An important part of maintainability is feedback to me
on this document, and the performance of ROKCAM.
This web site complements and extends the ROKCAM User Manual by Greg Doppmann
and Dan Lester by providing specific at-the-telescope instructions and resource
links. That latter document deals mainly with system characterization, and
is recommended as an auxiliary reference. The 47 page MS Word document is
available for downloading by anonymous ftp from marple.as.utexas.edu:~ftp/outgoing/rokcam_manual.hqx
and is also available in hardcopy form at the telescope. Note that while
it is reasonably current, it is not actively maintained, in favor of this
web page.
You can find a summary of the characteristics of ROKCAM , including filter
profiles, here, and some brief details about
the NICMOS array here. If ROKCAM is to be used
as an imager for CoolSpec,
the user should consult that web site for details about the operation of
the spectrograph.
If ROKCAM was requested on the rfs form, it can be assumed that the hardware
will be set up by the Scientific Support staff at the observatory. The observer
is responsible for software setup, calibration, and observing. This page
is organized roughly as an observers checklist for routine ROKCAM imaging
that can be followed on the day of the first observing night. You can click
on any topic to get more information.
Note that if all objects to be imaged are bright enough to be recognizable
in the ROKCAM field, and if no active guiding will be required, it won't
be necessary to set up the PXL guider camera. It's often not possible to
know this in advance, however ...
Software and Computer Setup
Initializing
the Instrument Control Software
Inititalizing
Snapshot
Inititalizing
the ATOG
Initializing
the PXL Guider Camera
Initializing
the Telescope Control System
Final Afternoon Verification and Setup
Verifying
ROKCAM
Presetting
Focus
Verifying
Tub Rotation
Topping
off the Dewar
Alignment Checks and First Light (even if it's partly
cloudy!)
Boresighting
ROKCAM to TCS
Boresighting
ROKCAM to the PXL Guider
Verifying
ROKCAM Beam Alignment
Focussing
the Telescope
Determining
Direction and Scale, and Setting the Orientation
Calibration and Useful Observing Info
Standard
Stars
ROKCAM
Sensitivity
Useful
ICEX Tasks for ROKCAM
End
of Night Checklist
Software and Computer
Setup
For general information on the computer systems at Mt. Locke, and
observer responsibilities in using them, consult this
page. Everyone will have their own preferences about how to set the ROKCAM
software up. Here is one way ...
- Have the onsite staff configure an account on atlas for use with ROKCAM
Initializing the Instrument Control Software:
- use df to find an Atlas data directory (e.g. /data2/atlas)
that has plenty of room
(note that each frame takes ~140K bytes, and you can use up several hundred
Mb in a single night)
Make a data directory. Suppose we're using, for example, /data2/atlas ...
- atlas>mkdir /data2/atlas/rokcam
open two xgterm windows, one for IRAF analysis, and one for ICEX commands,
which we will identify here as atlas1 and atlas2
- atlas>xgterm & make 1st window
- atlas>xgterm & make 2nd window
- atlas1>cd IRAF
- for convenience, edit the last line of the login.cl file in this directory
to default you to the data directory that you've selected
Now pop an ximtool window for data display, and start IRAF for data display
- atlas1>ximtool &
- atlas1>cl
- atlas1>cl>cd /data2/atlas/rokcam
- atlas2>cd ICE
- for convenience, edit the last line of the login.cl file in this directory
to default you to the data directory that you've selected
Note that ICE source code, tasks, and documentation resides at /home/miranda/ice/ice-1.5
Now start ICEX in the second window
- atlas2>cl
- atlas2>cl>cd /data2/atlas/rokcam
- atlas2>cl>icex
- Use eparam (or epar) to check and edit, if necessary,
the parameter file setups required for running ROKCAM. You can find starting
default values in the following links, and you can get an explanation for
all the quantities in these and other parameter files using the help
task (e.g. help instrpars). You can also use command line editing
as in, for example, obspars.sequence=26.
detpars - default detector setup and
microcode parameters
instrpars- default ROKCAM instrument
parameters
telpars- default telescope parameters
obspars- default observing format
parameters
Now you are ready to take data. At this point, you can skip down to the
section on verifying ROKCAM, or continue with the observing software setup.
