MDM Observatory's 2.4m Hiltner telescope is among largest telescopes in
the world that always operates without a night assistant.
This document is a guide to observing with the 2.4m. It contains
checklists for operating the equipment, designed to prompt experienced
users who haven't observed recently, and more detailed sections intended
to elucidate common procedures for novices. This document is
not intended to replace the more detailed manual in the control
room (also available in a
web-based version). When in doubt, "Read The Fine Manual".
If you're preparing for your first run at MDM, you should watch the
instructional videotape which I recorded back in 1996 or thereabouts.
Although many of the details are out of date, it's mostly still current,
and it gives a sense of the look and feel of the operation which you
can't get from a verbal description. There should be a copy of the tape
at each Consortium institution.
Note that there are quite a few instrument combinations available, some
of which I've never used personally. I'll try to make clear when an
item or a statement refers only to a particular setup.
- Hand paddle won't work after a move ...
- On occasion the telescope fails to complete a move and gets
hung up, usually very close to the object you're moving to.
The most obvious symptom is that the hand paddle is locked out, so
it seems as if there's something wrong with the paddle. To diagnose
this condition, look at the Hour Angle field of the NEXT OBJECT
line near the top of the TCS monitor. When a move completes
successfully, this field gets filled in with asterisks, "* * *".
If you don't see this, but rather see the hour angle of the
next object (which should be changing), the TCS is hung.
The fix is simple; in the xtcs window on hiltner , click on the
red STOP virtual button. (Do not press the red plastic stop
button on the black control panel, that will stop the drives!)
This will stop the endless wait state, and give you the telescope
back. Do SEND and GO again, and the move should complete.
- I've totally lost pointing!
- Calm down, slow down and follow the
step by step instructions
under How the telescope works .
"Moving with Autodome off" or "RA and dec encoders not responding" ...
- These error messages are almost always spurious. You can safely ignore
them unless there seems to be a real problem (dome isn't moving with telescope,
or encoder readings do not change when telescope moves.)
"OMG, the dome just went haywire!"
- Occasionally -- maybe every few weeks on average -- the dome will start
rotating for no reason, usually during an observation. I call it the dome's
"fugue state", after a psychiatric condition marked by aimless wandering and
amnesia. This is clearly caused by a
software bug in the TCS. No one knows a graceful way to stop it, other than to
sit and wait for it to snap out of it. This can take anything from a
fraction of a minute to quite a while -- I've had the dome make several full
revolutions! The best strategy is to wait it out, but if it does go on for a
while with no end in sight, you could turn off the track switch
on the black TCS panel (which should stop the tracking, you hope), note
the hour angle and dec on the TCS, and then
try using the TCS mouse and pull-down menu to exit the TCS program. The
risk here is that you'll lose pointing, and have to go do the
pointing recovery procedure, but
you may be able to avoid that lengthy process by re-setting the encoders
to the last known position of the telescope after you bring the TCS back up, and
then hunting for a nearby bright star to tweak up. In any case, it's a known problem.
[Contents]
Closing
[Terse version.]
- TV gain down, image tube power off, TV monitor off.
- Once you're no longer looking at the sky, turn 'em off. Leave the
guider PC on.
- take sky flats?
- See the earlier discussion under Opening for
details. If you've written down the exposure data on previous mornings
and evenings, it should be possible to get a fairly good exposure the
first time.
- slew to zenith.
- An easy way to do this enter coordinates
RA = sidereal time + 1 minute
Dec = 31 57 12
Epoch = now (in decimal years)
into the xtcs window, and slew.
- reset instrument rotator to 0 degrees if needed.
- If you've rotated the instrument and it's in some funny position,
you should set it back to zero, at least on your last night. Remember
that the mirror covers must be open to rotate the instrument!!
- close mirror covers (do not move rotator with mirror cover
closed).
- You should ideally be near the zenith to close the mirror covers,
though in an emergency you should just go ahead. To do this, flip the
silver switch on the black panel down. You'll hear the four of them
crash down, hopefully in order.
- set accurately to zenith; track off, drives off.
- It's worthwhile to get the telescope dead straight up, since then
you know exactly where it is if there's some kind of problem (e.g. a
lightning shutdown). Turn off the tracking and set the hour angle to
zero.
- all other switches down. Dome goes home.
- hit red STOP button on TCS black box.
- dome lights on.
- close instrument dark hatch.
- close dome.
- if dome did not return to contacts, hit the
yellow button and tweak manually.
- dropout must close before main shutter
- Note that the main shutter takes a while to close, so you can get it
started first, and then hold down the dropout close button until it
slams shut. If you need it, the yellow auto-dome enable switch
(discussed earlier) is in the computer room at the bottom of the
left-hand rack.
- If there's a problem getting power to the dome - for example, if the
rotation fails and you can't get it on the contacts - there is an
emergency cable which can be connected. See the full manual for
details.. Note that plugging in the cable with the dome on the
contacts will cause a short, so don't do that!!
- top off instrument dewar
- Doing this as the dome is closing is a good habit. Don't forget to
mark the time on the whiteboard. See earlier cautions on filling
upward-looking dewars and not filling warm dewars.
- close louvers and garage door
- verify dome fully closed
- back up the night's data
- If you lose the data, your work, and the telescope time, are gone!
It should be sufficient to copy the data to another computer (e.g. a laptop),
or, if you're using the 8k you may want to put it on tape.
- The Linux rsync command is your friend here. For example,
if you're using one of the ccdcom-controlled detectors, you might copy
your data onto agung by making your own directory, cd-ing into it, and
typing something like
rsync -azvt /data/krakatoa/obs24m/mydir/night1 .
This will make a copy of "night1" in your current directory. If
"night1" already exists, it will copy only new files or files that
have changed.
- There's a nice DVD-burning utility on Agung that you can
use to make copies of your data.
- If you're using the 8k, be sure to
back up on tape as you go along. Observers sometimes get into trouble
by leaving their copying
to the last day. Not only does this leave you vulnerable to disk or
computer failure, it can cause problems if you underestimate the time
required, leading to collisions with the next observer. Back up every
night, and be sure that your plans for writing tapes the last morning of
the run are realistic.
- check air conditioner; adjust if needed
- Once the mirror petals are closed, the air conditioner will try to
cool the mirror to a settable temperature, because the seeing goes bad
in a hurry if the mirror is more than 1 degree C or so warmer than
ambient. The mirror has a long thermal time constant (like 12 hours),
and the air conditioner is not particularly effective at cooling the
mirror, so if you want to adjust the temperature for the next night,
better do it now.
- How to decide on a temperature? The A/C control program has a
feature which displays the history of the ambient temperature, the
mirror temperature, and the temperature set point for the last 3 days or
so. Detailed weather forecasts are available on the web; the National
Weather Service Tucson office (http://www.wrh.noaa.gov/Tucson/twc.html)
has a terrific site which includes very technical details of the
forecast. If humidity could be an issue, you'll want to be careful not
to overcool the mirror.
