SA Notes
CCD WINDOW
Use
> window wyffos 1 "[1:4300,1:4200]"
to get the two CCDs mosaiced together
CCD ROTATION
If the spectrum appears to be rotated anti-clockwise, than you have to rotate
the cryostat COUNTERCLOCKWISE by approximately two mm for every 100 pixels
of rotation in the spectrum.
CCD FIBRE SEPARATION is about 25 unbinned pixel
FIBRE NUMBERING
Fibres run from 1 to 150 going from RIGHT TO LEFT in the display window.
To identify fibres, the fibre number can be overplotted using the ds9 region
files
/home/whtobs/AF2/ds9Reg/af2bin1.reg (no binning)
/home/whtobs/AF2/ds9Reg/af2bin2.reg (binning 2x2)
In ds9 select "Region/Load Regions...", then "Region/Select All" to overplot
the labels. You can move the labels left/right using the <- or -> arrow keys.
VIGNETTING
On the top of the CCD (red side), ~150-180 pixels suffer from strong
vignetting to complete obscuration by the relay mirror. Useful CCD pixels
are then Y<~3950 (unbinned) for reflection mode.
SPECTRAL RANGE, DISPERSION AND RESOLUTION
With AF2, BLUE is DOWN and RED is UP.
There is an overall curvature, with wavelength shifting to redder
wavelengths in the central fibres, in addition to the expected
wavelength shift within every trio of fibres (see sketch below)
<-- fib. 150 ------- ------------ fib. 1 --->
. . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . .
. . .
^
| RED
|
| BLUE
v
In reflection mode, the maximum distance (upper fibre of every trio on
the left/right side, compared to the lowest fibre in the centre) is
about 215 pixels, which makes a maximum wavelength shift of some 5% of
the total spectral range. The wavelength shifts between two consecutive
fibres in each trio is 60 pixels.
The following was measured on 20041003 (FWHM in unbinned pixels is from
arc lines)
Grating disp. Total FWHM FWHM R=lambda/FWHM Best
(ang/pix) range (pix) (ang) focus
R1200R 0.38 1490 3-4 1.1-1.5 6200-4500 (6800ang) -3800
R600B 0.81 3180 4 3.2 1500 (5000ang) -2900
WHEN MEASURING THE FWHM OF SKY LINES IN SCIENCE SKY SPECTRA, HOWEVER, THE
FWHM IS OF 5-5.5 PIXELS, I.E. SOME 50% HIGHER, CLOSER TO THE EXPECTED
RESOLUTION COMPUTED BY EXTRAPOLATING THE FIGURES FOR THE WYFFOS SHORT
CAMERA. THIS IS DUE TO THE DIFFERENT ILLUMINATION BEAM FROM THE SKY
AND FROM THE CALIBRATION LAMP UNIT.
CENTRAL WAVELENGTH
On 20040707, it was found that a Grating Offset Value of 6850 would
provide the required central wavelength in an intermediate fibre, so
that some fibres are shifted up to 2.5% toward the red, and the others
up 2.5% to the blue. But check it every time you setup a grating.
TRANSLATION OF FIBRES
Fibres can be translated so as to lose only one fibre in the CCD gap,
#83 which is currently disabled, if the rotation is precisely set (left-right
offset between top and bottom) of less than 5 unbinned pixels. However,
two fibres will then be at the edge of each CCD just out of the butted
region, and so if e.g. relative flux calibration is important, it is
generally better to lose two fibres (either #82 or #84 in addition to #83).
This is the normal setup for AF2/WYFFOS.
Increasing the slit translation value moves the spectrum to right: 1000
units corresponds roughly to the fibre separation (25 pixels).
SPECTROGRAPH TILT/FOCUS
There is still not a clear procedure for how to setup the CCD tilt and
focus. The aim is to use the Hartmann shutters, but for the time being
we just move the capstans and slit focus. Remember that the two CCDs are not
coplanar, so it is virtually impossible to have the same focus everywhere
over the entire mosaic. You should however be able to get a FWHM around
2 binned pixels for all 4 corners and the centre.
One hint:
- if the UPPER-LEFT AND LOWER-RIGHT corners are out of focus, this can
be corrected by tilting capstain A (the one below the cryostat) and
then change overall slit focus (if capstain A is turned clockwise,
decrease focus value).
Capstans:
C
B
A
TASKS TO BE DONE ON SETUP NIGHTS
There are several tasks which must be done following each instrument change
to AF2/WYFFOS. Some of these are done by the support astronomer, and others
are done by the OSA, but it is ultimately the responsibility of the
support astronomer to ensure these essential setup tasks are completed
following the instrument change.
In the afternoon (by the support astronomer):
- Put fibres in a circle (af2 allcircle)
- Setup spectrograph: insert grating with blaze arrow directed towards the
WYFFOS door (always ensure the grating cell is empty when moving from
echelle mode to reflection mode to avoid damage to the grating), insert
blocking and order-sorting filters if required, check rotation,
translation, tilt/focus, slit-unit translation, and check the actual
central wavelength and spectral range regisered on the detector
At night on-sky (by the OSA and support astronomer as specified below):
- Take a sky flat with all fibres in a circle, to measure the relative
throughput of fibres. This will be done by the support astronomer.