Now, here's a VERY IMPORTANT item. The filter complement in ROKCAM changes
frequently, as required to support varous needs. Of course, you've been
careful to specify your filter needs in your Request for Services (RFS)
form (and ideally well before that!). So the filters that you want should
actually have been physically installed. The definitive table of filters
installed in ROKCAM is taped to the side of the dewar. But you need to
check that the software is aware of these filters and the order in which
they are installed. This is not the responsibility of Scientific Support,
in that they don't know what ICEX account you're going to use. Edit the
fts parameter line in your instrpars accordingly. This
line should list the filter names in position order (#1 first, #8 last)
separated by commas. See the default parameter set linked above as an example.
Inititalizing Snapshot:
No big deal here, but it's nice to have a snapshot window open
to allow you to print screen images. Simply open a local window on Atlas,
and execute snapshot. Under print->options, set printer
to 107laser, if it is not already set as default. Note that full-array
ximtool images take several minutes to be printed on 107laser. Note that
in order to 'snap' an image, that image has to be on the same workstation
screen as the snapshot control window. If 107laser isn't available, use
hplaser, which is a much faster printer down on the 2nd floor.
It is worth noting also that if you're going to save ximtool images in a
file, don't do so in your home directory. These are large files, and your
home directory isn't that big. Most importantly note that IRAF gets upset
if you start to push your quota, so the way you'll find out that you're
close is when IRAF starts misbehaving. Do quota -v to check on
your status.
Inititalizing the ATOG:
If the ATOG (Automated Telescope Offset Guider) microprocessor has not
been booted, and you don't see an ATOG window, you'll need to do it now.
An ATOG manual is not currently available on-line. A hard copy is available
in the 107 Cass control room.
- use the pull-down menu in SunOS to open an atog window
- if it doesn't come up with a "-" prompt, type exit
to get one, and then start the program
-atog6u8 (note: the screen echo is funny on this, just type it
and hit return without looking at it!)
- aftger ten seconds or so, you should see the ATOG command window --
if you don't, get help
This is a complicated window, with a lot of options that you'll never use
for ROKCAM, and just a few that you will.
Set up the ATOG to the default diagonal mirror position for allowing ROKCAM
to look up past the primary out at the secondary, off the edge of the diagonal
mirror.
- -TV @ROKGUID
It is smart, at this point, to go out and check. Open the mirror cover,
bend over, and sight upwards along the side of the ROKCAM snout. You should
see the secondary mirror up at the top of the telescope through the primary
central hole. Close the mirror cover.
Initializing the PXL Guider Camera:
The setup and operation for the PXL guider is covered in the PXL User Manual. The ATOG includes a filter set that can
be used in front of the guider. For maximum guider sensitivity, be sure
that the guider is looking through a clear hole instead of a filter with
the following ATOG command line.
- -FILT OPEN
Inititalizing the Telescope Control System:
The Telescope Control System (TCS) handles pointing, tracking, and offsetting
in a user-friendly way. Consult the 2.7m
Telescope Control System Manual for instructions on how to initialize
and use this new system.
Final Afternoon Verification and Setup
Now that things are all set up, verify the performance of imager and
software.
Verifying ROKCAM:
Go to the ICEX window to take some test exposures.
- atlas2>ic>filtgohome
This homes the filter wheel. Go outside and check that the readout
on the filter position under the dewar is indeed roughly 0000 (which is
nominally the J filter). While you're out there, make sure the ROKCAM lens
cap is still on, such that ROKCAM is looking at something with emissivity
of unity.
If that's all working then go to broad band K (instrfi=@K above)
and get ready to do an integration.
Good opportunity to check that the filter really did move as commanded.
If atlas isn't talking to the CyberPak stepping motor driver, it may not
have been hooked up right. Check with Scientific Support.
Another very important item ...this is a good time to make sure that the
wire grid polarizer is in the position you want it in. If you're just doing
imaging, you want it OUT of the beam. The polarizer is removed from the
beam by turning the polarizer knob (that second knob on the bottom of the
ROKCAM dewar clockwise (while facing up from below) until it stops.
