Quick Guide to Direct Imaging
with the P60 CCD camera


by Michael E. Brown and Marc Kuchner


Planning Your Observations
Afternoon Preparations
Biases, Darks and Dome Flats
The Start of the Night
The Middle of the Night
The End Of the Night
Summary of Some Useful Commands
Time Saving Tips


Planning your Observations:

-Choose your CCD. Almost everyone uses CCD 13. It is occasionally in use on the 200-inch for the Norris spectrograph, so a good second choice is COSMIC. Curves of QE are found at the Palomar website. The best CCD, #9, is only available for use with the 60-inch echelle. Sorry.

-Choose your filters If you are using broadband filters, you will either go with UBVR etc., or griz etc. Details to come.

-Pick your calibration sources.

-Look up the object coordinates, twilight times, moonrise/set times, and sidereal time at midnight. XEPHEM, available on GPS unix machines, works nicely for all of this, or try the ephemeris at the Keck website (it's good for Palomar to within ~20 minutes).

-It's nice to prepare a printout of all the coordinates of your targets to give to the night assistant so you don't have to go yelling them back and forth.

-Plan to arrive at Palomar at least 3 hours before dinner. Don't forget to bring some exabyte tapes for your data.


Afternoon Preparations


-Familiarize yourself with the control room set up. There's OSCAR (the old VMS machine that runs the camera) on the table with the two monitors above it. The monitor to the left displays useful general information like the pointing of the telescope, the local sidereal time (LST) and the Universal time (UT). The minutor to the right tells you imformation about the CCD camera, like how long you have left in your exposure and the status of the chip. The cool new auto-guider is to the right of this table, and the two new Sun ULTRAs (thebe and leda) are behind you.

-Hanging on OSCAR's table will be a paddle that will allow you to move and focus the telescope. Note that when you move the telescope with the paddle (or the "focus" command), you will alter the offsets ("DRA" and "DDEC" on the monitor), but not the nominal pointing. These displayed offsets are corrected for atmospheric refraction, so they might not corrspond exactly to where you think you are moving the scope. To move the telescope faster, hold down the middle buttion on the paddle.

-Log in to Thebe or Leda (login "user", the password is written on the monitor) and print out some log sheets. Ben Weiss has a form prepared. Here's the .ps file.

Check that everything is working. -Look at the night assistant's video camera and make sure no lights are on in the dome (or you can go upstairs). Make sure flatfield lamp is turned off (retro metal box above the TV display terminal)

-Take a 1 second DARK using the DARK command (see below) and see if you have something that looks like a CCD frame. It take about 3 minutes to read out the chip. Yes, this is a pain in the ****.

-Check to see if you have a light leak by taking a 100 second DARK and comparing the results (should be about the same, only more cosmic rays in 100 seconds).

-make sure the right filters are in place. Type FZERO and the filter wheel will be homed and the filter names will appear (hopefully) as you requested them. If not, find someone to put in the right ones.

Everything above should be working correctly, but it is good to make sure.

-delete any old data on OSCAR (the VMS machine which runs the camera) by typing DCLEAR. This deletes EVERYTHING on the data disk. So be careful.

-set up your directory structure on OSCAR. You should be in directory SCR:[CCD]. If you would like to put different nights into different directories, make a new directory with the VMS command CREATE/DIR [CCD.NEW1] or whatever. To change to your new directory (so that all of your new data will be put there) type SET DEF [CCD.NEW1]. Also then use the DESTINATION command. If you are going to take 5 minutes to transfer all your data to thebe or leda at the end of each night, you can ignore this step.

-set up your CCD the way you want it. Binning will make the readout faster and reduce the readout noise, at the expense of worse sampling. Sub-arraying will NOT decrease your readout time but will give you smaller, more manageable images to deal with later. The CCD is accessed through the command CAMERAS, which will ask you a series of 3 questions, all of which should be answered in the default. You will then see with a menu of camera parameters, including the binning and sub-framing setup.

Go to Top


Biases, Darks and Dome Flats


-You will need biases, darks, and dome flats to reduce your images. In real life, the dome flats may be unneccesary, but they are nice for online reduction and are good to have around just in case.

BIASES:

These are 1 second exposures taken with the shutter closed. Take 10 at a minimum. Here's a simple way to do this: Make or DOWNLOAD a text file on thebe or leda called mdark.com, that consists of the following lines. Ftp it to OSCAR (remember to set the transfer type to ascii, by typing "ascii" at the ftp prompt). Then run it on OSCAR by typing @MDARK.

