Summary
If you are taking a short (typically less than 60 seconds)
science exposure, you will notice that FCS may stay in
the Lockout mode for the duration of the science
integration. This is normal and your science exposure should be
ok as soon as FCS was tracking before starting the
integration. The reason why FCS stays in the Lockout mode
in short science integrations is that FCS automatically stops
tracking 30 seconds before the readout of the science exposure
starts, while each FCS tracking iteration is approximately 20 to
30 seconds long. Therefore, if the science exposure is too
short, the signal for FCS to stop tracking may arrive before the
first tracking iteration is completed and the FCS mode will
be Lockout.
The FCS image as displayed in the default orientation of FIGDISP
is 200 rows × 1200 columns; however, the image is a mosaic
of the images read out from two CCDs which lie of opposite sides
of the science mosaic. For the left-hand detector (FCS CCD1),
movement along CCD rows (+X) corresponds to motion towards the
red end of the specrum (+Y on the science CCD display), and
motion on along columns (+Y) corresponds to "right" on the slit
(+X) on the science CCD display.
On the right-hand FCS detector (FCS CCD2) FCS +X corresponds
to science +Y, and the +Y direction on FCS is the -X direction
on the science mosaic. The following figure may clarify the
default orientations of these images as displayed on the
respective FCS and science array image display tools, FIGDISP
and ds9.
Orientation of FCS and science mosaic displays. The
(x,y) axis pairs depict the row and column directions of the
three detectors as displayed on FIGDISP (for FCS) and ds9 (for
the science detector) when the image display tools are set to
use their default orientations. The numbers denote the CCD
device numbers for the DEIMOS science CCD mosaic. A sample
spectrum shows the red and blue ends of the mosaic.
The current version of the fcstrack program depends
on some additional files which contain information about the
tent mirror "gains" (i.e., the amount of image motion resulting
from a given change in the tent mirror angle) as a function of
the grating tilt and grating dispersion. The need for (and
hence the dependency on) these files will disappear once we have
fully characterized this relationship for all gratings at all
tilts.
In the meantime, those combinations of gratings and tilts
for which a tent mirror gain calibration has been performed have a
"gain" file that currently resides in the
/local/u/kics/scripts directory on polo. These
gain files have names which correspond to the instrument
configuration to which they apply. Note that since the
filter has no impact on the tent mirror gain, the filter
name is not a part of the gain file name. Examples of
these file names are:
gain.1200G.slider3.at.7800.013672
gain.600ZD.slider4.at.7400.024902
In addition, there is also a default "gain" file for each
grating. These have names of the form:
gain.1200G
gain.900ZD
The default gain file for a grating will be used in the case
where a gain file matching the current configuration cannot
be located. The default files will contain an "default"
tent mirror gain for the given grating. This will enable
the FCS control loop to make a best-faith effort at
correcting flexure, although the result will not be as
effective as it would have been had the precise tent mirror
gain for the current grating tilt and dispersion been
available via a matching gain file.
While the dewar focus value has some effect on the intensity
(and perhaps also the position) of the FCS spots in an FCS
exposure, it is only loosely considered part of the instrument
configuration. In the current version of the FCS software, the
focus value is not used when constructing the file names for the
reference and gain files. Thus, the focus value is not used
when trying to match the current instrument configuration to its
corresponding reference and gain files. It is assumed that in
general there will be only one optimal focus for a given
instrument configuration.
However, when the fcsref program records the configuration of
the instrument into a reference file, its does record the
focus setting that was in effect when the reference image was
obtained. Later, when the fcstrack program is run, if the
current setting of the focus does not match that recorded in
the reference file for the current configuration, a warning
messages is written to the fcstrack program's standard output
and to the FCSMSG keyword (which is displayed on the fcsmon
GUI). Note: a warning is only issued in the case
where the current focus setting differs from the reference
focus setting by more than 100 focus units (i.e., motor
encoder counts), since such a small change in the focus
setting is barely perceptible.
It is possible to re-use the reference files and reference
images obtained on a previous night. To do so, simply copy them
from the "fcs" directory for the previous night to the "fcs"
directory for the current night. (Note: this is done
for you automatically by the newday command if the old data
directory has the same account name.) For example, if one
wanted to re-use the Dec. 25 reference files on Dec. 26, one
would use:
cp /sdata1003/deimos3/2002dec25/*.ref /sdata1003/deimos3/2002dec26
However, in such cases it is vital that any of
the reference images that are referred to by these reference
files are retained on polo's disks in the directory
locations in which they were originally written. If those
reference images were deleted, then new reference images will
need to be obtained.
