Phillips' DEIMOS Test Pictures

The images on this page show results of initial tests with DEIMOS in Nov/Dec 2000 and subsequently. Watch this space as progress continues.
NB: Final filters and baffling are not in the system presently.

NB: These images should NOT be viewed as final performance images!!

New Figures: Pinhole Shadow tests (08-10 Oct) ; Aligned TV images (06 Nov) ; FCS Scattered Light (07 nov) ; Final Pinhole Condensation Image (30 oct)

NEW! Camera Flexure Report

Perhaps you are looking for the DEIMOS reference page ?

Click on images for PostScript Versions.

click for PS Full 1200-line grating image with HgNeArKrXe lamps on (frame 1910). This is a near best-focus image with the "line-of-slits-and-holes" mask. The slits are 5.5 arcsec long and 0.5 arcsec wide; the holes are 0.5 arcsec square. (12/30)

click for PS Full 600-line grating image with HgNeArKrXe lamps on (frame 1507). This is a near best-focus image with the "line-of-slits-and-holes" mask. The slits are 5.5 arcsec long and 0.5 arcsec wide; the holes are 0.5 arcsec square. (11/21)

click for PS Full 900-line grating image with HgNeArKrXe lamps on (frame 1620). This is a near best-focus image with the "line-of-slits-and-holes" mask. The slits are 5.5 arcsec long and 0.5 arcsec wide; the holes are 0.5 arcsec square. (11/27)

click for PS Two views of a montage of images, taken from frame 1910 (above). In this case, a mask across the collimator produces the proper beam size for the 6 mask holes shown here. Eleven different wavelengths are shown, sampled from along the full spectral image. Each "postage stamp" image has been background-subtracted and flux-normalized. The tails that appear are believed due predominantly to an assembly spacing error between the field flattener and the rest of the camera, which are mounted independently. Such tails were not seen in the COHU test images taken last March, when the camera was in the optical shop testing tunnel (see below). A review of the assembly shows that a spacer was mistakenly omitted when mounting the dewar, to which the field-flattener is attached. This will be remedied when the instrument is reassembled in March.

The field-flattener has since been moved back 0.030in and image quality has improved remarkably ( mosaic and contour plot at 20 px spacings, from 07/24.) It is understood that the camera at room temperature should have somewhat comatic images; recall camera will operate near 0C on Mauna Kea.

click for PS Contour plot of the same spots above (postage stamps are 45x30 pixels in this case to give a more representative aspect ratio). The contours also show more clearly that images from chip 7 (top half of fourth column here) are significantly worse than the others. Chip 7 (third from left top in full mosaic) is an early-manufacture red-sensitive device.

click for PS COHU images produced by a 5-inch collimated beam source and the DEIMOS camera from March 2000. The image scale was magnified by a microscope objective in front of the detector. Each horizontal row is a series through focus at 30-micron spacings. Images were obtained on axis (top) and at a tilt of 10.2(?) degrees off axis at four cardinal rotations of the camera.

click for PS Central portion of frame 1966, showing low-level ghosts and other artifacts. This is a 1200gmm-grating image through my special-purpose calibration mask, which has (1) a slit-hole-slit pattern across where the chip-gaps should be (in proper alignment); and (2) a series of holes across the y-axis width of the mask. Large areas are left blank so that 0th-order ghost images (ie. re-imaged reflections off the CCD itself) can be identified [RED]. In green are shown identified lines of Ne and Ar, arising both from actual holes in the slitmask and from other sources, such as light leaks from around the mask form and possibly from scattering within the instrument. There is no collimator mask and the internal lamps were used here, so the beams are overfilled. One possible internal scattering is a thick controller cable that projects into the light path at present.

click for PS Fringing amplitudes from frame 1938. These are normalized quartz-lamp spectra taken at various positions, corresponding to the collimator mask holes (the first two are from adjacent slitmask holes). Note that fringing in Chips 1 and 4, engineering-grade high-rho devices, is both more rapid and lower in amplitude (about 2%) compared to the standard CCDs (Chips 2 and 3). This is as expected from the difference in thickness. The standard devices have fringing amplitudes up to about 7% beyond 9000A. Caveat: there was no order-separating filter in place, so there may be a small amount of blue light present, which would dilute the fringing; however, even at much longer wavelengths the 2nd-order blue spectrum appears to be very weak.