- record any messages for staff (e.g., filter changes) on whiteboard
- If you have any requests which will require timely action by the
staff, please be certain to write them out legibly on the whiteboard.
You cannot expect to wake up at 3 PM, saunter out of bed, and tell the
staff to drop whatever they're doing and swap out all your filters. In
this, and everything else, try to think ahead, both to assure that your
requests can be handled and to minimize the strain on our very small
staff. Your nightly observing report (next item) is another place
to post requests like this -- the staff reads these thoroughly.
- Fill out the Observer's Nightly Report Form.
- MDM requires that all observers fill out a brief report of the
night's activities. To make this as painless as possible, a simple web
form has been created that is accessed from the
mountain web server..
Even if you did not observe, you should fill out a report saying why.
Nightly observing reports are stored on the mountaintop server, and
emailed to a short list of recipients that includes the observatory
directory, mountain superintendant, and the MDM consortium
representatives from each member institution. Report data are used
to track how much observing time is lost to weather, problems, etc.,
as well as to keep a permanent record of activities.
Similarly, if you encountered any problems at all, even ones you
fixed yourself (i.e., not common mistakes but real problems that
required you take some special action to continue working), also
fill out a Trouble Report Form. There is no such thing as a "minor"
trouble report. Often a big problem starts as lots of little
problems or quirky annoyances. If we can see a pattern develop in
the trouble logs, we might be able to head off bigger problems
later. Remember, the run you save could be your own!
Some problems occur when using the data-acquisition software or other
computer systems. It will greatly help diagnosis and solution of a
problem if you can include the verbatim text of any error messages
printed when the problem occurred.
- go to bed!
- Your program may call for instantaneous data reduction - if so,
bring a collaborator or automate it so you can get enough sleep. It's
important not to get too tired!
[Contents]
At the end of your run ...
[Terse version.]
- Plan to depart as soon as practicable; keep staff informed
- Accomodations at the Obsevatory are very limited, and the next
observer should not have to work around you as they set up and settle
in. Accordingly, observers are required to leave the mountain as soon
as they can, ideally by the early afternoon of the day after their last
run. Please keep the staff informed of your plans. If you intend to
sleep for a while after your final night, be sure the staff knows when
you intend to get up and leave, or they have no idea what to tell the
next observer about bedroom availability and so on. The
instant-departure rule can be relaxed when there is a very good reason
(e.g., packing up a complicated user instrument at the end of a run),
but exceptions must be cleared with all concerned!
- Pay your meal bill (ideally on the afternoon of the last
business day before your departure.)
- If you've taken any meals up at the Kitt Peak cafeteria, you're to pay
for them before you leave. This is at the main office of the Admin
building up top. You can dial zero on the phone to tell Joanne
(usually) that you're departing and have her make up a receipt for you.
If it's after hours or the weekend, you can go to the admin
building and fill out a form with a credit card number -- they'll
charge you and send you a receipt. But you have to remember to
do this!
- Pay your room usage and/or vehicle usage fees.
- The MDM Board has (with regret) instituted a room
charge of $40/night to help cover the Observatory's budget, to be instituted
at the end of January 2008. This will probably be implemented on a
per-telescope (rather than per night) basis, so as not to discourage
training of students. We are also recharging observers $50 if they
transport themselves to and from the mountain in an observatory vehicle
-- this should just cover the actual expense, and is much less than the
cost of a rental car. In any case, make arrangements with the staff
to pay these fees, and get a receipt to claim the expense. A PayPal
account is in the works for these fees, in the meantime a check
certainly does the trick.
- Check with staff about whether to fill dewar.
- Sometimes the staff may want the dewar to run low because it makes it
more convenient for them to change the instruments.
- Make arrangements to get off mountain if needed
- If you're using the Kitt Peak shuttle you'll have to reserve a spot;
the schedules are kept in the Admin building up top.
MDM pays an annual fee so that observers may use the shuttle;
we are not charged by the passenger, so go right ahead.
MDM users may not drive Kitt Peak vehicles, so you can only use the
'U-drive' schedule options when someone else is driving.
- If you're using an MDM vehicle to
get off the mountain, you'll have to arrange it ahead of time with the
staff. Be sure to give the staff plenty of advance notice!
- If you're driving yourself off the mountain, be
very careful in this case not to overextend yourself to the point
where you're really sleepy while driving. Falling asleep at the
wheel is one of the most important causes of fatal crashes -- don't
put yourself in a position where this is a danger.
- Be sure all your data are properly backed up.
- Standard operating procedure is for the staff to simply wipe out
your data from the observatory computers as soon as you're gone (though
if you ask real nice they can sometimes be persuaded to leave your data
alone until you verify your tape is legible at home). It's therefore a
good idea to write more than one copy of your data. There are many
possibilities for this -- tapes (for 8k data), transferring the data to
a laptop, writing
a DVD, bringing your own USB-pluggable disk drive, etc. The observatory
internet service is now fast enough that modest-size
data sets (e.g., single-chip detectors) can be shipped home, but given
that the phone lines can sometimes go down you shouldn't
depend on this exclusively! It's a good idea to leave a backup at the
observatory in case something gets wrecked on the way home.
- Be sure to allow sufficient time to back up your data. You should
plan any tape writing carefully so that you're finished early enough to
allow the staff to start instrument changes when they arrive at 8 AM.
As noted earlier, you should be backing up your data as you go along
anyway.
- Tidy up control room; erase grease-pencil marks from TV
- Gather up all your charts, scratch paper, whatever, and leave a neat
workspace for the next observer. If you've marked up the guide TV
monitor with grease pencils, erase all those marks now -- they probably
won't apply to the next observer. (Incidentally, you should never write
anything
on a computer monitor, and you should be sure to use easily
eraseable grease pencils if you feel you must make marks on a TV
monitor. I never do this anyway, since there are cursors to mark
things, and yellow sticky pads too.)
- Throw away uneaten food and wash your dishes
- There's a tiny refrigerator and larder in the kitchenette -- space
is very limited, so don't leave behind anything which could go bad. Who
knows when that half-eaten sandwich was parked there, anyway? There are
vermin (verpersons?) at the observatory, so you should never let dishes
pile up, but if you have, better get 'em now.
- Strip your bed
- Stuff your used linens and towels into one of your pillowcases and
leave by linen closet.
- Leave your bedroom tidy
- There's no chambermaid. Behave accordingly.
- Be certain: bedroom windows closed, heater and lights off, no water leaks
- As noted earlier, water and electrical power are extremely expensive
on the mountain. No one else is going to check your room to be sure
you've turned things off, so be especially careful that you do. Be
absolutely sure that (a) the heater is off and (b)
the toilet isn't running!
