Previous: TV Focus
Up: Observing with one or two CCDs
Next: Acquiring data on the sky
Previous Page: TV Focus
Next Page: Acquiring data on the sky

Instrument Calibration Data

The standard observing mode with ISIS is to observe with one or more CCDs. Decisions which have to be made before the beginning of an observing run include:--

• The grating or gratings to be used, and the central wavelengths they will be used at.
• The dichroic to be used if both arms are required. Plots of the reflection and transmission curves of the mounted dichroics should be available in a folder in the control room.
• The size of the data window. The ISIS slit illuminates about 700 pixels on the spatial direction on an EEV CCD; slightly less on the Tektronix CCD. Windowing the chip saves readout time, and observers should not normally use more pixels than the slit illuminates. Many observers, particularly those interested in point sources, will require much less of the slit, and in many cases a hundred pixels (33 arcsec) will suffice.
• Whether on-chip binning is required to reduce the effect of readout noise. it is possible using the BIN procedure to set any binning factors in either direction; binning factors of up to 4 in each direction are normal.

Procedures for configuring the spectrograph for the beginning of observing are described in Part iv. The centre of the slit on the CCD can be determined by taking a short exposure of a tungsten lamp through the narrow dekker (dekker position 1); and measuring the centre in pixels with the DMS command Y-FIND. Before beginning to take exposures with a CCD it is necessary to set the CCD up with the ICL command SETUP <ccdname>. This will prompt the user for the size of the CCD data array; this includes underscan and overscan regions and is 400 by 590 for the FOS and GEC5 CCDs; 1280 by 1180 for EEV P88300s; and 1124 by 1124 for the Tektronix CCD. If windowing or binning is required this can then be set up with the ICL procedures WINDOW and BIN. There are a number of faults that can occur when reading CCD data into the DMS buffers:

• If the CCD fibre optic interface card (at the top of the second blue cabinet from the left hand end of the control desk) displays a light labelled G.O.A.T. then there is probably a fault with an optical fibre cable. A duty tech should be called.
• If there is an error light on the DICI card; or a FIFO FULL light on the CCD fibre optic interface card; then the DICI should be reset and the exposure repeated.
• If the data are clearly scrambled; or show a series of diagonal stripes then the CCD controller thinks the data frame is a different size to the DMS. The DMS should be reset, and the SETUP and WINDOW commands should be repeated at the ICL interface FOR EVERY CCD IN USE.
• If the previous procedure does not work then the CCD controller should be rebooted (with the REBOOT command rather than switching it off and on) and the procedure repeated.
• If the CCD shutter fails to open this will send a message to the ICL interface. This could be due to a fault on a shutter cable; to low nitrogen pressure to the telescope; or to a mechanical failure of the shutter. However before calling for technical assistance it is worth typing UNJAM at the relevant CCD controller to see whether this cures the fault.

Once the CCD has been set up to obtain images of the correct format, and the spectrograph has been set up correctly then the observer can begin taking calibration exposures. Normally the following calibration exposures are required:-

• A number of ``Bias'' frames, which are dark exposures with zero exposure time. These can be taken with the BIAS command. These are often not required because the bias is structureless, and its absolute level can be determined from the overscan and underscan regions associated with each exposure.
• A single long dark exposure, which can also be taken with the DARK command. The dark count is very small, and a dark frame is usually used to measure the cosmic ray rate.

  If the ICL interface is not running, dark frames can be taken with the following procedure:--

  1. Set up the correct buffer with the correct size in the DMS with one of the commands CRUN, GRUN or OPN; if the latter then the correct arguments must be used.
  2. At the CCD controller invoke the Network vocabulary with the command:-
     NETWORK.
  3. At the CCD controller set the exposure time with the command:-
     

     <headno> <exposuretime> SEX,
     where the time is in milliseconds.

  4. Begin the dark exposure with the command <headno> 0 DAR. Note that although the network command to take a dark frame is DRK, the CCD controller command is DAR.
  5. The exposure will read out into the DMS buffer, and can be stored with KEEP from ICL.

  It is very difficult to setup windows or binning without access to the ICL interface.

