Up: Rotation, tilt and focus
Previous Page: Rotation
Next Page: Tilt
The best method for determining focus is to use Hartmann masks. These work by blocking off alternate halves of the pupil. If the optics are perfectly focussed onto the detector, the image of a point source will be the same when either Hartmann mask of a complementary pair is employed. The size of the relative image shift between each Hartmann pair is a measure of how far away the best focus is. A number of routines are available on the DMS to calculate the Hartmann shifts from pairs of exposures employing alternate Hartmann masks.
LDSS-2 is equipped with Hartmann masks located in the filter and grism wheels (Table 8). They have the same mounting arrangement as the filters so that they can be mounted anywhere in the filter wheel or in the round holes in the grism wheel.
Up/down etc. refers to the half of the aperture which is obscured in the coordinate system of the detector. Up = positive Y direction, Left = negative X-direction etc. (+X is the red end, +Y is north on the sky for PA=0). See Fig. 6 for details of the coordinate system.
Although the Hartmann masks may be mounted in any available position within the filter and/or grism wheel, there is a particular arrangement which is strongly recommended. HL and HR should be mounted in the grism wheel so that they can be used to provide information in the dispersion direction for a particular filter. If desired, HU and HD could also be mounted in the filter wheel to provide information in the cross-dispersion direction in conjunction with a pin-hole mask and grism.
To do Hartmann tests for focus determination in the dispersion direction, you will need to use the focus mask (Section 1.2) which consists of three parallel long slits. This should be illuminated with white light from the calibration system with appropriate neutral density obtained using the ND filters in the comparison system. Note that some DMS routines are confused by pixel values >32000, so ensure that the maximum count is less than this in your focus images. Typical exposures through a ND5 filter should be roughly 10s in V, and 5s in R.
The DMS focus routines require sampling regions to be defined within each image. The regions are set up with the following DMS commands, typed in the DMS window on the Sparcstation:
25 50 75 VERT 1 SET-LINEThe first command defines three positions in the Y-direction (i.e. along the slit) where the direct image of the slits will be sampled. On Fig. 9 these positions are labelled A,B, and C. They should be well spaced to sample the full length of the slits symmetrically about the fields centre. In the example given these positions are 25%, 50% and 75% of the extent of the detector in the Y-direction which suits the length of the slits quite well. The second command allows the user to interactively define where the three slits are in the X-direction.
Make sure that the exposure in the DMS was taken with the HL mask in beam, and enter the DMS command FOCUS-LEFT. This will determine the centroid of the slits averaging over the X-direction at the specified Y positions. Then put the HR mask in the beam, take another exposure and type FOCUS-RIGHT in the DMS window. This will again determine centroids for the slits and compare to the HL values to give Hartmann shifts for each of the three positions on the three slits. This gives a total of 9 Hartmann shifts which sample the on-axis position and 8 off-axis positions symmetrically disposed around it. Fig. 9 shows the geometry of the situation, and Table 9(a) shows some typical results for the case where both the focus and tilt are wrong. Experience has shown that the Hartmann shifts have an associated uncertainty when measured in this way of about 0.1. Note that you must use the FOCUS command before FOCUS-LEFT and FOCUS-RIGHT for each pair of Hartmann images.