- Look around to be sure you haven't forgotten anything.
[Contents]
Acculturation for New Observers ...
I trained as an observer at Lick in the 1970s. At that time new
observers were walked through procedures by seasoned staff observers.
This certainly got the job done - we learned the equipment - but in the
long run, the most important lessons we learned from this were not about
which buttons to push, or how to develop plates. They were instead
lessons about the experience of observing, and the attitude to bring to
the telescope. As an old curmudgeon I think some of these lessons have
atrophied over the intervening years, as overworked faculty pack
students off to observe with minimal preparation. Here's a distillation
of some of that acculturation, as refined through many hundreds of
nights of observing experience since then, most of them at MDM.
- Telescope time is precious, and using a large telescope is a
privilege. You should feel a strong obligation to use your time
well.
- Accordingly, it's important to be efficient. This means giving some
thought to your procedures. It also means having a well-thought-out and
well-prepared program. I remember one famously efficient observer who
was said to have spent three nights of large telescope time efficiently
observing the wrong list of objects - he'd grabbed the "already
observed" list instead of the "to be observed" list, and hey, galaxies
all look alike!
- Be careful with the equipment. It's expensive stuff and
carelessness is not an option. Leave everything in good working order.
Report problems completely and carefully so that they may be fixed.
Leave the control room in better order than you found it. And when
conditions become dangerous to the telescope (blowing dust, too-high
humidity, etc.), close immediately!
- Work very hard, but try to take care of yourself. If you get too
sleep-deprived, you're liable to make really dangerous errors.
- Most primary programs require excellent conditions. What if your
conditions are usable, but mediocre? What if the seeing is 2 arcsec
instead of 0.7 arcsec? What if you require photometric conditions but
there are high clouds? This happens a lot - have a backup program! If
you can't think of one yourself, ask around your department and
collaborators - someone else is sure to need something.
- You should take a workmanlike attitude toward your data. Pictures
should be in focus and centered. Spectra should be in focus and
properly calibrated - and they should be of the object you intended!
Take a few moments to get things right.
- On the other hand, better is the enemy of good. I remember a
meticulous student who hardly ever took data because the conditions were
never perfect enough for his high standards - he had the most wonderful
dome flats, though. I actually witnessed a memorable scene in which an
instrumentally-minded investigator spent most of a beautiful night
fine-tuning equipment endlessly as the stars wheeled overhead and a
roomful of European collaborators flown in for the occasion became
increasingly exasperated. Telescope time is precious. Use it.
- Even though you should feel a strong obligation to use the time
well, you'll be more efficient in the long run if you don't hurry.
Hurrying leads to really big mistakes, like that poor guy observing the
wrong objects.
- Finally, have the discipline to avoid excessive distraction by
the TV, the stereo, and so on. Sometimes observing gets boring
and it's fine to let part of your mind do something else, but it's
important to stay on task. Observing is like a dance in time,
and it either happens or it's gone -- pay attention!
[Contents]
Computer System Overview
This little blurb isn't complete but might help get you
started. A description of the many improvements wrought by the
August 2005 upgrade is given
here .
There are three computers at each telescope. At the 2.4-meter telescope,
these are:
- hiltner:
- This is the main Observing Workstation where the observer logs in
for most things. It is a fast Linux box.
- agung:
- This is an identical PC running Linux configured as a Data Reduction Workstation.
It is usually where any second observers will be able to login and work.
- krakatoa:
- This is a Sparc 10 running Solaris configured for special data-acquisition
tasks. krakatoa runs the MDM CCD cameras with a
special interface. It is anticipated that krakatoa will be retired soon
and replaced with a Linux machine, ideally in the summer of 2006.
Hiltner and agung run the K windowing environment,
and have been set up to have a common look and feel which will be
familiar to most observers. The
window system has four desktops available - there's a little
cartoon in one corner showing them all.
On all machines you login to a visitor account named obs24m.
Logging onto hiltner as obs24m puts you into the
directory /lhome/obs24m. The data all go to directories
under /data, where the next item in the path is the
machine name (e.g., hiltner or krakatoa), followed by the user
name (obs24m), followed by whatever you want to put there.
The /data directories are transparently visible to
all themachines.
The telescope and the MIS (Multiple Instrument System, used for
everything except the MDM 8K camera) are controlled from windows on
hiltner. There's one window called xtcs, which
runs the telescope control system, and another called xmis,
which runs the MIS. These controls are thoughtfully designed and should
be fairly intuitive. The
web-based manual has sample displays of these windows, in color.
The usual MDM CCDs (Echelle, Charlotte, Templeton, etc.) are run by a
program called ccdcom. This has a text-based interface and
runs on krakatoa. There's a menu choice on
hiltner for a ccdcom window, which automatically
logs you onto krakatoa and gives you an xterm with yellow
letters on a black background. You now cd to the directory
you want your data in, and type ccdcom to invoke the ccd
control program. There's a good manual
describing ccdcom, and a later section
of this Guide gives a little advice on the program.
Note that you can run everything from hiltner's terminal -
just ssh to the other machines as you wish.
It's probably a good idea to use agung for any heavy
reduction, to keep hiltner free for observing.
[Contents]
How the Telescope Works
The
official manual contains a lot of detailed information on the
workings of the telescope. Here's a short overview which may be helpful
but which is necessarily very general. (Note also the
gotchas listed elsewhere, all of which are
known telescope bugs.)
The telescope is generally configured as an f/7.5
Ritchey-Chr�tien - an f/13.5 secondary is seldom used. It has an
equatorial fork mount, built by DFM engineering in the early 80s. The
drives are unusual in that they do not have worm gears - rather, the
telescope is driven by large steel wheels with smaller driving wheels
pressed up against them.
The Telescope Control System, or TCS, is a PC containing a fair amount
of custom hardware. In particular, it contains counters that listen to
and interpret the pulses from encoders, and boards that issue signals
to the power electronics "muscle boxes". These in turn send power to
the stepper motors which run almost everything. The TCS dates from
1995, when the original custom computer which came with the telescope
was retired. Because of limitations of the multitasking software
available for PCs at the time, the operating system of the TCS PC is
OS/2, which will be unfamilar to almost everyone; luckily, observers
should never have to interact with the OS (except if they need to
turn it off for a lightning shutdown, but this is straightforward).
The user interface for computer control of the telescope is through a
program on hiltner called xtcs. This is fairly easy to use
-- to slew, just enter the coordinates and hit `send' and `go'. As
noted earlier, this can be made more efficient by putting your target
list in a file ahead of time and calling the objects by name. The
xtcs window also allows users to reset the encoders, adjust
track rates, and so on. It may be helpful to keep in mind that
xtcs is just a user interface -- the actions are mostly
occurring in the separate TCS computer.
