MDM Observatory's 8K CCD array project based at Columbia Astrophysics Laboratory: BASIC STARTUP AND SHUTDOWN PROCEDURES
Arlin Crotts e-mail: (
Telephone: (212) 854-7899


A warm, well-pumped vacuum should read at most 20 mTorr. When liquid nitrogen fills the system, this reading should stabilize at less than 0.5 mTorr (with pressures less than 0.1 mTorr appearing as "-0.0") on the Granville-Phillips gauge display. When the liquid nitrogen fill is complete, the aluminum cap connected to the 8K by surgical tubing should be attached to the bottom of the 8K nitrogen feed tube (by turning the attached clamp). (The other end connects to the purge tube on the 8K, in the plate holding the CCD window, on the side of the instrument towards the shutter cassette. This keeps a weak flow of gaseous nitrogen past the CCD window, keeping it from fogging.) Don't do anything odd that would force liquid nitrogen through the tube, and be cautious of nitrogen pouring out of the CCD as the telescope points to large zenith angles. The 8K's hold time is between 12 and 24 hours.


(The first three paragraphs of this section cover procedures usually performed by the MDM staff.)

The instrument should be mounted on the 2.4-meter with the filter wheel protruding to the north and the dark slide positioned to the east. There is a pointer and fiducial mark on the instrument rotator and telescope corresponding to 0 degrees. (One cannot trust the rotator encoder to better than a few tenths of a degree.) These are labeled and are found just west of the north edge of the rotator. For the 2.4m/8K, the balance settings are 1550 on the horizontal weight and 2530 on both the north and south vertical weights.

On the 1.3-meter the 8K dark slide must be removed before mounting the 8K on the telescope. Once the dark slide is removed, it may be necessary to tape closed the slot into which the dark slide usually fits, in order to prevent light leaks. The secondary baffle of the 1.3-meter should be covered when no observations are underway in order to keep dust out of the 8K. There is a red cover provided which fits over the central baffle chimney. The telescope is balanced by bolting one large weight on each side of the east truss (for a total of 2), and two small weights on the bottom ring in the #1 position (due south). The north and south DEC weight setting is 1125, and the RA weight setting is 750.

The power strip 120VAC should be the only power connection required on the instrument. (If the system has been de-cabled, do the following: CCD power supplies should be connected to MASTER and SLAVE controller boxes via the black MS-connector cables. Connect MASTER/SLAVE clock and sync SMB coax cables. Connect ACE to MASTER and to FW connector and shutter connector. Connect ACE power supply Molex connector.) The two fiber optic cable pairs should be connected between the acquisition computer "etna" and the MASTER/SLAVE connections if the identities of the MASTER versus SLAVE jacks on the back of the computer are known. If not, connecting them in the proper permutation will be described below, in which case one starts with the SLAVE fiber unplugged.

CAUTION: do not strain the fiber cables, either by pulling on their ends, by bending them radically, or by crushing them with heavy objects. In some instances, lengths of fiber cable may even stretch across the floor; be especially careful not to crush these.

The data and software are actually resident on the computers "vesuvius" (at the 2.4-meter) and "etna" (at the 1.3-meter), but the 8K control program DetCom can be started from the telescope control computers "hiltner" or "mcgraw" by starting a window there, as follows: 1) with the mouse pointer in the background field on hiltner or mcgraw, hold the left mouse button down until the "Main Menu" appears, then slide the mouse pointer to select "Data Acquisition" and "MDM8K". 2) At this point an xterm window labeled "/home/visitor" should appear, and you should enter the command mdm8k to cause a DetCom window to appear on hiltner at the 2.4m (or, alternatively "mdm8k agung") or on mcgraw at the 1.3m. If you need to turn on the acquisition computer "etna", switch on all peripherals first, then the computer.

The "detcom" window will open and various messages in various colors will flash on the screen. The process should start automatically in the directory /data/vesuvius/visitor (or /data/etna/visitor), which is a good place to store your data, with up to 70 Gigabytes free. There will be various complaints about lack of communication with the controllers (because they are not powered on yet). If, however, the CCD is powered on and fiber communication is already properly established, you may see no error messages. Some error messages in red may appear saying that the 1st and 2nd timing or utility board TDL failed plus additional warnings in yellow, but these are probably safely ignored unless they persist beyond startup (see next two paragraphs). You should also start an xterm window using the Terminal icon (in the middle of the right edge of the screen) in order to keep track of files produced by DetCom. Change directories (cd) in this window to the directory you intend to use to store images if not the default /data/vesuvius/visitor or /data/etna/visitor.

