Previous: Setting up the CCD Detectors
Up: Setting up the CCD Detectors
Next: Tilt and coarse Focus
Previous Page: Setting up the CCD Detectors
Next Page: Tilt and coarse Focus
The long axis of the CCD should be aligned fairly precisely with the dispersion direction of the spectrograph, with care it should be possible to do this to an accuracy of 1 minute of arc, which corresponds to 0.3 pixels slope of the spectrum from one end to the other. The value of doing this alignment well is in the signal-to-noise ratio of the final spectrum when extracting spectra of faint point sources, and in the effective spatial resolution of long slit spectroscopy of extended objects in good seeing. This alignment will probably only have to be done when the cryostat has just been put on the spectrograph, but it should be checked more frequently. As the dispersion direction depends upon the angle of the grooves on the grating this alignment could in principle be grating dependent, although with the current set of gratings it is thought not to be.
To check the alignment take an exposure with GLANCE of the tungsten lamp through the narrow dekker (dekker 1). Filters, exposure time and grating angle should be adjusted to give a fairly uniform count level along the spectrum. The count level should not exceed 32767 anywhere along the spectrum. If the spectrum displayed on the DMS screen is vertical, as it will be for an EEV P88200 chip, the X position of the spectrum should be measured as near as possible to each end with the DMS command X-FIND. If the spectrum on the DMS screen is horizontal, as it will be for an EEV P88300 chip, the Y position of the spectrum should be measured at each end with the DMS command Y-FIND.
At the DMS control window type:--
X-FIND or Y-FIND
Locate the cursor on one end of the spectrum with the DMS mouse, then press the space bar on the keyboard not one of the mouse buttons to accept the cursor position. A profile across the spectrum will be displayed on the DMS display screen, and on the alphanumeric screen the position, which is the raw cursor position in the spectral direction and the centroid in the spatial direction will be displayed. For the centroiding algorithm to work well the peak of the profile should be near the centre of the display, if it is not then press the space bar to return to the main display and repeat the process. This should be repeated at the other end of the spectrum, and the difference between the two centroids noted. If this is more than about 0.3 pixels then the rotation alignment should be adjusted.
The CCD mounting ring is fixed to the mounting ring on the spectrograph by seven Allen screws which pass through slots in the CCD mounting ring into the spectrograph mounting ring. There are two further Allen screws which hold the spectrograph mounting ring to the spectrograph, these lie on a horizontal line through the centre of the mounting ring, and are recessed more deeply than the other screws. These two screws should not be loosened. If the seven screws are loosened the cryostat and mounting ring can be rotated a small amount, the limit is set by the length of the slots in the CCD mounting ring. There is a micrometer gauge fixed to the CCD mounting ring to the left of the cryostat; this can be brought against a block which is fixed to the spectrograph mounting ring to give a measure of the relative position angle of the two rings.
The cryostat should be rotated by backing off the micrometer; loosening the screws; rotating the cryostat manually; tightening at least three of the screws; and then measuring the position angle with the micrometer. The micrometer gauge should not be used to push the cryostat around. The mounting ring on the red arm is stiffer than that on the blue arm, and needs to be pushed around quite hard.
A 1 pixel slope from end to end on the CCD will require a movement of the cryostat position angle equivalent to slightly under 0.2mm on the micrometer. The direction of the slope on the DMS display will depend upon which CCD on-chip amplifier is being used, so this will have to be established by trial and error. Getting the rotation right is largely a matter of trial and error, and can be a time consuming process.
When the rotation is correct all seven screws holding the two rings together should be tightened.