When the TCS computer program is running normally, the TCS computer's
monitor displays the status of the telescope. It's designed to be
similar to the old TCS monitor, so the comprehensive
manual in the
control room gives a fairly good idea of its function. I'd recommend
that you peruse the monitor display carefully on your first day - you'll
be looking at it a lot, so you'll want to understand what you're seeing.
Under most circumstances, the telescope points to about 15 arcsec rms or
better. This figure can change depending on how recently the pointing
errors have been mapped and modeled; the position as displayed has been
corrected using a model of the telescope errors. The position displayed
is also corrected for refraction, nutation, aberration, and precession,
so it should approximate the mean coordinates for the specified equinox.
The model (and the corrections) are handled correctly when the
coordinates are reset using a known star. The telescope readouts
sometimes drift slightly in declination during the night, especially if
you do a lot of long north-south slews; apparently a few pulses from the
dec encoder get lost.
Note that the telescope has some pointing limits. It can't get
extremely close to the horizon, or various hard and soft limits are
triggered (see the comprehensive
manual for where these limits are and
how to back out of them if you get into them). The RA is limited to
+/-6 hours to avoid cable wrap problems. Observing under the pole is
not really supported, though I hear it's been done.
There are manual control paddles to move the telescope. These have
directional buttons NSEW, and two buttons labeled SET and SLEW. The
actions of these buttons are quite standard. Holding a directional
button down moves the telescope very slowly in GUIDE rate (typically one
or two arcsec per second of time - the guide rate can be adjusted with
the xtcs window). Holding down the SET button results in a
much faster rate, about 1 arcmin per second. Finally, the SLEW rate is
full-speed, around 1 degree per second, used for major repositioning of
the telescope. If you need to slew manually, you must keep an eye on
the telescope to be sure you know what it's up to.
Restoring lost pointing.
Occasionally the telescope may lose pointing entirely. Hardware
failures may be responsible; more frequently, it's due to pilot error
(updating pointing carelessly and then slewing away before the error is
detected). There's only one way to reliably set pointing from scratch,
and that's to slew to the zenith and reset the encoders. Follow these
instructions exactly and pointing will be restored. The whole
procedure takes 10-15 min at the most, if you try to improvise you'll
waste much more time than that, so don't.
- Turn off the TRACK switch (on black TCS switch panel).
- Get a working flashlight, and go out in the dome, turning on
the red dome lights as you go. There's a telescope hand paddle out in
the dome, it's usually hanging near the southwest side of the polar
axis. Get it.
- Look at the telescope. Figure out which way you should move it to
point at the zenith (east or west? north or south?). Then, using the
paddle, manually slew the telescope
to near the zenith. Keep looking at the telescope as you slew to be
sure you're moving the right direction and not driving toward the
horizon!!
- Go back in the control room. If you have another observer leave
them in the dome.
- Within a few degrees of the zenith, tiltmeters
mounted on the telescope will read how far off the zenith you are.
The readouts are on a rack in the computer room. Move the telescope
with the control room handpaddle to set the tiltmeters to exactly
zero, or to whatever the current zenith reading is, in both axes.
(If you have another
observer in the dome have them alert you
over the intercom if you go way off by mistake.) Note that sometimes
the tiltmeters seem to drift away from the correct zenith; there will
often be a handwritten note giving the tiltmeter readings for the
correct zenith; if so, use those.
- In the xtcs window, set the TCS Equinox to
the current epoch, within half a year or so, e.g. 2006.2.
- Find the Sidereal Time display on the TCS monitor.
Note the time and write down the sidereal time 1 or 2 minutes in the
future, to allow time for typing and so on.
- Type the coordinates of the zenith into the
appropriate fields in the xtcs window on hiltner; they are
RA = sidereal time (the one which you just wrote down).
Dec = 31 57 12
Equinox = right now (decimal year)
- SEND the coordinates to the TCS.
- Pull down the setup menu in the xtcs window,
and note the set RA and dec encoders item. Watch the
sidereal time readout on the tcs. At the moment the sidereal time
matches the sidereal time you've set up, let your finger off the
mouse button
to reset the encoders.
- Verify that the hour angle is reading within a couple of seconds of
time of zero, and that the dec is within one arcminute of 31 57 12
in the present-epoch coordinates.
- Turn the TRACK switch back on.
- The xtcs window has a button labeled Get Coords;
click on this and select Nearest Bright from the pull-down menu.
This will load the Yale Bright Star catalog entry nearest the zenith.
- Slew to this bright star. Since all the bright stars are around 6th
or brighter this should be a huge unmistakable star. (It's a common
mistake to set up on the wrong star. Don't be fooled by cheap imitations,
this sucker will be bright!) The zenith
tilt meters are not necessarily accurate enough to get the star in the field
of view (which for some instruments is very tiny, like 90 arcsec), so you
will probably need to scan the telescope back and forth a little bit
in order to find it. Note that
the best way to pick up the bright star will be instrument-dependent, and
requires a little thought on your part; here are some suggestions:
With the MIS: In xmis, find the pink "Preset" menu button for the guide
probe position, and pull down "center". This will put the
upward-looking probe in the center of the field. Don't
forget to move it away when you're done, since it will
block your science instrument there.
Retrocam: Put in the retrocam and take a quick shot. It's accurately
boresighted with the telescope and shows a much bigger
field of view than the MIS guider.
With the 8k: Take an image (probably binned pretty heavily to minimize
read time) and the very bright star should be obvious.
- Use the directional buttons on the telescope handpaddle to put
the bright star near the center of the field of view of your instrument.
- With the telescope pointing right at the star, go to the
pink "Setup" menu in xtcs and pull down
the "set RA/dec encoders" item. This should reset the telescope
position to that of the bright star.
- Before moving away, make absolutely certain that
the coordinates have set correctly on
the TCS; look at the original coordinates of the star (recall it again in
xtcs) and verify carefully that you have reset the coordinates correctly.
Be aware of the coordinate "epoch" (more properly equinox); if the
telescope "epoch" is not 2000, click on the
"TCS Equinox" button in the xtcs window and adjust the telescope's
display to 2000 (this affects the display, but not the actual
position), and then the telescope should agree with the star's catalogued
position. If you've used any offset commands in centering the star, be
aware of the following insidious way to
make a mistake : the
set RA/dec encoders command sets the encoders to the
values in the Next Object field. But, if you send offset commands to the
TCS, this over-writes the Next Object values, so when you set, it will
be to the wrong coordinates. The cure is to use the Send Coords button
in xtcs to reset the Next Object field to the correct coordinates
before setting the encoders. Bottom line: don't move away until you're certain you
have it right.
- Select another bright star nearby and slew to it to
double-check. You can do this by slewing randomly by 10 degrees in
any direction and doing "Nearest Bright" again.
- Don't forget to turn off the red lights if you turned them on!