- Setup telescope: rotator centre offset, pointing, offset between
gripper centre and mobile probe centre, focus telescope with star
in a fiducial bundle. These tasks will be done by the OSA, with input
from, and communication with, the support astronomer. It is crucial
that these on-sky setup tasks are done on-sky on the first setup
night (weather permitting), since collectively they can take an
hour of sky time, occasonally even more. If it's not possible to
complete these tasks due to e.g. bad weather, then they will have
to be done on the next, clear night (usually a science run, possibly
with no support astronomer present).
- Using an astrometric field (UCAC3, APM, Sloan) with stars acquired in
all fiducial fibres well spread around the FoV, check that they are
well centred. This is a fast check to make sure that prime focus
distortion coefficients are correct. Acquisition of the astrometric
field will be done by the OSA, and acceptance of the distortion
coefficients is the responsibility of the support astronomer.
- Use the astrometric field to focus the telescope; the best focus is a
compromise between the centre of the field and approximately 20-arcminutes
distant from the field centre (there is considerable image aberration
beyond 20-arcminutes from the field centre, and image size can be up
to 4-arcseconds at the edge of the field). The compromise focus
is close to 82.45mm. The image of the fiducial stars in the inner parts
of the field should be circular; if they are not, adjust the focus and
if this doesn't improve image quality, check that the ADC is running.
Acceptance of the focus value is the responsibility of the support
astronomer.
- Throughput measurement: make throughput measurements as advised by
the instrument specialist. This will be done by the support astronomer.
- Every 1-2 years (or if a problem is found), do a "mapping" of the
prime focus distortions. This will be requested (and probably done
by) the AF2/WYFFOS specialist.
UPDATING FILES IN THE WEB AND AF2_CONFIGURE SOFTWARE
The main html documentation is in
/docs/Astronomy/instruments/af2/
Each time that the small_fibres.dat file is changed (e.g. when a fibre
is disabled), or the wht_prime.dat file is changed (e.g. after a new mapping
field analysis), the new version should be copied from /wht/var/ to
/docs/Astronomy/instruments/af2/
The small_fibres.dat and wht_prime.dat files in the local installation
of the af2_configure package are symbolically linked to the versions in
/docs/Astronomy/instruments/af2/, which therefore should have ugo read
permissions.
Configurations files used to physically configure AF2 must have been
created with the current versions of small_fibres.dat and wht_prime.dat
files.
MAPPING
- HOW TO CREATE A MAPPING FIELD
1) select a field [input.dat] with good astrometry for relatively bright
(12-16 mag) stars. Use UCAC3, APM or Sloan.
2) create .fld file [Ainput.fld] with any SKYPA and all the above stars
defined as fiducials.
3) run af2_configure with or without p.a. scanning, park with the
cursor all fiducials, except for #112 OR #128 which should be put on a
fiducial star near the edge of the f.o.v. of AF2. Then accept max 99
'spare fiducials', that should be well distributed in the
FoV except for the vicinity of the used fiducial star 112 or 128.
(expect some iterations if you have many stars, in order to create a
nice grid).
- HOW TO RUN A MAPPING FIELD
SYS> af2 setup ....
Run the setup created as above with telescope in position (if known,
add the typical rotator mount offset for that run).
- Start autoguiding on fibre #112 or #128.
- Switch off the mobile probe autoguider.
SYS> af2 map
Enter autoguider name (ag2), exposure time for each star (e.g. 5 sec),
n. of integrations per star (3), and the approximate x y position of
stars in the autoguider, and the proposed value for the autoguider
rotation in AF2.
- HOW TO ANALYSE A MAPPING FIELD
> cp ASTROM_AF2.DAT astrom.dat
and then run starlink task
> astrom
This puts the results in astrom.lis.
In this file, you need to check 7 coefficients results (it does not
give the 9 coefficients result as in the AF2 manual, but we think to
understand from starlink manual that the `astrom' program chooses
between 7 and 9 coeff. solutions depending on the number of stars).
In the 7-coeff solution, you will find output lines like the following:
----------------------------------------------------------------------
1Plate solution: 7-coefficient
-----------------------------
Radial distortion has changed by -20.34 to +336.75 (std dev 1.74)
Plate scales (in measuring units): X 17.6420 arcsec
Y 17.6439 arcsec
mean 17.6430 arcsec
Nonperpendicularity: +0.008 deg
Orientation: +0.004 deg and laterally inverted
----------------------------------------------------------------------
The values to be updated in file wht_prime.dat from the example above
(June 2005 mapping) are:
17.643 --> Linear plate scale component (arc sec/mm) from af2 map
336.75 --> 2nd order term of plate scale from af2 map