1200/7500 Gold 831/8200 Gold 900/5500 Al 600/7500 Al Also, there is a 1200/7500 Al grating in Santa Cruz which is NOT mounted
Please note that, due to the large acceptance angle of the camera, all gratings may produce zeroth-order ghosts at some grating tilts. The 831 and 900 gmm gratings are particularly prone to this at typical grating angles.
The following limits in mask design keep the mill out of the "forbidden zones" around the edge of the mask. Note that this is for the slit corners. Violations of hard limits result in the slit being rejected, ie., not milled.
X-limits: -502 to +502 (arcsec) Y-limits: 186.8 to 479.6 (arcsec) Radius: 600 (arcsec) Diagonal corner: X(arcsec) <= -0.98273 * Y(arcsec) + 833.0(*) (*) approx; it would be best to remove another, say, 2.5 arcsec from this constant as the mask form probably occults to this point anyway.NB: There's currently an ambiguity about the relief cut around the slitlet. This currently is _included_ in these limits, but as its size is not known to the user a priori, it makes life difficult. I am arguing the relief cut is under the control of the milling folk, and hence should be handled by them, not the mask designers.
Slider 3 (2002/04/23): G3TLTRAW = 2500 * angle - 28906 Slider 3 (2002/04/28): G3TLTRAW = 2500 * angle - 29094 Slider 4 (2002/04/28): G4TLTRAW = 2500 * angle - 40934PRE-SHIP VALUES (obsolete)
Slider 3 (08/26): G3TLTRAW = 2400 * angle - 41020 Slider 4 (08/06): G4TLTRAW = 2500 * angle - 22840
Angle Lambda (A) Anamorph. graphical form 2 5879 0.706 3 6127 0.693 4 6372 0.681 5 6615 0.668 6 6857 0.656 7 7096 0.643 8 7333 0.631 9 7568 0.618 10 7800 0.606 11 8030 0.593 12 8258 0.580 13 8483 0.567 14 8706 0.555 15 8926 0.542 16 9143 0.528
Angle Lambda (A) Anamorph. graphical form -8 4429 0.831 -7 4778 0.818 -6 5126 0.805 -5 5471 0.793 -4 5815 0.780 -3 6158 0.768 -2 6498 0.755 -1 6837 0.743 0 7173 0.730 1 7507 0.718 2 7839 0.706 3 8169 0.693 4 8496 0.681 5 8820 0.668 6 9142 0.656 7 9461 0.643 8 9777 0.631 9 10090 0.618
Angle Lambda (A) Anamorph. graphical form -6 5552 0.806 -5 5927 0.793 -4 6300 0.780 -3 6671 0.768 -2 7039 0.755 -1 7406 0.743 0 7770 0.730 1 8133 0.718 2 8492 0.706 3 8849 0.693 4 9203 0.681 5 9555 0.668 6 9903 0.656 7 10249 0.643
Angle Lambda (A) Anamorph. graphical form -12 4533 0.883 -11 5063 0.870 -10 5592 0.857 -9 6119 0.844 -8 6644 0.831 -7 7167 0.818 -6 7688 0.805 -5 8207 0.793 -4 8723 0.780 -3 9237 0.768 -2 9747 0.755 -1 10255 0.743
CCD# Xnom delX Ynom delY Theta 1 -3203.8 -13.42 -2056. -17.67 0.2367 2 -1067.6 -8.75 -2056. -19.88 0.1476 3 1067.6 0 -2056. 0 0 4 3203.8 13.70 -2056. 12.74 -0.1173 5 -3203.8 -34.09 2056. 12.22 0.1493 6 -1067.6 -14.91 2056. 4.26 0.2574 7 1067.6 -23.59 2056. 24.47 0.0380 8 3203.8 -10.58 2056. 46.50 -0.0107
CCD# Xnom delX Ynom delY Theta 1 -3203.8 -20.05 -2056. 14.12 -0.082 2 -1067.6 -12.64 -2056. 7.25 0.030 3 1067.6 0 -2056. 0 0 4 3203.8 -1.34 -2056. -19.92 -0.1206 5 -3203.8 -19.02 2056. 16.46 0.136 6 -1067.6 -9.65 2056. 8.95 -0.060 7 1067.6 1.88 2056. 1.02 -0.019 8 3203.8 4.81 2056. -24.01 -0.082Note that while the rotations, x- and y-spacings are close to nominal, the y-centers move systematically to form a slight stair-step pattern.