- If you've put the MIS guide probe in the center of the field,
better withdraw it to where it's not blocking the beam.
That's the procedure. I hope it's worked for you!
Here are a few more aspects of the telescope you'll want to know about:
The focus control: The telescope is focused by holding down the IN and OUT buttons on the
paddle. This moves the focus rather slowly. If you also hold down the
SET button, the focus moves much more quickly. The focus numbers are
arbitrary and their zero point varies widely - your numbers from the
last run mean nothing. Focus numbers tend to decrease as the
temperature goes down.
The Instrument Rotator: There is an instrument rotator at the
back of the telescope. It has its
own encoder, which is not particulary reliable these days; you can set
the encoder using xtcs. There are various pieces of tape
marking different position angles. Note the following about the
instrument rotator:
- You must have the mirror covers open to rotate the
instrument. The mirror covers bear upon a moving part of the
instrument rotator! Metal shavings and paint flakes don't do much
good for the mirror!
- The rotator paddle is in the dome. Be sure to turn the speed all
the way down before turning the rotator on or reversing direction.
- While rotating the instrument, use a flashlight to watch carefully
for any cables which might be hanging up. It's a helluva thing to
have a cable catch on some critical switch or knob, pull it, and then
get ripped out ... ruins your whole night.
- To avoid problems with cable hangup, the rotator angle should be
kept within +/-90 degrees.
What's the "slit angle"? :
On this last, note that there's a little quantity called the "slit
angle" displayed on the TCS, in the upper right-hand corner. The
story behind this is as follows. If
you do slit spectroscopy away from the zenith, your data suffer from
atmospheric dispersion - the star is smeared into a little spectrum in a
direction perpendicular to the horizon. You can capture all the
wavelengths in the slit by orienting the slit perpendicular to the
horizon. The position angle of an arc connecting a given point to the
zenith is called the parallactic angle, because it's
(anti)parallel to the direction of topocentric parallax displacements.
Ordinarily, with the roator angle at zero the MDM spectrographs are
oriented with their slits north-south, so that they're on the
parallactic angle for any object crossing the meridian. Once you're far
away from the meridian, you want to rotate the slit to the parallactic
angle, but keep the rotator angle within +/-90 degrees. Because the
slit is indifferent to 180-degree rotations, but the rotator isn't, the
"slit angle" is the rotator angle which, for the present position of the
telescope, will put the slit on the parallactic angle and keep the
rotator within its travel. For instance, if the parallactic angle is
-150 degrees, the "slit angle" will be +30 degrees. It's important not
to interpret the "slit angle" as an actual readout of the rotator
position. The rotator position readout is the "Rotator Angle",
displayed near the middle
of the TCS monitor about 1/3 of the way from the top.
If you
intend to track an object over a large range of hour angle, you
may want to check out the parallactic planner button on the
GUI skycalc; it computes the optimal rotator setting over a
range of time for your object and gives a graph showing how
badly you do if you're off by 10 or 20 degrees on either side.
This feature is only on the skycalc versions installed at
MDM, not in the usual distribution. Like the
guide star selector, it pops a
bright blue
temporary xterm window and
a white-on-black pgplot Xwindow (which may cover up the
xterm window); hitting ENTER in the blue xterm
window makes the whole thing disappear.
[Contents]
About the MIS
All the commonly used instruments except the MDM 8k camera are
mounted on an adapter called the MIS (Multiple Instrument System). This
provides a number of commonly-needed utilities. Some useful diagrams
can be found in the
manual.
There are three parts to the MIS. Working downward from the telescope
they are
- The finder unit:
- This is a gold-anodized box about two feet square and 10 inches
deep. It contains a prism that can slide in and out of the beam to feed
an Apogee CCD camera called RETROCAM, so called because it
was retrofitted to the MIS by Chris Morgan of OSU.
The finder unit also houses a set
of comparison lamps - an incandescent flat bulb, and Ne, Hg,
Ar, and Xe discharge lamps - for calibrating spectra. The prism that
feeds the RETROCAM also diverts the comparison lamp light
downward to the instrument, and the optics are arranged so that
they provide an approximate f/7.5 beam to match the telescope. Note
that you can either see comparison lamps or the sky, but not both at the
same time, as they call for different positions of the RETROCAM
prism.
- The guider unit:
- This is about 5 inches deep; it bolts to the back of the finder
unit. It has a pickoff mirror which feeds optics leading to another TV
camera. The pickoff mirror can be moved around on a precision stage,
which I'll call the guide probe. The TV fed from this mirror is used
for offset guiding; one finds a guide star by moving the guide probe,
and then starts the autoguider to keep the telescope locked in position.
At one position of the stage, the TV looks into a slit-viewing
microscope, which is mounted at the end of a push rod. If you're
doing slit spectroscopy, the staff will push the slit-viewing optics
into place as part of the setup procedure. If you're working direct,
the slit viewing optics should be retracted. The guider TV has a fairly
small field (like an arcminute). The Ohio State CCDS has its own
slit-viewing arrangement. This allows one to view the slit and
offset-guide at the same time, which is a great advantage over the
arrangement used for the Modspec and Mark III, where one
must guide using light reflected from the slit jaws.
- The filter wheel:
- The filter wheel is only used for direct imaging.
There are two wheels, one of which takes 2-inch
filters and the other 4-inch filters. The 2-inch filters vignette the
20482 SITe chip to about 16002. The older
MDM 4-inch filter wheel is (as of this writing) still in use at the
2.4m, but a far superior version is being built by OU and OSU --
it's a copy of the existing "Buckeye" filter wheel at use at
the 1.3m. The old 4-inch FW is a bit cranky, but behaves OK if
you only advance 2-3 filters at a time, and always in the
forward direction.
The MIS is operated using the xmis program on
hiltner. This issues commands to the MIS control computer,
which is an ancient piece of equipment residing in the computer room
racks. The controls are fairly straightforward to operate. There are a
number of preset positions for the guide stage (slit, center, etc). You
can also type in an X or Y coordinate for the guide stage and hit return
to send them; an indicator blinks red while the stage is moving, then
stops blinking and shows the new position when it gets there. You can
also get relative steps using the "Delta X" or "Delta Y" fields. Other
buttons or menus allow you to turn comparison lamps on or off, move the
RETROCAM prism in or out, and so on.
There are a number of things to be aware of when using the guide probe.
- The GUI version of skycalc running on agung makes it almost
criminally easy to pre-select guide stars for your target, using the
USNO A2.0 star catalog as a database. The full explanation
is given a little later in the document. This is highly
recommended, even if you're working at low latitude where
guide stars are plentiful, since the software automatically excludes positions
in which the guide probe might obscure the field center (see
the next item on the list). Also, for some programs it's possible to use a pre-set
guide star to greatly speed up field acquisition and centering.