$dark 1 dark \ nodisplay
$dark 1 dark \ nodisplay
$dark 1 dark \ nodisplay
$dark 1 dark \ nodisplay
$dark 1 dark \ nodisplay
$dark 1 dark \ nodisplay
$dark 1 dark \ nodisplay
$dark 1 dark \ nodisplay
$dark 1 dark \ nodisplay
$dark 1 dark \ nodisplay
$rename dark0.dst;1 dark1.dst
$rename dark0.dst;2 dark2.dst
$rename dark0.dst;3 dark3.dst
$rename dark0.dst;4 dark4.dst
$rename dark0.dst;5 dark5.dst
$rename dark0.dst;6 dark6.dst
$rename dark0.dst;7 dark7.dst
$rename dark0.dst;8 dark8.dst
$rename dark0.dst;9 dark9.dst
$rename dark0.dst;10 dark10.dst

The first 10 lines take 10 1-second dark exposures and save them with filenames like "dark0.dst;4". The next 10 lines rename the files to dark1.dst, dark2.dst and so on.

DARKS:

These are exposures taken with the shutter closed that are as long as your target exposures. For imaging you probably don't even need to bother. In fact, we just measured that you get less than 1 dark count per ten minutes of exposure, so it hardly amounts to anything.

DOME FLATS:

In the afternoon, you should take dome flats. Turn the lamp on by flipping the switch on the gray box above OSCAR to "on". You can adjust the current in the dome lamp to change the amount of light, but don't go above a current of "600" and remember that the power goes roughly as current squared. Turn on the telescope by pressing "telescope" ON and "dome rotation" AUTO on the night assistant's console. You can now drive the telescope north and south but not east and west. Move the telescope down to the platform on the north side of the dome, climb up the stairs and remove the cover. Lean the cover against the railing to your right (as you look down the telescope tube) and secure it with the little metal arms. Then move the dome so that the telescope points to a blank section of the wall by typing "90 godome" on the night assistant's terminal and pointing the telescope to a declination of 75 degrees north.

NOTE: if you have just removed the cover, you must move a good bit SOUTH to reach the right position as you have gone north through the pole already and need to come back. You can now take dome flats.

Put in the filter that you are using by typing FMOVE and then typing the name of the filter. You want your dome flats to have ~16000 counts. Stephen Odewahn, Reinaldo de Carvalho and Roy Gal put together a booklet called "POSS-II Observing Procedures", available as a postscript file, which includes the following suggested exposure times, with the lamp at 600 for Gunn filters:
CCD 13
Filter
g
r
i
Exposure time (s)
30
5
3
counts
15000
15000
15000
CCD 16
Filter
g
r
i
Exposure time (s)
30
7
4
counts
10000
14000
14000

You want about 10 flats per filter that you are using. A simple way to do 10 in a row is to make or DOWNLOAD a VMS file called MDOME.COM with 10 identical lines reading

SNAP 0 0 domeflat \ nodisplay

Ftp this file to OSCAR and then simply type @MDOME.

You should finish all of this before the night assistant shows up so that he can get the dome opened as early possible to start temperature equilibration.

Go to Top

Eat dinner. Very important.


The Start of the Night


Twilight calibrations:
- There is only a narrow window of opportunity--when the sun has just set and the sky is still bright--when you can get twilight flats. This part is exciting and requires quick typing. General philosophy is that the twilight sky is reddish and darkens (or brightens, in the morning) exponentially. Take the telescope out of focus. When the sun has just dipped below the horizon, start by taking 2 second exposures (the minimum you should ever use for a flat) until the image is no longer saturared and gives about 16000 counts. Keep adjusting the exposure time up as the sky gets darker so that you keep getting about 16000 counts. Move the telescope between each image, just in case there are some bright stars in there somewhere. Ben Weiss has found that if you multiply the exposure time by 2.2 each time, that will be just about right, if you are reasonably quick at the keyboard. If you have more than one filter to do, do the bluest filters when the sky is brightest. You can buy some time by starting your flats with the telescope pointed toward the East and swinging the telescope towards the West as the sunset progresses. You will need at least three flats in each band.

- Focus. This can be done when twiflats are finished. Look in your twiflat images and make sure that there are stars that are going to be visible (and not saturated!) in a 5 second exposure. Then focus by using the command FOCUS, and take 7 exposures of 5 seconds each. Type FOCUS MOVE=20 to have the telescope move 20 arcsec between each exposure. Set the focus (using the telescope paddle hanging to the right of the keyboard) to the current focus minus 60. Whenever setting the focus, go down to a value about 30 below where you are setting, and come up to the value. Now take your 7 exposures, moving the focus up by 10 between each one. From Thebe, ftp over your .dst file from OSCAR and convert it to a .fits file with the fig2fits command. View the image with saoimage and find the best focus in your image by looking at the close-up in the little screen in the top right corner. You can tell which focus position corresponds to which image because the star jumps the most between the last two images. Reset to the best focus (again, by coming up from below).