• A number of flat fields taken with the Tungsten lamp. The grating can be moved to give uniform illumination along the spectrum, because the tungsten flatfield should not be used to take out large scale sensitivity variations. It cannot be used to take out the large scale variations along the slit because the illumination of the slit by the calibration optics is not exactly the same as the illumination by the sky. A flatfield taken with a wide slit can be used for correcting for small scale sensitivity variations along either dimension of the CCD. A flatfield taken with a narrow slit will contain additional small scale variations of illumination in the spatial direction, which will appear as horizontal stripes on the image. These are due to variations in the slit width; the slit jaws are not precisely straight and parallel.

  To obtain flatfield exposures the following steps must be followed:-

  1. Put the acquisition and comparison mirror into the beam with the ICL command AGCOMP or AGMIRROR ACQCOMP.
  2. Turn on the tungsten lamp with the ICL command COMPLAMPS W.
  3. Adjust the comparison neutral density filters with the ICL commands COMPFILTA and COMPFILTB.
  4. Turn out the lights in the dome if possible.
  5. Take trial exposures with GLANCE and adjust the exposure time and filters to give a maximum level in the region of 20000 - 40000 ADU.
  6. Take the final exposures with RUN or FLAT or keep the trials if the level is right.

• An exposure of the twilight sky, with the spectrograph set up precisely as it is intended to use it during the night. In particular the slit width should be the same as that intended for the night-time observations. Twilight sky exposures can be taken about 20 minutes after sunset, with the telescope pointing at the zenith. They are used for calibrating the sensitivity variations in the spatial direction.

  To take twilight sky exposures:-

  1. Take the acquisition and comparison mirror out of the beam if it is in with the ICL command AGSLIT or AGMIRROR SLITVIEW or AGMIRROR OUT. If a comparison lamp is on you will need to turn it off with COMPLAMPS OFF first.
  2. Verify that the dome and primary mirror covers are open.
  3. Take trial exposures with GLANCE and adjust the exposure time and main filter slide filters, with the ICL command MAINFILTND to give a maximum level of no more than 40000 ADU. The twilight sky exposures do not require a very high signal level, because they will be collapsed in the spectral direction before being used anyway.
  4. Take the final exposures or keep the trials if the level is right.

• Exposures of comparison sources (usually copper-argon or copper-neon) interspersed with observations on the sky. The comparison lamps are really rather faint in the blue, and exposure times of between one and five minutes, depending upon the dispersion and slit width, will be required to give adequate signal in the blue. The copper argon lamp usually gives the most usable lines here.

  At wavelengths longer than 5850 Å the copper-neon lamp has a number of strong clean lines, and at wavelengths longer than 6965 Å the copper-argon lamp does too. Thus it is usually possible to obtain a high signal-to-noise comparison spectrum, containing a number of good lines, using one or both of these lamps in an integration time of between 5 and 30 seconds.

  When using both arms it is possible to make the best of the disparity in the required exposure times for comparison lamps in the blue and red channels by completing the entire red exposure, including clearing, exposing and reading out the CCD and keeping the file to the Vax disc while the blue CCD is integrating. At the ICL interface:-

  1. Ensure that the acquisition and comparison mirror is in, if it is not put it in by typing AGCOMP.
  2. Switch on the required comparison lamp(s) with the command COMPLAMPS.
  3. Start the blue exposure with ARC BLUE <exposuretime>, where the exposure time is 100-300 seconds.
  4. When the packet collection is complete start the red exposure with:-
     ARC RED <exposuretime>,
     where the exposure time is in the range 10-30 seconds.
  5. Wait for the red exposure to finish reading out; when it has finished the icon for the red CCD frame will appear on the DMS display screen; and the message ``PICTURE action complete'' will appear at the ICL interface.
  6. Wait for the blue frame to finish reading out; again waiting for the icon to be redrawn and the ``PICTURE action complete'' message to appear at the ICL interface.

  If the difference between the two exposure times is less than about 60 seconds then the blue frame will begin reading out before the transfer of the red data to disc has taken place. This is not a problem at all; the DMS is capable of having data written to one buffer by a CCD controller whilst data is being read from another.