|
Support Astronomer Notes
Hartmann test
The best method for determining the spectrograph focus involves the use of Hartmann masks. They work by blocking off alternate halves of the pupil. If the optics is perfectly focussed onto the detector, the image of an arc line will remain stationary when
either Hartmann mask of the complementary pair is employed. In out-of-focus conditions, the shift of the relative peak introduced by each Hartmann mask is a measure of how far away the best focus is (see Fig. 1).
Rotation, tilt, and focus
The optical system and detector must be aligned so that:
- The optimal focus position is within the range of the focus adjustment on the camera.
- The detector is not rotated with respect to the slit and dispersion directions (i.e. the columns/rows of the CCD are aligned with the slit/dispersion direction).
- The detector is not tilted with respect to the focal plane (i.e. the focus is uniform over the full extent of the detector).
To check for rotation, tilt and focus, we use a Hartmann test. At the ING there are some IRAF scripts that help you do so. But, do you know how these scripts work exactly?; and moreover, what do the numbers that the scripts provide mean?
Rotation scripts
The isis_rotation.cl IRAF script extracts the spectra of an arc image in two windows separated spatially and cross-correlates them to find any pixel shift. The shift is then converted into a detector rotation offset in micrometer units.
A common output of the rotation script is:
pixel shifts between top and bottom : 1.08
corresponding angle in degrees : 0.0935
corresponding micrometer offset in mm : -0.282
The pixel shifts are the important numbers!!!! The other two numbers are estimates from a factor conversion and may be ocassionally wrong (e.g. if the cryostat is rotated with respect to the nominal position).
Pixel shifts between top-bottom (or left-right) should be less than 0.5 pixels. But sometimes larger values are also fine depending on how demanding is the science programme. In any case, try to get the best possible value.
Tilt scripts
Two spectra of the arc lamps are the input for the Hartmann test, one with the left Hartmann shutter closed (S1), and the other one with the right Hartmann shutter closed(S2). It does not matter which one is done first. The isis_tilt.cl IRAF script is given the positions of three arc lines uniformly distributed along the detector, and then extracts spectra in nine small windows, three by three in both the spatial and spectral directions across the detector, cross-correlates the spectra corresponding to the two Hartmann exposures, and finally gives the pixel shifts (S1-S2) for the nine positions on the detector. These are then used to compute offsets for the capstan positions.
A common output of the tilt script is:
Hartmann shifts are in pixels :
-1.56 -1.54 -1.69 THESE ARE REALLY
-1.70 -1.69 -1.83 THE IMPORTANT
-2.08 -2.02 -2.08 NUMBERS !!!!!!
corresponding x,y coordinates are :
(103,645) - (592,645) - (840,645)
(103,326) - (592,326) - (840,326)
(103,7) - (592,7) - (840,7)
average Hartmann shift in pixels is : -1.80
top-bottom Hartmann shift in pixels is : 0.46
left-right Hartmann shift in pixels is : 0.09
Hartmann shift converted to tilt angle in degrees:
top-bottom : 0.2279
left-right : 0.0363
to correct for overall Hartmann shift turn capstans as follows
capstan offsets, positive is clockwise, in units of full turns :
A: -0.48 B: -0.48 C: -0.48
to correct for top-bottom Hartmann shift turn capstans as follows
capstan offsets, positive is clock wise, in units of full turns :
A: 0.00 B: 1.16 C: 0.00
to correct for left-right Hartmann shift turn capstans as follows
capstan offsets, positive is clock wise, in units of full turns :
A: -0.14 B: 0.00 C: 0.14
The pixel shifts are the important numbers!!!! The others are estimates from a factor conversion and, as in the case of rotation, may be ocassionally. To correct for detector tilt, I recommend an step-by-step procedure:
First: correct for the overall Hartmann shift. This step may involve just a collimator focus adjustment, so check the focus with the focus IRAF script (isis_focus.cl). In case the new collimator focus is out of the optimal range, you must move the three capstans (same direction and several turns). If not, try with the values recommended by the tilt script. Then take new exposures with the Hartmann shutters alternately closed and run the tilt script again. Check the pixel shifts for the nine windows and compare with the previous values. You can easily see if the capstan offsets were right or not (see the Precision focussing of the spectrograph section to know when you can consider you are in focus).
Second: Correct for the left-right or top-bottom Hartmann shifts depending on which is worst. Always check the pixel shifts (nine points at the starting point of the script output) to be sure you are moving in the right direction!!!!
WARNING!!!: If some of the pixel shifts are absurdly large, you have likely selected the wrong spectral line (few/much counts). Be careful to select apropiate spectral lines and well distributed along the chip.
Precision focussing of the spectrograph
The ideal case is to get (S1-S2)=0 for all the chip. However, in practice we always obtain some small pixel shifts. It has been estimated an overall pixel shift (S1-S2) < 0.5 pixel as a good value. The degradation in resolution increases rapidly for large values of (S1-S2) (see Fig. 2).
SPECIAL CASE: The pixel size of the EEV12 on the blue arm (13.5 microns compared to 24 microns of a TEK CCD) means that a slit width of approximately 1 arcsec will project onto ~4 pixels (FWHM) on the detector. In this case, the worst acceptable pixel shift will be (S1-S2)=1 pixel, but this can be significantly improved after few iterations!
|