- Go take a standard. Move to a standard field, take something like 4 exposures while dithering around. You can use the focus command to do the dithering automatically, but remember it does not reset your pointing back to where it was before! There is a P60 Standards binder on the shelf above OSCAR where you can find lists of standard stars. If your standard saturates the CCD (more than 20,000 counts) in a 2 second exposure, you might want to defocus a little. Don't forget to refocus again (and do that overshoot thing).


The Middle of the Night



-Autoguiding. Your night assistant will set up the autoguider. Be sure to disable it when you slew by clicking on the two "enable" buttons (they will turn yellow, indicating they are disbled). Ckick those buttons again when you are done slewing to re-enable the guider. To dither between exposures while keeping your guide star in the guider, set the autoguider to drift, move the telescope, watch the guide cross hairs move to the guidestar's new position, then set the autoguider back to guide.

- Give the operator the first set of coordinates and move there. Start taking data (you will probably primarily use the SNAP command). Make sure you are guiding if you are taking a long exposure. Write papers. Become famous. Hire some of Mike's grad students. Or at least take the time while you are doing long exposures to listen to some good music (the CD player and tape player are on the night assistant's side of the control room).

-Refocus regularly. You will need to refocus at least once during the night because of flexure. Also, each filter gives you a slightly different focus. The Johnson filter set generally has a focus setting -10 from the Guinn r filter, for example.

-Keep an eye on the weather and the seeing. Note changes in your log.

The End of the Night


-You might do standards and twiflats and even domeflats if you didn't get them in time.

-Transfer all your data to a Sun workstation. Make your own directory in the scratch space on thebe or leda; for instance, when I (mjk) observed on November 13 I made a directory called /scr/user/mjk/nov13. Then ftp to oscar and do an "mget *.dst" to transfer all of the .DST files. Change the .dst files to FITS format by typing (on the Sun station) "fig2fits *.dst" and then rm the .dst files. Insert blank tape. Type SETENV TAPE /dev/rmt/0n. Check the tape status with "mt stat" and finally do a "tar cvf $TAPE ." to write everything to tape. This may take a while. Pick it up on your way out the next afternoon. Do an "mt offline" to rewind the tape and eject it.

Go to Top

Summary of Some Useful Commands


Data taking commands:
SNAP take an image
If your exposure is longer than a few seconds, snap will run in the background, and allow you to use OSCAR for other things while you are exposing. Note that for exposures shorter than 2 seconds, the shutter timing error creeps in (see the chart on the wall).
DARK take an image, but don't open the shutter
TEST take an image, but don't save it to disk
IMAGE display an image that has been saved to disk Image also allows you to look at an array of DNs surrounding the location of the cursor on the chip; choose option "P". This is a good way to see whether a star is saturated, or estimate the seeing.
FOCUS take a series of focus images
MODEXP change the exposure integration time
STOPALL stop an exposure once it's started (throws data away, unless you then type KEEP which saves it, or REGO which restarts it. If you want the data to be thrown away, type READY)

Filter commands

FMOVE change filters
FSTAT Show the names of all the filters in the filter wheel. The names are insensitive to case! Make sure to give all your filters starkly different names. After you do an fstat, the filter wheel is reset to position 1, so you will have to reselect your filter.
FZERO home the filter wheel

Other Commands

EXTRACT make an row extraction, save it to disk
YSTRACT make a column extraction, save it to disk
SPLOT plot up one of these extractions.
COLOUR GREY fix the color scale when it gets screwed up
FRAMES resets the frame number, so after a night of making files s1-s5, say, you can start again at s1.

Time Saving Tips



-To quit from a command, hit control-Y. Never quit using Control-C. It could crash the computer.

-When you are answering questions in, eg., the image command, answer with a backslash "\" to skip over all the questions and
accept all the default values.

-Use the up-arrow to repeat commands you have used earlier.

-Don't do anything involving OSCAR while it is reading the chip... it could crash. Do use this time to slew the telescope.


When You need More Information


Manuals galore are on the shelves above the monitors in the observing room. Other useful resources are experienced P60 CCD observers like Roy Gal and Ben Weiss.

Go to Top