It's on a long screw, so if it needs to move, you'll need to turn that
knob many turns before it stops.
- atlas2>ic>test
The "test" mode takes frames without storing them. Select
a single exposure, and an expopsure time of 1 second. For this first integration,
you'll see some reports that the microcode is being uploaded to the ROKCAM
FDISH. In broadband K, you'll see a roughly uniform background displayed
in the ximtool display. You should get a background level of ~2000-4000
DN (see the ximtool DN display at the bottom right of the ximtool screeen),
depending on the ambient temperature.
This is a good opportunity to strategize about how the cosmetic quality
of the array will affect the science program. You can see where the bad
pixels are, and where they aren't.
- Now go to blankoff (@blank above) and do it again.
You should now get values of order 10 DN, which is essentially the dark
count. If the dark count is much higher than this, it is possible that
the dewar has not equilibrated (use a DVM to measure the on-chip temperature
using the TEMP BNC connector (Vtherm) and the posted temperature-voltage
conversion -- it should be 85K or less), or that one of the cans has run
out of cryogen.
Note that ROKCAM takes at least ten hours of cooling from first fill before
the "dark current" is small, and up until about five hours after
the first fill, when Vtherm gets to ~970mv, the dark current will saturate
the array. The dark current drops quickly after that, and by seven hours,
when Vtherm~990mv, it should be <60DN/second, well below that of the
sky background in a broad-band filter.
- Now verify that the nodding is enabled. The skysub routine
does a sky-star-star-sky set of exposures, which we call a "quad".
One quad corresponds to four stored frames. While there isn't any way to
set skysub up so it won't save to disk, you might identify what
is just playing around by setting rootname to "junk" (see below).
atlas2>ic>skysub2
As above, select an exposure time of 1 second, and enter nod offsets
of 60" in both RA and Dec directions. Watch the telescope encoders
on the handpaddle to make sure that, as the quad is taken, the telescope
is moving between star and sky. (By the way, the telescope has to be turned
on, and the brakes off, to do this.) Note that the absolute encoders that
are displayed on the paddle ("old" TCS) are not accurate enough
to measure small moves. They may or may not indicate 60" moves, but
they should change quickly and substantially. If you are using the new
TCS, you should see the moves displayed more faithfully.
- This is a good time to set the rootname and sequence number for the
files (see below). The rootname is usually
date-coded (e.g. 7jul), and the sequence number is usually just set to
1 for a new night.
Presetting Focus:
One of the last things to do before the sun sets is to verify the focus
presets. For ROKCAM, the secondary encoder should read (see above for display
initialization) ~21900. The PXL Guider should
have been positioned coaxially with the beam at 43.5 cm for rough parfocalization
with ROKCAM, if it is being used as an imager. It's sufficient to check
with Scientific Support to make sure that this was done.
Verifying Tub Rotation:
This is a good time to make sure that the ATOG tub is rotated properly.
The standard orientation of the tub is 180 degrees, such that the TV camera
housing is on the south side of the tube when the tube is east of the polar
axle. In this orientation, the vanes on the cold pupil mask are supposed
to block the warm spider vanes, in order to minimize emissivity.
Topping off the Dewar:
You should start the night with the outer can full, and the inner can
with at least 1" of LN2. Make sure that this is the case, and that
the fill tubes have 90-degree angle vents that are tightened down directed
well to the sides, such that LN2 spillout won't fall onto the reimaging
lens at the top of the snout (if ROKCAM is being used as a direct imager).
You can assume that with a full outer can, some will spill when you get
far off the zenith.
The outer can has a nominal hold time of 9 hours. Within an hour or so
after the outer can runs out, the background count will start to rise in
a way that is highly non-uniform across the array. So although the array
itself is still cold, and reading out properly, the system performance
will degrade. On a long winter night in particular, you'd be best off refilling
at midnight. As long as the outer can hasn't run out, the inner can will
hold for several days.