- It's important to note that there's nothing to stop you
from blocking the telescope beam with the guide probe. You
must keep the guide probe away from the ! There's a
very confusing chart in the manual about this, but a rule of thumb is
that if you keep the X coordinate less than a few thousand you'll be OK.
If you have a cloudy night you might want to experiment by taking dome
flats and moving the guide probe around until you can see it when you
divide one picture by another - I've never mapped out the "safe" area
myself, and someone should do it.
- The guide probe camera needs to be focused so that when the
telescope is in focus, the guide probe is also. The guide probe focus
is controlled by a rocker switch on a little aluminum box a couple
inches square. If
you're doing spectroscopy, and guiding off the slit jaws, you focus the
guide camera during spectrograph setup, by looking at the slit jaws
illuminated by a flat lamp (you'll need the TV gain way down), and
focusing the guide probe until the slit looks as sharp as possible.
Then when you get a star on the slit jaws, you simply focus the
telescope until the star looks as good as you can get it. The slit jaws
are in focus in the TV, and the star is in focus in the TV, so therefore
the star is in focus on the slit jaws, which is what you want. In
direct work, you focus the telescope on the CCD somehow (classical step
'n shoot focus test), then find a guide star and focus the guide probe
on it.
NOTE: the focus in the slit viewing microscope is way different
from the focus when the guide probe is looking up into the telescope as
for offset guiding. This means that it's impractical in spectroscopy to
park a star on the slit, then hunt for an offset guide star; by the time
you've found a guide star and refocused the guider, the telescope has
drifted enough that you can't be sure your program star is still in the
slit. The OSU CCDS solves this problem by having a separate
slit-viewing camera of its own, freeing the MIS guide TV for what it
does best, namely offset guiding.
- If the guide probe is driven past the end of its travel, it loses
track of where it is. I've also seen it lose track when it gets
close to the end of its travel in Y, say at 10000 or greater.
In any case, it takes only
a minute or so to reset the coordinates, by selecting
Origin in xmis. This drives the probes back to
its zero position and resets the counters.
- The guide probe travel is larger than the unvignetted field of the
telescope. You'll find that you can't see anything at Y<2000 or so,
for example.
- Easy pre-selection of guide stars!
The GUI version of skycalc available on Agung contains an
MDM-specific tool for selecting guide stars from the USNO A2.0 catalog.
The gui has its own manual .
Instructions for starting the GUI skycalc program are here.
Once you have the program up, it's easy to select guide stars:
- Enter your object's RA and dec in the main window (upper left). (If you
have read in an MDM-style pointing list, this is especially painless:
just double-click the name on the list, or type the name in the top
entry box and hit return.)
- If the instrument rotator isn't at zero, open the "Guide Star Config"
window and enter the appropriate value.
- If you're south of +40 to +50, hit "UCAC2 guide*", otherwise
(of if UCAC2 doesn't show any stars) hit
"Get Guide Stars" (which uses the USNO A2.0 catalog).
This pops a bright blue
temporary Xterm window, and a big black Xwindow with a diagram of the
guide field and a cursor. Click on a star near the left-hand edge
of the field (for best results) and its guider XY coordinates will
appear in the blue Xterm window. Stars between magnitudes 12 and
14 appear to work best, but brighter and fainter also work.
- When you're done, type "q" in the black Xwindow and
the whole apparition -- Xterm and Xwindow -- will disappear.
- Fast target acquisition using the guide probe.
The guider XY stage is very accurate, and so is the
USNO A2.0 catalog (though some stars may have moved in the 50 years
since the POSS-1 on which it was based). Therefore centering the preselected
guide star in the appropriate location on the TV should put you
very close to your target, provided of course that you're
setting on the right star! This can save enormous amounts of time.
To develop this procedure, start with a target you can set up
accurately. Select a guide star using the program (described
immediately above), set up the guide
probe at the XY coordinates predicted by the program, aim the
telescope accurately at your target, and
note where the guide star lands on the TV. If you keep the
guide cursor there as you set up on new fields, and center
the (correct!) guide star in the box, you should be
very close. Be careful of the
a HUGE GOTCHA here, explained a
little farther up:
the guide probe can lose track of where it is when it is
driven to Y-values greater than about 10000. It does
this silently -- no error message is printed, and it all
looks fine, but the numbers are all wrong.
Just reset the probe coordinates by sending the probe to
the present "Origin" if this happens.
Also, you'll want to be sure you know the rotator angle accurately
[Note: A previous standalone version of the code, called gs24,
proved to have a bug in it which introduced 10-20 arcsec
inaccuracies. This has been fixed and the procedure described
above has been field-tested with very good results; with care
one can get within a few arcsec typically.]
- It's also possible to hunt and peck for guide stars if
you need to (the USNO A2.0 is no good in some crowded fields, for
example). Here's how to do it.
(The units used here for X-Y are the default, raw counts; you can
make them arcsec or mm instead.)
- Center up on the object
- Be sure gain is up on the guide TV
- set the guide probe to X = 0, Y = 2000
- set Delta-Y to 1000
- repeatedly hit carriage return and watch
the TV to look for guide stars
- stop when Y = 10000; if you drive beyond that, the guider
will silently lose track of its Y-coordinate.
- At the top of the Y travel, set X = 1000
and Delta-Y = -1000, then step back down ...
and so on, raster scanning until you have a guide star.
In principle you can scan across the low Y range and up the far side of
X without vignetting the chip, but this is seldom necessary. The
field maps produced by the skycalc tool give an idea
of the useful field.
More about the MIS TV cameras
The TV cameras currently used in the MIS are tricky devices. It's
important for many applications to understand them well.
The cameras consist of Gen I image tubes coupled to video-rate TV
cameras. Since image tubes are less common than they were a while
back, here is an explanation. An image is projected onto the front face
of a vacuum tube, where there is a photocathode. Electrons liberated
from the photocathode are accelerated greatly by a high voltage.
The fields in the tube are arranged so that the electrons from a given
spot on the input all land in the same spot on the other end, thus
preserving the image. Where the electrons land, there is a phospher
which glows when excited by these high-energy electrons. Because
the incoming photon has only about 2 eV, and the electron has several
keV, there's much more energy than there was to start with, so the
intensity of the image is greatly amplified.
Our image tubes are military-type units which are fed from a 6-volt
power supply -- the high voltages needed to run the tube are generated
internally. The voltage feeding the tubes is controlled by a power
supply in the control room -- it has a knob which can feed zero to
a little over 6 volts to the image tubes. The higher the voltage, the
higher the voltage in the tubes, and the more the image is amplified.
There are several important gotchas with these image tubes:
- The tubes discharge voltage with a very long time constant
(like 10 minutes). Once you turn the voltage up, it is up -- unless
the tube is discharged somewhat more quickly by bright light.