To turn on the CCD power supplies: turn on the MASTER first, then the SLAVE. The switches are on the grey controller power supplies. Three green lights should be visible through the window on the back of the CCD controller boxes.

Hit <return> at the DetCom prompt. Any time you hit <return>, the program will attempt to communicate with the controllers. If the fiber optic cables are properly connected, you will get a green message indicating that all boards have passed the link test, hence communications are OK between the CCD and computer. If you persist in getting a different error message in red, there is most likely a problem with the fiber optic cables: click here for instructions to fix it, or you may need to reset the controllers (using the black pushbutton on the grey controller power supply box---choose the appropriate one: master or slave). For now you can ignore the yellow warning messages: ACE not connected? No heartbeat detected and CCD temperature is out of normal range.

At this point or later one should start communications between the 8K and the Telescope Control System (TCS). If there is no xtcs window running on the computer console, move the mouse so that the cursor is on a blank field, then pull down the Telescope Control menu and select the "xtcs" option with a left mouse button release. (Alternatively, type xtcs4 & in an xterm window.) When the xtcs window appears, click on "INITIALIZE". Next, move back to the DetCom window and type tcs on. At any time you can check the link by typing tcs status in the DetCom window. If all is well, you will receive a printed summary of the current TCS info on the DetCom screen. If not, you have some options. If the 8K/TCS link is dropped (or you forgot to start it), you will be warned in the DetCom window when it tries to query the TCS before an exposure starts. Once the exposure starts, you have no choice but to wait until readout to disk before starting the 8K/TCS link with "tcs on". To disable the 8K/TCS link, type tcs off in the DetCom window. The operation of the 8K/TCS link, including other tcs commands, is described elsewhere by Rick Pogge.

Power on the controller high voltages by typing pon at the detcom prompt. Near the bottom of the DetCom window you should see two rows of voltages that read approximately 35, 16, -16 and 5 volts. A bank of 6 green lights (and one yellow light) should be visible through the window in the back of each of the two controller boxes.

MDM staff should connect the shutter power cable when installing the instrument. If at startup (or any other time) the shutter starts making a prolonged, chattering noise for more than about 15 seconds, unplug the shutter power (by pressing the clip on the bottom of the white, 4-pin connector with two red and black wires) and pulling) and then plug it back in. If the shutter chatter still continues for more than 15 seconds, it is necessary to manually reset the shutter.

Next, in the DetCom window perform a short test exposure e.g.

filename test
etype o
etime 10
raster full 8

The shutter should make a quiet, humming noise, followed by another, 10 seconds later.

IN THE UNLIKELY EVENT THAT THIS TEST EXPOSURE DOES NOT WORK or if exposures begin to fail during a run (also unlikely), try the rescue procedures. IF ALL THREE OF THESE PROCEDURES FAIL, GET HELP.

You can now operate the CCDs and begin a normal observing session. There are three DetCom startup commands the observer must consider: (Remember, "command help" will get you a help entry for that command.)

cd - change directory into area where images will be stored (currently /data/vesuvius/visitor or /data/etna/visitor by default).

mef <on/off> - sets multi-extension FITS format versus 8 single FITS files (stored in a single subdirectory with name determined by "filename").

flush - flushes accumulated CCD charge manually, should be executed 5-10 times at start of run or after exposure to bright light. (Note that the up-arrow will give you a DetCom command history to repeat previous commands, and that you can type commands ahead.)


Before you can observe with the 2.4-meter/8K, retract the dark slide near the top of the 8K and either clamp its outer edge so that it rests near the outer diameter of the telescope rotator ring, or remove the dark slide completely (which may require rotating it 90 degrees as it reaches the end of its travel upon being retracted). On the 1.3-meter, simply remove the cover bag over the baffle chimney.