If you have any doubt about the cryogen level in the cans, do not
hesitate to use the dipstick that is usually kept under the penthouse electronics
box. If the inner can runs out, it may take hours to equilibrate once it
is refilled. The symptom of the inner can running out is a rapid rise in
dark current, followed by flat zero readout.
Don't let that happen.
Before you fill, and if this is your first night, make sure that Scientific
Support has left the secondary in the f/18 position. If not, and the telescope
needs to go to service position to change it, much of the outer can will
get dumped, and youll just have to fill it again.
The blowoff from the inner can is especially slow, so the side port on
that vent should have a constriction that keeps at least slight positive
pressure in that can, to keep water vapor out and prevent an ice plug.
If you fill the cans, BE SURE THAT THE LENS CAP IS IN PLACE ON THE SNOUT
when you do, and be sure that the telescope is not tracking.
.
Alignment Checks and First Light
Many of the alignment checks can be performed well before it gets dark
and, indeed, immediately after the sun goes down. For information about
powering up and using the telescope, consult the 2.7m
Telescope Control System Manual.
- power up the telescope and set to a bright (V<1) star overhead
If you're confident that the star is going to be somewhere in the FOV,
you can go after a fainter star that is suitable for focus (see below).
If you do, be very careful to identify the right star, though, as life
will get miserable quickly if you boresight ROKCAM to the telescope control
system with the wrong one.
Boresighting ROKCAM to the Telescope Control
System (TCS):
- Be sure to remove the ROKCAM lens cap, and check by sighting along
the snout that ROKCAM has an unobstructed view of the secondary (and the
sky).
- Set to the J filter. Since the background is lower than at K, things
are a bit less complicated.
- Do a 1 second test exposure to look for the star.
If you see the star, that's great. It'll be way overexposed and bleeding
all over. Center it up roughly and zero the telescope encoders .
If you don't see the star, take a white piece of paper and look for it's
image at the focus near the top of the ROKCAM snout. You will need a star
with V<1 to do this easily. If the sky is reasonably clear, you should
be able to see the image of the star projected on the paper. Get the image
centered on the snout, and then try again.
Once this is done, ROKCAM is boresighted with TCS, and you should be able
to acquire any star using coordinates. Leave the bright star at the center.
Boresighting ROKCAM to the PXL Guider:
If you plan to do active guiding while taking data, you'll need to set
up the PXL guider camera.
- In the ATOG control panel set
TV @ROKFLD (position of diagonal mirror; default position is @2820)
XT @ROKFLD (X-position of PXL camera; default position is @2854)
YT @ROKFLD (Y-position of PXL camera; default position is @2178)
This moves the ATOG diagonal mirror over ROKCAM, and directs the telescope
beam to the PXL guider camera.
- Take a very short exposure with the PXL and acquire the optical image
with PXL. The PXL User Manual may be useful here.
Note that it may have to get dark enough so that the PXL isn't saturated
on scattered sky light.
- Center the image on the PXL by moving, if necessary, the PXL in X and
Y, using the ATOG XT and YT commands. Move the camera incrementally using
offset command lines like XT +100 or YT -50.
- When PXL is boresighted, read off the XT and YT encoders, and reset
the ROKFLD location definitions if necessary. For example, if the best
XT turns out to be @2801
SELECT XT
MODIFY ROKFLD @2801
Once this is done, ROKCAM is boresighted with PXL, and you should
be able to acquire any star in ROKCAM by centering it in PXL.
- If you have extra time, you might want to take this opportunity to
boresight the finder telescope too. It's unlikely that you'll need to use
it, but if you lose the COSMO boresight somehow, a boresighted finder will
make it easy to get back.
Verifying ROKCAM Beam Alignment:
The telescope mounting bracket for ROKCAM should hold it in position such
that the f/18 beam coming out of ROKCAM is pointed straight at the secondary.
It should be verified that this is the case, and that the beam is not vignetted.
Chances are, unless someone has been loosening screws that shouldn't have
been loosened, everything will be fine.
- Go back to ROKGUID positions to let ROKCAM look at the sky, past the
ATOG diagonal mirror.