- If the gain is up, bright star images completely saturate. You
really can't use them to focus or to assess the seeing. When you first
set on a bright star, advance the gain slowly to avoid saturating
it. Because of the long time constant (item 1), once the gain is too
high, you can't bring it back down!
- The power supply is set up to "crowbar" at a little over 6 volts.
If you turn up the gain beyond that, it abruptly resets to zero, and
you have to turn the gain all the way down to get control of it
again. It's like you dropped it, and have to go down to pick it up
again. Before you walk away and let the telescope autoguide while you
grab lunch, be sure you haven't let the gain drop to zero, or your
guide star will -- gradually disappear!
Adding the DTI box to the mix adds another level of complexity,
but it can be worth it because you can see much fainter. Here's a
recipe for getting the most out of the DTI while viewing the slit:
- Retrocam OUT (so you're looking at sky).
- Check TV selector set to GUIDER.
- On the DTI, set the black toggle switch to DIRECT.
This passes the signal straight through.
- Advance the image tube gain while watching the monitor. Stop when
it starts getting medium-bright (but way short of white).
- Switch the DTI to INTEGRATE.
- Set the DTI time constant to 1/4 sec or less.
- Spin the BLACK and WHITE knobs on the left side of the DTI (they
have many turns) until the contrast looks right (greyish sky with white
stars if they're present).
- If the image is too dim, slowly advance the image tube gain while
watching the TV carefully.
- When it looks optimal, increase the time constant on the DTI. Time
constants up to 2 sec are useful, beyond that you're just amplifying noise.
- Be aware that you're now looking at an integrated image, which will
respond slowly.
[Contents]
About the autoguiders
The autoguider programs reside in a big old Gateway 486 PC which sits on
a table just inside the computer room (it had been in the control room,
but it was too noisy to bear.) There are two distinct programs, namely
- The "New" or "Seitzer" guider (TVGUIDER)
- This is named after Pat Seitzer, who engineered it. In this guider,
the video signal comes right into the PC where it's averaged on a board.
You put a little box around the guide star, adjust a few things, and
start the guider. It works really well for offset guiding - there are a
bunch of things you can tweak, but you generally don't have to do much.
Pat has also written a very
user-friendly manual, which you should read carefully before using it.
- The "Old" or "DTI" guider (PCGUIDER)
- This uses the old cream-colored Digital Television Imagery ("DTI")
box to integrate the signal. This integrated signal is fed into the PC,
which then generates a guide signal, using a program originally written
by Mark Metzger and later tweaked slightly by myself. This program does
not offset guide as well as the Seitzer guider, but it has compelling
advantages when
you are guiding from an image reflected from the spectrograph slit.
I wrote a manual for this guider which is quite complete, and it has
on-screen help and default settings. (Older versions of this manual
warn that the default guide rates are too aggressive, but in 2008
January I changed the default guide rates in the code to more
appropriate values.)
The Seitzer guider program works only with the integrator in the PC, and
has nothing to do with the old DTI integrator. Conversely, the program
that uses the old DTI integrator knows nothing about the
integrator in the PC. Since the two pieces of software use
different hardware, switching between the two
guider programs involves plugging and unplugging cables, which must
be done by the staff (hence during the day).
The PC programs for the two setups have confusingly
similar names (TVGUIDER = the new, Seitzer guider, while PCGUIDER = the
old, DTI guider).
The Seitzer guider is much preferable for direct work -- it's simpler
to use and guides better -- but the DTI
guider has compelling advantages for spectroscopic work (except with
CCDS which uses its own slit-viewing camera).
Most importantly, the Seitzer
guider cannot display an integrated image and guide at the same
time. Furthermore, the integrated images in the Seitzer guider
aren't as smooth or as deep as those in the DTI. If it's important to
see and guide on faint objects on the slit jaws, the DTI rules. As
noted earlier, it would be better if there were a separate camera to
view the slit jaws (as in CCDS), freeing the guide probe to roam around
and guide the telescope with the Seitzer guider, but that's life.
Some TIFKAM observers have noted a way in which a subtle difference
between the DTI and the Seitzer guider can be important. When the
DTI-based program starts guiding, it notes the position of the star in
its little box, then tries to hold the telescope exactly on its initial
position. The Seitzer guider, on the other hand, tries to keep the star
at the exact center of the little box, even if it didn't start out right
at the center. The steps by which one moves the little box on the TV
amount to a substantial fraction of an arcsecond; with a 1/2 arcsec
slit, you can lose much of your flux just because of this silly
digitization problem. A better strategy for fine adjustment of the
telescope position when you are offset guiding -- as with TIFKAM,
CCDS, or direct imaging -- is to move the guide probe slighlty. The
Delt x and Delta y input fields in the xmis control
window make it convenient to do this.
The DTI integrates with a "leaky memory", which gradually forgets the
oldest signals; the image changes continuously, unlike the Seitzer
guider, in which the new image appears abruptly. The abrupt updates can
cause headaches after hours of staring.
There's often a lot of confusion about how to set the adjustable
parameters of the guiders in various situations. TVGUIDER (the Seitzer
program) wakes up with good parameters for direct imaging, and the
manual should be helpful for fine adjustments. PCGUIDER (the DTI)
is rather more complicated. It has a
menu of defaults for different setups, which should get the
parity and rotation right. In old versions of the program, the
default guide rates had been set way too small, but this has
been changed so it should work "out of the box".
The rates do also
depend on the guide rates set in the TCS, which control
how fast the telescope moves when the button is pushed.
The refresh rate and number of frames averaged are also adjustable
in PCGUIDER;
I recommend setting the refresh rate to match the DTI integration
time, and averaging a few frames together, because the guide signal
can be glitchy.
A simple way
to test any guider is to get it going on a fairly decent star and push a
guide button to throw the telescope off a little bit.
The error you introduce should be big enough to be obvious, but
small enough so that the star stays within the guider's field
of regard, which is about the size of the little guide box on the TV.
Then, watch to see if the star is guided back to the middle
of the box. If it goes off
the wrong way, there's something wrong with the rotation or parity
(try both axes to be certain); if the star overshoots, your guiding is
too aggressive; if it doesn't move, the guider may somehow be disabled
(is the little switch on the TCS panel in the up position? The
bottom of the TCS monitor has a display which reports guide pulses --
are the commands getting through?) The speed at which the telescope corrects
will of course be a function of your guider parameters and the guide
rates set in the TCS.
Nearly all autoguider problems are due to observer error or
unfamiliarity, but occasionally something may break.
It's worth noting here that for nearly all programs, an autoguider is
not essential - it's only a convenience. Back in my day, when we walked
uphill to school both ways, we guided all night by hand. A
dysfunctional autoguider is a pain, but not a show-stopper.
If you can't get the damn guider to work after a reasonable
effort, don't close down -- suck it up and hand guide! There is
a stereo, after all.