A Postscript plot of the non-normalized spectra shows the relative sensistives. In the plot note that the spectrum from Chip 3 (blue) passes onto Chip 7, which is an early red-sensitive device; the others all continue onto standard devices. Chips 1 and 2 have significantly higher red sensitivity above 9000A despite their lack of AR-coating.

click for PS Zeroth-order ghost lines compared to primary lines. The primaries are from frame 1948 (1 sec exposure), the ghosts from frame 1947 (60 sec). Both are 1200-line grating images. Spectra from 4 different slitmask holes, close in X by staggered in Y across the mask, are plotted. The ghosts have been scaled by 1000 (that is, 16.67 x 60) relative to the primaries. Ghost intensities appear to be of order 0.1% or less (but I suspect this is a function of grating, as well as precise angles). A different spectral range is also available.

click for PS "First-light" TV Guide Camera image (PXL). The image shows a piece of paper taped to the slitmask. Another annotated picture displays the image at different contrast to show the slitmask and pickoff mirror.

click for PS This shows how we can measure the Z-position of the CCD using a double-pinhole source in front of the camera. Double shadows of dust specks on the front and back of the Dewar window are cast onto the CCD; knowing the thickness and index of refraction of the window, combined with the ratio of the shadow-pair separations, allows us to determine through simple geometry how far behind the window the CCD is located. (The double-pinhole source has 18.8mm separation, and is located approximately 50mm in front of the camera. The "pinholes" are actually screw holes, and so only features very near the CCD cast reasonably sharp shadows.) The PS version includes an image section showing typical features from the front and back of the Dewar window.

click for PS This figure shows the test setup for the pinhole "shadow" tests, and how displacement of an element in the camera will result in different motions for images in collimated light and for shadows of "dust" located on the moving element. We had previously identified several pixels of image motion (flexure) coming from the camera. In these tests, a few "shadows" were noted to have motions of order 20px.
Example of moving spots.

By slightly moving the pinhole source, it was possible to identify which lens surface each "dust" feature is associated with, similar to the picture above. Two moving shadows from Element 3 are associated with a physical displacement of 0.005 inch o that element, which would produce about 6 pixels image motion. Thus, Element 3 is the dominant source of flexure in the camera.

click for PS Typical image from pinhole "shadow" test, identifying various features. Most of the features seen are either dust or condensation on the front of the Dewar window. The majority of the remaining features are on the back of Body 4, which is open to the outside and close to the moving shutter, so we would expect this surface to be relatively dirty. There was no filter in the system during these tests (the filter would be the other relatively dirty surface). A GIF version is also available.

click for PS This pair of images was taken Sept. 13 (top) and Oct. 03 (bottom) with illumination from the pinhole light source. It illustrates some of the temporal changes seen in the Dewar window condensation. Notice the droplet (top) that disappears later on, probably leaving a new "ice disk" in the vicinity. Also, notice how the "frost" at the upper-right grows. Most of these features are located on the front of the Dewar window (see above for how this can be determined). Some "dust" features appear to move, possibly being transported as droplets roll across them. Larger JPG images are also available.

Each image shows the central part of CCD#2 in the mosaic.

Oct. 25 TV tests -- These are "flats" divided by each other: 13/14 ; and 07/10 . The later has a slight (sub-pixel) shift due to flexure, so features appear there that do not appear in the first ratio. Both have had "bias" subtracted before division. The amplitude of the large-scale features is about +/- 2%. (Currently no PS available)

Apparently this is caused by plasma waves in the neon lamp source producing a spatially-variable illumination on the hatch in the 25ms exposure. Flats taken with daylight on the back of the garage door divided perfectly.

click for larger JPG A post-alignment TV Guider image. Small piece of tape are on the pickoff mirror surface, while slits (and scratches) are seen on the slitmask. Adjusting the fold mirror has produced a small rotation of the field, <1 deg, which will be corrected later ( larger JPG). Note the small line and box (3x3mm) near the top center of the slitmask, made by milling partway through the slitmask with a 12thou end mill -- that's 0.4 arcsec on the sky) shown in detail below:

click for larger JPG

click for larger JPG Difference of two 10min "darks" with shutter open but no lights. The first of these "darks" had all FCS lamps on; the second had all lamps off. This difference shows scattered light from the FCS spectra onto the science mosaic. Image is a logarithmic stretch.

click for larger JPG Final "pinhole" image 10/30 (center, 3000x3000 pix). Most of the condensation has now turned to platy ice. Note that the concentric rings on the back of Body 4 are now visible through the ice. The condensation patch is approximately 40mm in diameter.


Upon disassembly, it was found that the "ice" was actually cracking of the AR coat on the front of the Dewar window. Apparently, the water sitting on the AR coat for a long period of time somehow damaged it. Even water drops that did not remain very long produced some damage, hence the "ice disks" mentioned above.
JPG picture of dewar window front upon disassembly; anotated.
Last modified: 07 feb 2002

Andrew C. Phillips / Lick Observatory