- Move the telescope secondary way in or outside of focus, and look for
a symmetrical donut. If the star is too bright to do this easily, go to
a somewhat fainter one.
If you see a symmetrical donut, that's great. You're done.
If the donut is clipped off on some side, it's probably because ROKCAM
isn't pointed exactly at the secondary. The ROKCAM mounting bracket allows
ROKCAM to be loosened and pivoted in both directions. Do this one direction
at a tme, recentering the star after each motion, and remove the clipping.
Tighten the bolts carefully when it is done.
Note that the pupil stop inside ROKCAM is very close to the filter wheel,
so a misaligned filter wheel can vignette the pupil, clipping off the top
or bottom of the donut as seen with ximtool. So before you start
loosening bolts on the ROKCAM mount, home the filter wheel again, and make
sure the digital turns indicator at the bottom of ROKCAM is at 0000+/-15.
Focussing the Telescope:
- Go to a star that doesn't saturate the chip (peak <10,000DN, >8th
magnitude), perhaps just an anonymous star near the position the telescope
is on already. The advantage of using a much fainter star (>10 or so)
is that focussing efforts don't end up being confused by speckles.
- Use imexamine "r" option to do radial fits on the star images
for different secondary positions. Converge on the best ROKCAM focus, and
read off FWHM. Note that the star should be set on a cosmetically good
part of the array to get a successful fit.
- Another way to do it is to set ximtool display for a logarithmic stretch,
such that the image core sharpness is accentuated. This won't tell you
what the FWHM is, though.
- Go to ROKFLD and assess the parfocalization of the PXL.
If parfocalization is good, then you're done.
If parfocalization isn't good, parfocalize using instructions in the PXL User Manual .
Determining Direction and Scale, and Setting
the Orientation
The scale at the PXL guider should be reasonably close to 0.27/pixel.
If you need to check it, you can do this conveniently using the TCS to
offset the telescope precisely.
- While on a star in ROKFLD, measure PXL pixel position of star centroid
- Use TCS to move 60"N, measure PXL pixel position, compute PXL
scale
- Use TCS offsets or paddle buttons to determine directions of N and
E
- If direction bar and scale bar in PXL image are incorrect, reset them
manually using instrument->manual menu selection in PXL window
(see PXL User Manual ).
- Use instrument->view menu selection in PXL window (see
PXL User Manual ) to orient the frame consistently
with the paddle.
- Change to ROKGUID, and note orientation of ROKCAM field in ximtool
display.
Note that for standard tub orientation (180 degrees), the display defaults
with south to the right, and west to the bottom.
Calibration and Useful Observing Info
Standard Stars
There are many classic lists of near infrared standards that are relevant
to ROKCAM imaging work. TCS worklists are available for the indicated bulleted
lists.
- IRTF
Standard Star List
- UKIRT
Standard Star List
- ESO Standard
Star List
These stars, while mostly too bright for ordinary ROKCAM observing, are
of high photometric pedigree. Some can be used by defocussing the telescope,
and summing over the defocussed image. The following secondary standards,
most with K>7, can be used routinely for in-focus calibration. Be aware
that these lists represent different photometric systems that, while very
similar, should be reconciled for the most accurate work.
- Elias Faint Standard Star
List
Now with finder charts!
These stars are on the Caltech (CIT) photometric system. Note that K'~Kshort
data for these stars can be found in Wainscoat and Cowie 1992 AJ 103,33.
- UKIRT
Faint Standard Star List
These stars, from the 1992 JCMT-UKIRT Newsletter (Aug 1992, p.33), run
from K=8 to 14, and are well observed at J, H, and K. Several of these
are included in the 2MASS calibration star set below, though are assigned
slightly different magnitudes corresponding to the different system. K'
band finding charts for many of these stars are available here.
- Carter and Meadows Faint
Southern Standard Star List
These stars extend the faint standards into the southern sky, though many
are more or less equatorial. This page is from 1995 MNRAS 276, 734.
OK, but I really just want a good set of well-pedigreed
standards that I can reach anywhere in the sky, that don't saturate the
array, with a range of color that allows derivation of color terms, and
that puts me on a single, self-consistent photometric system that lots
of other people will be using!