At either telescope, the observer can select between two similar guide cameras: #1 (south of CCD field in the usual orientation), and #2 (north of CCD). These are intensified CCD cameras AND SHOULD NOT BE SUBJECTED TO BRIGHT LIGHTS WHEN TURNED ON. Avoid illuminating them brightly even when turned off. At the 2.4-meter, the guider camera is selected by a switch on the guide camera control box marked "1", "OFF" and "2" to select which guide camera is powered on, AFTER the power to the box is turned on with the red, square button. One most also select the camera output on the box: "FINDER" for camera #1 and "GUIDER" for camera #2, using the black and silver knob. At the 1.3-meter, the control panel is still under construction - consult with the staff for the current configuration. (Currently, the camera power and signal co-ax must be unplugged and replugged at the telescope in order to switch cameras at the 1.3-meter.) After turning on a camera, type "tvguider" to start the guider program. If the tvguider program complains repeatedly "No camera present - attach camera or hit any key to exit", this means that the power/output selection on the guide camera control box are incorrect (or the camera power or signal co-ax are not connected). The default timestep and movement parameters should work, although they might be adjusted to your situation. The usual rotation angle is 270 degrees for camera #1 and 90 for camera #2, and there is a reflection. Note that the orientation on the guide camera video monitor will be rotated by 180 degrees when switching between cameras.

Here are some reasonable parameter values:
1.00 (seconds, integration time)
2.00 (r.a. guide rate)
1.00 (dec. guide rate)
0.2 (fraction of guide to perform)
0.2 (minimum guide motion)
0 (off slit)
270 (degrees, rotator angle) - for camera #1; 90 for camera #2
300 (counts, minimum signal)
1 (auto sky level)

Now follow the usual PC Guider procedures. The focus of the 8K and guide cameras are mechanically fixed. If you feel you MUST adjust the focus, use the knob on each guide camera housing around the threaded rod extending from each guider housing. You must view the guide camera output on the monitor in the dome, or by having someone else focus while you watch the monitor in the control room. Don't forget to turn off the camera power before dawn, by flipping the switch to OFF on the guide control control box in the 2.4-meter control room.


At dawn, when lights are on in the dome, or any time observations are suspended for long, be sure to power down the 8K guider cameras flipping the switch to OFF on the power control panel in the 2.4-meter control room or the corresponding switch at the 1.3-meter.

If you are not taking calibration exposures when not observing for long e.g., during the day, it is best to power off the CCD high voltages. This is also required during lightning shutdown. Type pof at the detcom prompt. This disables the CCD high voltages and brings things down gently. (It may be necessary to type pof twice to actually refresh the voltage readout, which now should give values all within 0.1 volt of zero.) You should now end the DetCom session by typing detach at the detcom prompt. This will allow for a fresh start of DetCom (including reloading of the DSP code to the instrument) when you return to use the 8K.

If on the 2.4-meter, insert the dark slide (or cover the baffle chimney on the 1.3-meter) and unplug the guide camera power, if you have not done so already. Refill the liquid nitrogen, if need be. (NOTE: if the dark slide becomes stuck, it is because someone re-inserted it at a crooked angle, left-to-right. The best thing to do is slide it back out and try again, more carefully. If it is too stuck to retract, try vibrating it up and down by "sproinging" it with your finger. Anything more violent will just make the problem worse. Pushing the slide in firmly during the last centimeter of travel will assure it stays inserted until someone retracts it again.)

At this point, the 8K can wait until the next night's use, when you can restart the system by initiating a new DetCom session. In the event of a longer hiatus, or if lightning threatens, you should turn off all power as follows. Check that only three of the seven lights (in a line) are on in each controller window (as they should be after a "pof" in DetCom). Turn off the CCD power supplies, SLAVE first, then MASTER.

Turn off the Granville-Phillips power supply and unplug the 8K power strip only if lightning threatens, or if the instrument is being removed from the telescope.

If you need to end the login session of the user "obs24m" or "obs13m", end the DetCom session by typing detach at the detcom prompt. After terminating any other processes that should not be interrupted abruptly, exit the window system (and log out) by pushing the left mouse button while in the background field of the monitor, simultaneously moving to the "Quit session" option at the bottom of the first menu that pops up. To log out completely as "visitor", type exit at the console prompt after the windows system disappears.

To shut down the data acquisition computer, you can simply check that there are no disk processes or other programs running that should not be interrupted. Login as user "shutdown" and use the prominently posted password, then turn off the computer, followed by the peripherals starting with the disk drives. If lightning threatens, unplug all peripherals and the computer. Leave the fibers alone.