ALFOSC MOS: Guidelines for making a new distortion map
For the design of new MOS masks, the MOSPLATE software uses a
transformation matrix to correct the image distortion between mask
(telescope focal plane) and the detector (ALFOSC focal plane).
When optics or the detector in ALFOSC are changed, a new distortion
map should be made. In order to make the distortion map one needs a
rectangular/square grid of pinholes with 2.000mm spacing, made in an
ordinary MOS plate. This grid plate should be made with the same CNC
machine as that on which the science MOS plates will be cut.
The distortion maps are stored as IDL transformation matrices in files
such as
sandra:/home/mos/mosplate/coord_trfm/trfm/NAME_OF_CCD.trfm or
${MOS_DIR}/etc/NAME_OF_CCD.trfm
so you can easily swap between transformations.
Note that you have to update ${MOS_DIR}/etc/mosplate.par to make the
software point at the
correct transformation file.
Notes for MOSPLATE users are found
below (at the bottom)
How does NOT staff make a new distortion transformation matrix?
To make a new transformation
you have to do the following steps:
- Put in the grid mask, take out any filters inside ALFOSC,
and make 3 images of the grid mask with dome lamps.
- Run DAOFIND or STARFIND on the images of the grid mask.
(See below for details.)
- Copy these files to sandra:${MOS_DIR}/coord_trfm/input_coords,
calling it something sensible like 'ALFOSC_NEW.coo'.
Any empty lines in the file should be removed.
- Before running PIXEL2MM_FIT, update file
sandra:${MOS_DIR}/etc/pixel2mm_fit.par
to correctly reflect the new instrumentation and grid pitch.
- Run the IDL programme PIXEL2MM_FIT, located in the directory
${MOS_DIR}/coord_trfm. Running this programme may require a bit of
help from an experienced IDL user. There are some prompting for info
which is commented out and substituted for hard-coded values first in
the programme. You would like to hard-code something else or
uncomment the prompting. But otherwise everything should be straight
forward. (See below for details.)
This creates a transformation matrix .trfm file in directory
sandra:${MOS_DIR}/coord_trfm/trfm/
Note: JHT runs PIXEL2MMJHT_FIT
because it does not expect all gridpoints to be detected, and hence is
less strict. (It now also contains a vital BUGFIX.)
- Verify the orientation of the transformation. If the CCD is read
out in the same way as the previous one (north and east orientation
the same as for the previous CCD, using the same sky position angle as
before) it should be OK.
- Before running MOSPLATE, update file
sandra:${MOS_DIR}/etc/mosplate.par
to correctly reflect the new instrumentation, and the whereabouts of
the new transformation matrix.
- Let the users know that they have to update their MOSPLATE setup!!
About IRAF STARFIND
> images
> imcoords
> starfind jht00009 default 1.3 1000
> lpar starfind
image = "jht00009" Input image
output = "default" Output star list
hwhmpsf = 1.3 HWHM of the PSF in pixels
threshold = 1000. Detection threshold in ADU
(datamin = INDEF) Minimum good data value in ADU
(datamax = INDEF) Maximum good data value in ADU
(fradius = 2.5) Fitting radius in HWHM
(sepmin = 5.) Minimum separation in HWHM
(npixmin = 5) Minimum number of good pixels above background
(maglo = INDEF) Lower magnitude limit
(maghi = INDEF) Upper magnitude limit
(roundlo = 0.) Lower ellipticity limit
(roundhi = 0.3) Upper ellipticity limit
(sharplo = 0.2) Lower sharpness limit
(sharphi = 5.) Upper sharpness limit
(wcs = "") World coordinate system (logical,physical,world)
(wxformat = "") The x axis world coordinate format
(wyformat = "") The y axis world coordinate format
(boundary = "nearest") Boundary extension (nearest,constant,reflect,wrap)
(constant = 0.) Constant for constant boundary extension
(nxblock = INDEF) X dimension of working block size in pixels
(nyblock = 256) Y dimension of working block size in pixels
(verbose = no) Print messages about the progress of the task
The above creates a similar output as daofind, but is smart enough not
to include any cosmics. However, not all pinholes are detected, but
not all should be detected because some are clearly not very good.
Tune the 'hwhmpsf' parameter to get best results. Make sure to
delete the empty lines from the output .obj file! Check the result on
your image display with:
> tvmark radius=15 color=203
About running PIXEL2MM_FIT in IDL
On sandra as user mos:
- Edit sandra:${MOS_DIR}/etc/pixel2mm_fit.par
to correctly reflect the new detector size and the grid pitch.
- > cd mosplate/coord_trfm/
- start IDL:
> idl
- IDL> .run pixel2mm_fitJHT
- IDL> pixel2mm_fit
answer CCD-NAME (e.g CCD8) and then FILENAME (e.g. jht09.obj)
- Check the output with:
IDL> restore, "trfm/CCD8.trfm"
IDL> print, TRFM_COEF
About running MOSPLATE with the new transformation matrix
- Copy the new file transformation matrix to the directory
${MOS_DIR}/etc/
- Update file
${MOS_DIR}/etc/mosplate.par
to correctly reflect the new instrumentation, and the whereabouts of the
new transformation matrix.
- Follow the MOS User Manual to
design new science masks.
I am a MOSPLATE user: what should I do?
- Get the latest distortion transformation matrix .trfm file.
- Copy the new file transformation matrix to the directory
${MOS_DIR}/etc/
- Update file
${MOS_DIR}/etc/mosplate.par
to correctly reflect the new instrumentation, and the whereabouts of the
new transformation matrix. The current version of mosplate.par at NOT looks like this.
- Follow the MOS User Manual to
design new science masks.
|