Fortunately, efforts on 2MASS,
the near-infrared all-sky survey now being put together with NICMOS arrays
like ROKCAM, have recently resulted in a well pedigreed, well distributed
list of high quality standards based on stars from the UKIRT list, the
Elias list, and the NICMOS list. The 2MASS catalog will be the definitive
source of near-IR photometric data, and (now that you've read down this
far) the standard stars for 2MASS arewhat you probably want to use. Refer
to the 2MASS
Calibration Strategy 4.2 for questions of selection and pedigree.
- 2MASS Standard Star List
Now with finder charts!
---------------------------------------------------
- Temperature/Color/Spectral
Type Conversion
This table, from Tokunaga's compilation in Astrophysical Quantities (4th
Edition), is normalized to the CIT photometric system.
ROKCAM Sensitivity and System Characteristics
- The following representative point source sensitivities have been measured
for ROKCAM direct imaging with the 2.7m telescope. Sensitivities are dependent
on sky background (scattered light and auroral emission for J and H, and
ambient temperature and telescope emissivity at K), and seeing. So J will
be worse with a moon up, and K will be worse in the summer and better in
the winter.
J=21.0
H=20.3
K=18.3
These point source magnitudes correspond to S/N=1 in 1 minute of integration
on, after coadding a 6x6 pixel
(2.1x2.1 arcsecond) subframe around it, and subtracting the background.
- The system emissivity of ROKCAM + 2.7m telescope is ~30%. This rather
large emissivity is mainly due to the flip cage (in which the coude secondary
is oversized compared with the f/18 cass secondary), and the third mirror
stow position.
Note that background seen with ROKCAM in the K band is mostly thermal.
For emissivity=1 it is
S (Ks, T=70F) ~ 3300 DN/sec
S (Ks, T=50F) ~ 1650 DN/sec
Where T is the dome temperature. This is roughly what you'd see with the
dome or mirror cover closed, and represents a simple closed-dome sensitivity
test for the system.
With the dome and mirror cover open looking at clear sky near the zenith,
you should expect to see
S (Ks, T=70F) ~ 1000 DN/sec
S (Ks, T=50F) ~ 500 DN/sec
and note that S(K)/S(Ks) ~ 1.6, pretty much independent of temperature.
Obviously, the optimal K band sensitivity is obtained at the lowest dome
temperatures.
- The background level for the J and H bands is determined mainly by
moonlight and OH emission.
With no moon, at the zenith, representative background levels are
S (J) ~ 50 DN/sec
S (H) ~ 200 DN/sec
These numbers may vary quite a bit depending on auroral activity.
The background rate will usually determine the integration time that
you want to use in order to keep ROKCAM in a nominally linear regime (~10,000
DN or less). As a result, determination of the actual background rate is
one of the first things that you will want to do. Where short integrations
are required, be sure to consider using the ave or sum
options that are discussed below, in order to minimize the overhead involved
in nodding (and the resulting errors in guiding) and to maximize operational
efficiency.
ICEX Tasks for ROKCAM
There are a number of ICEX (/home/miranda/ice/ice-1.5) tasks that are
of special interest for ROKCAM data taking. See the ICE
manual for more generic details about this software, bearing in mind
that this manual has a lot of Kitt Peak-specific items. These tasks can
be commanded with shorthand labels that are the first few letters of the
full command name that can be interpreted unambiguously (e.g. obj
instead of objects, or g instead of grid). This
shorthand is awfully nice if you already know what is going to be considered
unambiguous. If you don't, you'll just have to try it out.
In the following descriptions, the following lexicon is used:
The ICEX command language is based on the observe command which
is somewhat flippantly termed in the ICE manual as "the only command
you really need". That is, all other ICE commands are shorthand for
what you could do, if you gave a complete enough command string with observe.
It must be noted that the ROKCAM implementation of ICE does not allow for
exposure pausing (there is no cold shutter!), and there is no graceful way
to abort an integration or a sequence. You can do cntl-C to get out, but
you will have to redefine home on all the CyberPak motors.