Alternatively, longer-lived/archival reference images and
reference files can be obtained by having the support
astronomer copy both the reference image and its corresponding
reference file into the directory:
/home/dmoseng/fcs_reference
In that case, the copy of the reference file will need to be
edited so that line containing the directory name is changed to
/home/dmoseng/fcs_reference. Note: this is
accomplished automatically if you employ the
save_fcsref command to copy the reference file and
image into the archive directory and modify the reference file
appropriately.
In order to support this capability, fcstrack has been
modified to search for reference files by looking first in the
current FCS directory (i.e., the one pointed to by the FCS
keyword OUTDIR), and if a matching reference file
is not found there, it will search in the
/home/dmoseng/fcs_reference directory. It is hoped
that over time, the /home/dmoseng/fcs_reference
directory will become a semi-archival repository of fcs
reference images and reference files.
The statistics logfile is a simple ASCII text file, containing
one line for each FCS correction cycle. You can monitor this
file in real-time using the tail -f command. For
example, if the current instrument configuration is 900ZD
grating in slider 4 at a tilt of 8500.052734 and with an OG550
filter, then the corresponding reference and logfile names would
be:
- Reference file:
fcsref.900ZD.slider4.at.8500.052734.OG550.ref
- Log file:
fcsref.900ZD.slider4.at.8500.052734.OG550.log
To monitor the contents of the logfile from an xterm in which
the current directory was set to the "fcs" directory (e.g.,
/sdata1003/deimos3/2002dec25), one would use the
command:
tail -f fcsref.900ZD.slider4.at.8500.052734.OG550.log
Each line of the logfile begins with a time stamp, followed by
28 fields of data, which are delimited by spaces. Note that each
line of the file is longer than 80 columns, so you will need an
extra wide xterm to view the file if you don't want the lines to
wrap around on your screen.
A typical line from a log file would look like this (note: I've folded
the line by inserting some line breaks for this example so that it will
print out on a standard 80 column wide printer page.):
Mon Dec 23 09:03:34 HST 2002 1595 59.99943924 -17.706856 -2.248045
4611 -848 -26 0.0 0.0 -17 -18 -17.50 -2 -2 -2.00
-17.486109 -17.927603 -17.706856 -2.357826 -2.138264 -2.248045
-3574 5447 -1082 2 0 OG550 -6201
The contents of each field (following the time stamp) is as follows:
- The frame number of the most recent FCS exposure
(this corresponds to deifcs keyword FRAMENO)
- The current physical position angle of DEIMOS, in degrees
(this corresponds to the deirot keyword ROTATVAL, and
can range from -420 to +310 degrees.)
- The computed flexure correction in the FCS X axis (DEIMOS
Y) (this corresponds to the deiccd keyword
FCSCORXM). Typical values should range from
between -20 to +20 pixels. FCSCORXM is the X
correction that results from averaging the cross- correlations
from the two FCS CCDs. (In non-engineering modes, this should
be the same value as in field 18).
- The computed flexure correction in the FCS Y axis (DEIMOS
X) (this corresponds to the deiccd keyword
FCSCORYM). Typical values should range from
between -10 to +10 pixels. FCSCORYM is the Y
correction that results from averaging the cross- correlations
from the two FCS CCDs. (In non-engineering modes, this should
be the same value as in field 21).
- The computed value for the tent mirror raw position that
(in conjunction with the grating tilt offset value in field 7)
should result in a correction of FCSCORXM pixels,
if fcstrack commands the tent mirror to that
position (i.e., if fcsmode >= 2). The tent
mirror raw position values range from 0 to 9418.
- The computed value for the dewar X translation stage that
should result in a correction of FCSCORYM pixels,
if fcstrack commands the translation stage to that
position (i.e., if fcsmode >= 2). The dewar X
translation position values range from -2250 to 750
- The computed value for the grating tilt offset that (in
conjunction with the tent mirror value in field 5) should
result in a correction of FCSCORXM pixels (if
applied). Typical values should range from about -70 to +70
Gurley encoder counts (where each count is 1.44 arsec of
tilt).
- For non-engineering modes, this is always 0.0
- For non-engineering modes, this is always 0.0
- The integral-pixel shift in FCS X (or DEIMOS Y) as derived
from FCS CCD 1.
- The integral-pixel shift in FCS X (or DEIMOS Y) as derived
from FCS CCD 2.
