry of the targets has to be of the highest accuracy
obtainable to minimize the offset errors in the fiber positioning.
Astrometric catalogues as UCAC3 and SDSS have proved to provide good
results. The astrometry should also account for proper motions of the
fiducial and science target stars. Even those stars with small proper
motions, e.g. 20 mas/year, might have coordinate differences of 0.3
arcsec after 13 years, given that the astrometric coordinates are
usually referred to Epoch 2000.
- The astrometry of the fiducial stars and the science targets used
in the configuration files should be from the same source. If they are
from different catalogues, there is a risk that although an
acquisition is good enough and fiducial stars are well centred in the
fiducial fibres, the relative offset in the astrometry of the science
targets can result in them not being well centred in the science
- At least 3 acquisition (fiducial) stars are required for an
aceptable acquisition, the more the better, but to be safe we suggest
more than 4. They should be well distributed over the AF2 field of
view, e.g. some fiducial stars in the centre of the field, and some in
the outer areas around 20 arcminutes from the centre. Avoid using fiducials beyond the 20 arcmin radius, as the distortion are stronger there
and might bias the field acquisition.
- The coherent fiducial bundles, allow
one to see, in the acquisition/guiding camera, stars as faint as
magnitude V~16 with a nearly full moon, good seeing (1~arcsec), and
some integration. However, you should select stars a bit brighter
than this limit (V≤15), to account for possible poor weather
conditions. Ideally the fiducial stars should all be bright enough (V
< 15) and, if possible, of similar brightness, in the range 13 < V <
- Always make sure that you have the most up-to-date versions of files
small_fibres.dat and wht_prime.dat.
- Using a buffer size of 1.5 when creating the configuration file with af2_configure is mandatory,
otherwise your setup will likely be rejected by the AF2
system computer. Even if such a configuration is accepted, the risk of the
mechanism losing a fibre rises significantly for smaller buffer sizes, leading to potential
problems which could take hours to fix, or, even worse, could cause permanent
damage to the fibres.
- Remember to set the parameter FIBRES SMALL in your field configuration
- Note that no more than 99 "spare" fiducial stars should be
contained in the output file (.cfg) ("spare" fiducials are those listed
at the end of the file and numbered starting from 901), or the setup command
given at the Instrument Control System (ICL) may fail.
- Fibres can be positioned over a field of 1 degree diameter; although there
is some vignetting at radius >20 arcmin, and image quality outside the 25 arcmin
radius is poor, the PSF showing a some elongation that causes light to be lost for
targets in this outer area. We thus recommend users to avoid putting science fibres
in the annulus outside the 25 arcmin radius, and consider there there will be some
moderate light loss due to the optics also at radii between 20 and 25 arcmin from the centre.
- Try to avoid to assign the important science targets to the bad fibres, with the lowest throughput and the worst image quality.
- It is not necessary to provide distinct configuration files
for a range of hour angles. For a given configuration, the appropriate
sidereal time is input when the field is configured at the telescope;
this overrides the hour angle specified in the configuration file.
- It can be useful to determine in advance if a rotator limit will be reached during the observations of a given fibre configuration. This can help to plan the observations during
the night, as there is a blind spot of 3 degrees (the mount PA has a range from -163.37 to 193.63 degrees), and when it is reached the rotator has to be moved back into limits, with
the subsequent loss of observing time. We can determine the time when the limit will be reached from the sky PA in the configuration file or, alternatively, we can avoid to choose a
range of sky PA for which the rotator will reach the limit in the period of time when the fibre configuration will be observed. See instructions on how to determine the rotator limits
- If many different fibre configurations are going to be observed
during a night, it is recommended to take sky twilight flats in circle
configuration, in the evening or morning, hence they can be used with
all the configurations when reducing the data. Circle configuration
flats are also useful to calculate the relative fibre througput.
- Sky substraction is a tricky issue and should be addresed in
different ways depending on the science program.
bright objects, the sky substraction should be good enough if just
assigning some 'sky' fibres next to the target fibres and then
calculating a median sky spectra to be substracted.
targets, the sky substraction need to be more accurate, hence the best
strategy is to move the telescope back and forth between objects and
neighboring sky (beam-switch), so that the sky spectra are measured on
the same fibres as the science objects.
The spectral range of interest is also very important when
deciding the best strategy. At the very red end, the sky emission
lines are very strong and also vary very quickly, hence the
'beam-switch' strategy might not be the best option, and the sky
should be measured simultaneouly with the science spectra.
- AF2 suffers
ghosting caused by undispersed light. Its intensity depends mainly
on the source brightness and the exposure time. The recommendation to
keep a low level of ghosting is to avoid mixing in a same fibre
configuration objects with a wide range of magnitudes, but create
e.g. a configuration file for bright targets and another one for faint
- It is recommended not to observe fields at very low elevation, as
the fibres would need frequent re-configurations to account for the
differential refraction in the field. It takes around 20 min each
time, which would increase substantially the overheads.