Also note that if you use "observe", you can set the imagetype
to whatever you want (it doesnt make any difference in the way the data
is taken) but be aware that the exposure time is recorded in the
file headers only when the imagetype is set to ir. The scripts
for these commands are available in ~dfl/ICE, and they need to be declared
in the login.cl file, as given in that directory.
- atlas2>ic>test
Takes a specified number of frames, saving them as test.pix and test.imh
(preexisting test frame is overwritten). Good for system checkout, since
it doesn't fill up the disk. Prompts for number of frames to be take, integration
time, and image type (which may be ignored here). Entries in response to
prompts become defaults. Part of the observe family (see below).
- atlas2>ic>observe
Parent command of test above, except the frames are saved
individually to disk (filename.imh and filename.pix for each integration),
and the observer is prompted for an object title and image type. The object
title is put into the top of the .imh file, and is diplayed at the top
of the ximtool screen with the filename. This is the basic command for
a family of commands (including tests above), which are functionally
identical to each other except in the image type that is saved. The (minor)
advantage in using the following commands instead of observe is
that the specification of the image type is at the command line level instead
of in response to a prompt.
flats - imagetype set to flat
darks - imagetype set to dark
objects - imagetype set to object
ir - imagetype set to ir (for some reason, only this imagetype
records exposure time properly!)
The ICEX file structure for a single "observe" integration
is as follows. Two files are written for each frame (header and pixel value
files).
- prnnnn.imh and prnnnn.pix, where
- p is a filename prefix string (see skysub2
below)
- r is a file rootname string (rootname
in obspars)
- nnnn is a sequence number (sequenc in
obspars)
-
-
- atlas2>ic>ir
This command works just like observe, in that it records
a single frame. Unlike observe, however, it writes a filename
in a somewhat more useful format.
nnn-ssss.imh and .pix
where nnn above is a three digit
running number that points easily to the running number on the log sheets,
and is presettable in obspars. ssss
is the image root name in obspars.
Note that ssss above is a string of indeterminate
length. You can have as many characters as you want.
Note that while nnn above defaults to three alphameric characters
(e.g. file number 4 will get the prefix 004), you can go over that number
of digits. That is, you can have a file that is called 1045-15sep.imh.
But if you go over three digits, it will make file sorting and sequential
listing a bit more inconvenient. Thus it is suggested that this sequence
number be reset for the start of each night.
- atlas2>ic>skysub
These are special sequences of the observe command. Takes
a specified number of star-sky data quads, saving each of the frames in
the quad separately, and displaying successively each pair in the quad
as it is finished, each quad as it is finished, and at the end of the sequence,
the average of all the quads in the sequence.
skysub can start the quad with the star frame or the sky frame.
One can select between these modes in skysubpar parameter file. The default
mode is to start with the star frame.
The beam switch distance (in arcseconds of RA and Dec) can be set in the
skysubpar file.
Note that the final displayed averages are now saved. These are all written
over upon starting a new quad.
sum: the sum of all the differenced quads in this sequence
difn: the object-sky difference for quad n (e.g. dif1,
dif2, dif3)
obj: the sum of the two "object" or "star"
integrations for the last quad
sky: the sum of the two "sky" integrations for the last
quad
The ICEX file structure for a "skysub" sequence follows
that for ir. Two files are written for each
frame (header and pixel value files).
nnnfssss.imh and .pix
- f is a frame-type flag that is - for single frames (from ir
above). Although the character can be modified in the skysubpars file,
the default flag reads "s" for
a "sky" frame and "o"
for an "object" or "star" frame.
-
- atlas2>ic>skysubt
This is a "test" version of skysub, in that the data is
not recorded. This is a good way to check out skysub, and make sure it
is doing the right thing. This sequence leaves behind the files "skysubt1",
"skysubt2", and etc., which are the results from each quad in
the sequence. These are written over on starting a new skysubt.
- atlas2>ic>ave n
Selects a number of frames for averaging (e.g. ave 5 defines
an image as the average of 5 successive integrations.) This is useful in
high background situations.