- The average of the integral-pixel shifts in FCS X (or
DEIMOS Y) as derived from FCS CCDs 1 and 2.
- The integral-pixel shift in FCS Y (or DEIMOS X) as derived
from FCS CCD 1.
- The integral-pixel shift in FCS Y (or DEIMOS X) as derived
from FCS CCD 2.
- The average of the integral-pixel shifts in FCS Y (or
DEIMOS X) as derived from FCS CCDs 1 and 2.
- The actual pixel shift in FCS X (or DEIMOS Y) as derived
from FCS CCD 1. (Keyword FCSCORX1)
- The actual pixel shift in FCS X (or DEIMOS Y) as derived
from FCS CCD 2. (Keyword FCSCORX2)
- The average of the actual pixel shifts in FCS X (or
DEIMOS Y) as derived from FCS CCDs 1 and 2.
(Keyword FCSCORXM)
- The actual pixel shift in FCS Y (or DEIMOS X) as derived
from FCS CCD 1. (Keyword FCSCORY1)
- The actual pixel shift in FCS Y (or DEIMOS X) as derived
from FCS CCD 2. (Keyword FCSCORY2)
- The average of the actual pixel shifts in FCS Y (or
DEIMOS X) as derived from FCS CCDs 1 and 2.
(Keyword FCSCORYM)
- The raw value of the nominal grating tilt for the
currently selected slider, in units of Gurley encoder counts,
as recorded in the FITS header for the most recent FCS image
(whose frame number is given in field 1). For slider 3, this
corresponds to the deimot keyword G3TLTNOM, while
for slider 4 this corresponds to the deimot keyword
G4TLTNOM.
- The tent mirror raw position that was recorded in the
FITS header for the most recent FCS image. (This corresponds
to deimot keyword TMIRRRAW.)
- The dewar X translation stage position that was recorded
in the FITS header for the most recent FCS image. (This
corresponds to the deimot keyword DWXL8RAW.)
- The value of the deiccd keyword FCSMODE that
was in effect for this cycle of the control loop. Values that
will appear in this log file will range from 1 to 4.
- The value of the deiccd keyword EXPOSIP that
was in effect for this cycle of the FCS control loop. If the
value is 1, it indicates that this exposure cycle occurred
during a long science mosaic exposure. If the value is 0, it
indicates that the science mosaic controller was idle during
this cycle of the FCS control loop, and that the shutter was
open without a science exposure in progress.
- The value of the deimot keyword DWFILNAM that
was recorded in the FITS header for the most recent FCS image.
This is the name of the filter that was in place when that
image was taken.
- The value of the deimot keyword DWFOCRAW that
was recorded in the FITS header for the most recent FCS image.
This is the value of the focus that was in place when that
image was taken.
The current version of the fcstrack and
fcsref programs support some engineering and
compatibility modes. The compatibility mode allows
fcstrack to be used with pre-existing reference files
(provided that the reference images to which they refer are
still in their original locations on polo's disks).
The engineering mode enables one to track the individual
centroids of up to 9 FCS spots distributed in any manner between
FCS CCDs 1 and 2. For a detailed description of these modes,
see the in-line comments at the start of the fcsref
script, which can be found in the /local/kroot/bin
directory on polo.
The fcswhich command will search for and identify all
FCS reference files (looking first in the current FCS OUTDIR
directory and next in the /home/dmoseng/fcs_reference
directory) that match the current instrument configuration, and
will indicate which one takes precedence. It will also check
whether or not the FCS reference image that is named by that
reference file is still accessible by the account from which the
command is being run. It uses the same algorithm as
is now used by the fcstrack command.
This command should provide a simple means for determining
whether or not a reference file and reference image is already
available for the current configuration. If one is, then the
observer can decide whether or not they want to use it or take a
new reference image of their own.
You can also use the FCS GUI to search for and examine the
currently available FCS reference files. On the main panel
of the FCS GUI, click on the Setup
button, which will bring up the "FCS SETUP" sub-panel. Next,
click on the Browse Reference Files
button, which will bring up a browser tool that enables you
to explore the current inventory of FCS reference files.
Bob Kibrick has written a script to move FCS spots by an
indicated amount, as referenced to the coordinate system of FCS
CCD 1. The script is called fcsmove and it takes two
arguments:
fcsmove x y
where x is the desired number of pixels to move the
spot in the FCS x coordinate, and y is the
desired number of pixels to move the spot in the FCS y
coordinate (of FCS CCD 1).
See Also