[Contents]
Some advice about CCDCOM
The standard old MDM CCDs are run by CCDCOM. The Ohio State instruments
CCDS and TIFKAM are run using an entirely separate progam called
Prospero, and the MDM 8K camera uses yet another control program, on a
different set of computers.
CCDCOM has its
own manual, which I am not repeating here. The purpose of this
chapter is to draw attention to some useful features, and a
hidden gotcha or two.
One of CCDCOM's most useful commands is source, which takes
as an argument the name of a file (that resides in the directory you're
currently using for the data). This causes the commands stored in the
named file to be executed by CCDCOM. Thus if you have a complicated
set of repetitive actions, you can simply edit them into a file (called
foo, let's say), and type
source foo
to execute the commands. Here are some ways you can use this to
make your life much easier.
First, the readout format of the chip is set by the sf ("set
format") command, which is followed by a bunch of numbers describing how
the chip is to be binned and which columns and rows are to be read. If
you've set up to read a subset of the chip (as for spectroscopy, say),
you'll want edit up a little one-line file like
sf 2 2 300 400 500 600 700 32
or whatever the numbers are, so that if the camera has to be restarted
you can quickly and accurately reproduce the settings. If
you're off by one column, your flatfields don't work!
The source command can be very helpful in direct work, too.
The read time of a big chip can be very considerable, so sometimes it's
useful to read only a subarray. I like to set up files
bigform and smallform, which are 20482
and 10242, so I can switch quickly. I also like to have a
findform, which is the central 10242 binned 2x2,
to give a very short read time useful for verifying the centering. You
can nest source commands (a source command can
occur in a file you're going to source), so I like to have a
command which moves to focus mode, sets the chip format to small,
changes the file prefix to something like "scratch", turns off clearing
of the chip, and so on and so on, and then another command which goes
back to full size, object mode, clearing the chip, and recording the
data. Commands like this save a lot of time and prevent a lot of
errors.
Here's another application. As a radial-velocity spectroscopist I must
keep very close track of the wavelength calibration, which means taking
comparison spectra before and after every sequence of exposures. The
comparison lamp set at MDM is pretty poor down in the blue. There's a
Hg lamp which has some very important bright lines, but there's a desert
between 4358 and 5461. Xenon has some nice lines there, but the Xe lamp
at MDM is extremely faint compared to the Hg. Furthermore, the Hg lamp
has a long warm-up time, like 60 sec.
To get good comparison lamps every time, I combine the source
mechanism with the tel command, which can command the
telescope and MIS. Using a source'd file, I turn on the
lamps to warm up, move the guider mirror away from the microscope to
avoid frying the TV, move the Retrocam prism into the beam to reflect
the comparison light down into the instrument, wait a little while with
the tel sleep command, set the CCD to focus mode
to get a multiple exposure, clear the chip, take a 0.1 sec exposure to
get the bright Hg and Ne lamps, turn off the bright lamps, expose again
for 60 sec to get the Xe, then turn off the lamps, read the chip, move
the retrocam prism and pickoff mirrorback into place, and reset CCDCOM
into object mode to take real data. This is error-prone, not to mention
tedious, if you do it manually.
A word of warning: I've had troubles driving the 2-inch filter wheel
with the CCDCOM tel mechanism. Better run it by hand with
xmis. The OSU 4-inch filter wheel may work better.
You can interrupt CCD exposures by typing Ctrl+C (i.e.,
holding down the Ctrl and C keys simultaneously).
It is a very bad idea to type Ctrl+C while the
chip is clearing or reading out - wait until it's exposing.
Once the
exposure is stopped you can adjust such things as the name of the
object, the exposure time, and so on. As for the exposure itself, you
have three choices:
- go <n> - continue the exposures, optionally
including the number of exposures to do
- rc - read the ccd and store the data.
- cl - clear the chip, throw away the exposure so far.
Note the little gotcha in the go option: if you've started a
sequence of (say) five exposures, and interrupt the second one, then
typing go only starts one exposure, so you stop at two! To
restart the sequence you'd need to type go 4 to get the
remaining four exposures.
When you're taking sky flats, especially, the istat command
is needed. It does some quick statistics on the last image taken.
With this you can see if the exposure level is correct. Ideally you
want between 10K and 30K counts.
CCDCOM has an "IRAF" flag, which now defaults to "on". This writes the
images (in FITS format) in such a way that IRAF properly understands
them as unsigned short integers (0 to 65535). This way, the
digitizer uses all 16 bits without wasting a bit on the sign, since the
raw numbers are all positive.
CCDCOM does not display images or do any analysis (though the
istat command can be used to check mean values).
For displaying images, I recommend
using ds9 directly, that is,
use the "File" menu in ds9 to select your image, and
do not use the "display" command in IRAF. The reason for
avoiding the IRAF "display" command
that it automatically rescales the displayed image for 8 bits of
depth, thus destroying most of the information; it also destroys
almost all the header information (e.g., wcs if you have it).
If you display directly using ds9, you can still
use the imexamine task (and others) in IRAF, amazingly
enough.
[Contents]
What time should I start?
OK, so the weather's great -- low humidity, photometrically
clear, and the wind is a gentle zephyr. So you can open. But
when should you plan to start observing? Here are a few remarks
on that.
It obviously depends on what time it gets dark.
The
GUI skycalc
installed at MDM (and also its
Java version)
has a "Nightly Almanac" button, which
pops up a window with the information you want. Instructions
for how to launch the program appear
immediately below.
If you can do it, it's a good idea to open the dome maybe 15-20 minutes
before sunset so you can have your dewar topped off and ready to go by the
time the sun actually sets (that way you can enjoy the sunset and
see the green flash). There are constructive uses for twilight time.
If you're doing direct imaging in the optical, sky flats can start
in the U band shortly after sunset, within 5 minutes or so, and the
window of opportunity for well-exposed sky flats is pretty short.
Bright stars can be found almost in broad daylight to verify
telescope pointing. Certain kinds of spectroscopic standards can
be done in quite a bright sky -- e.g. 10-th magnitude flux standards
should be do-able within 20 minutes after sunset. Just be sure to
keep the TV gain low when the sky is really bright!
[Contents]
How do I start GUI skycalc?
The GUI skycalc, with its guide-star selector, planetarium display
and so on is available only on agung, not hiltner (and at the
1.3m it's on hill, not mcgraw). You have (at least) two ways to
get at it:
- On Agung's (or Hill's) console, it's installed in the launch menu
(the red hat in the lower left).
But -- you'll have to scoot over to agung to consult it.
- To lanch it on hiltner, click on the little terminal icon in the lower
left to get a terminal. Into this terminal type the following:
ssh agung
skycalcdisp.py
and it should fire right up. The "disp" means it has the planetarium
display implemented; the ".py" indicates that it's in the Python
language (which I believe everyone should learn, as it is more
useful than, say, algebra).
[Contents]