- atlas2>ic>sum n
Like ave above, except the sum, rather than the average is
used.
- atlas2>ic>daut
Sets the ximtool display to autoscale the image. Note that the autoscaling
will probably be dominated by bad pixels for single frames. Autoscaling
works better for difference frames.
- atlas2>ic>dsc n m
Sets the ximtool image display range to n (minimum) to m (maximum).
Since the ximtool display (and the handy pixel intensity number at the
lower right corner) divides the 64K DN image range into only 200 discrete
levels, this command is useful for both contrast stretches and for giving
more resolution to the pixel intensity readout.
- atlas2>ic>dtest
A quick way of doing display test.
- atlas2>ic>dsum
A quick way of doing display sum.
- atlas2>ic>grid
This is also a special sequence of the observe command.Takes
9 images spaced on a 3x3 grid. The task additionally prompts for the grid
size in RA and Dec, and the number of frames to be taken at each grid position.
Imagetype is prompted, but defaults to "ir".
Note that the default observing mode is one integration per frame. You
can set nframes in detpars to
"ave n" or "sum n" to sum or average
n integrations into a single output frame. (Actually, the shorthand "a
n" or "s n" works too.) Defaults to one frame
if the parameter field within detpars is blank.
This is useful in situations where the sky background is high, and the maximum
integration time is short compared with the time overhead required for quad
moves. In this case, you'd rather spend more time at each telescope position,
but the background won't let you take longer integrations. As long as the
background is filling up the wells, there is no penalty from readout noise
by averaging several integrations to make an output frame and, as a bonus,
disk space is conserved. Rapid changes in the background (e.g. clouds moving
by) however, should discourage any efforts to spend more time between star-sky
moves.
- atlas2>ic>filtgohome
This command homes the ROKCAM filter to filter #1.
- atlas2>ic>filt n
Sets to a particular filter number n (where #8 is "blank", for
example) , once the drive is homed.
If you want to use the assigned filter names given in instrpars,
you'll need to do the longhand
atlas2>ic>instument instrfi=@blank for example or @n
The standard filter set is identified by the names (? above= J,H,K,Kshort,H2,CO,Kcont,blank.)As
noted above, the observer should NOT assume that the list in instrpars
is current. The definitive list is physically attached to the ROKCAM dewar.
All filters, including the blank are cold.
If you want to tell the system which filter it is really on then use
- atlas2>ic>filtis n
where n is the filter number (1-8)
- some miscellaneous other handy not-so-shorthand
...
obsp.r="rootname" sets the file rootname (e.g. obsp.r="10jul")
obsp.seq="seq" sets the file sequence number, which
gets incremented
detp.n="ave n" or detp.n="sum n" to
change nframes
instru.i=@f to change filters where f is the filter label (e.g.
instru.i=@Kshort)
End of Night Checklist
Shutdown procedure for the 2.7m is covered in the 107"
telescope operations manual. Don't forget to fill out the on-line night
report using xreport on atlas. If you do not want to fill the cryogen
cans, be sure that you have informed Scientific Support that they need to
do it.
Make sure that the PXL shutter isn't cycling continuously. For ROKCAM itself
...
- if using ROKCAM for direct imaging, cover the field lens at the end
of the snout, make sure the blowoff fixtures are directed well to the side
why is this very important?
the field lens is unfortunately close to the focal plane, so specks of
dirt will cover stars, and you don't know who might move the telescope
during the daytime (perhaps to service position), and otherwise dump LN2
on the field lens.
- top off outer can
- back up data onto exabyte tape, or ftp them somewhere else.
The files are small enough that the latter is an acceptable alternative.
See this page (from Lauren Likkel) for instructions.
If you use ftp, be sure to set mode to binary (image). Check to
make sure that complete files were sent. The .pix files should all have
length of exactly 133120 bytes. We've had some trouble in the past
where incomplete files were sent, and such corrupted files are unusable.
We have the capability for making CD-R copies of the data back in Austin.
Let us know if you'd like one made. Random access, robust, long-lifetime,
standard formats, 650Mb capacity ...
|