name = ABCDE
where A = a (pre-) or b (post-grating);
B = m (mask) or s (sky);
C = f (forward) or r (reverse) mapping;
D = 2, 3, 4 (slider)
E = "grid" (pairs of points) or "map" (fit-description)
In addition, the following info is now (2003mar05) valid:
Grating/slider gmm "roll" "o3" "mu" Mirror/2 - 0. 0. -19.423 600ZD/3 600 0.145 -0.008 G3TLTVAL * (1-5.6e-4) + -0.182 831G /3 831.90 ? ? [G3TLTVAL * (1-5.6e-4) + ? ] 900ZD/3 900 0.141 0.008 G3TLTVAL * (1-5.6e-4) + -0.134 1200G /3 1200.06 0.145 0.055 G3TLTVAL * (1-5.6e-4) + -0.181 600ZD/4 600 ? ? [G4TLTVAL * (1-6.9e-4) + ? ] 831G /4 831.90 -0.034 0.060 G4TLTVAL * (1-6.9e-4) + -0.196 900ZD/4 900 -0.064 0.083 G4TLTVAL * (1-6.9e-4) + -0.277 1200G /4 1200.06 -0.052 0.122 G4TLTVAL * (1-6.9e-4) + -0.294A few notes:
I am informed that average grating spacing tolerance is 0.01%, and I have taken the liberty of adjusting some grating spacings slightly to produce better fits. Also, the 831G and 1200G gratings are BK-7 blanks, which will shrink slightly (I've assumed a temp difference of 17C). This is why the two gold gratings have slightly different spacings than nominal (830.77 and 1200)
Why there's a slight correction to the GxTLTVAL (besides the zeropoint offset) remains a huge mystery.
The "roll" and "o3" values for the gratings are now approximately constant, because the grating mis-alignment has been split into two parts, one for the grating itself and one for the mis-alignment of the axle on the tilt stage. This latter mis-alignment has been incorporated into the mappings, which is why there are now individual mappings for each slider. It is reasonable that the "roll" and "o3" values vary from slider-to-slider, but we would expect the differences to be similar (eg delta-roll between the 900 and 1200 gratings should be similar in both slider 3 and 4, etc.). This appears to be true, and thus the two missing values (831/3 and 600/4) can be extrapolated until measured. [For anyone who cares, this is what I found for the slider/axle misalignments:
Slider theta-y theta-z 2 -0.059 - 3 0.073 -0.057 4 -0.120 -0.116where values are in degrees.]
The optical model is still evolving and enhancements are likely to take place occassionally.
Sense & Sensitivities: To understand the effect of the numbers above, keep in mind that "roll"/theta-y moves images in x and "o3"/theta-z produces a skew in x, whereas grating tilt moves images in y. The order of magnitudes shifts are 0.001-deg / pix for tilt and roll. Theta-z is harder to interpret, but 0.01-deg should cause a 1 pixel shift of top relative to bottom.
NB: The neon lamp does not produce acceptable flats in short exposures (eg. 25ms) -- apparently there are waves in the plasma that produce spatially non-uniform illumination at the level of +/- 2%.
Normalized surface fits to a flat field, illuminated by light on the garage door, are expected to be close to the actual flat-field shape. Note that the slitmask has response around 25-30% of the (peak) mirror in the vicinity of the longslit. However, also note that this particular slitmask was badly scratched and so may not be a good representation. Also note that this data was collected Oct.27 before good TV alignment!
Plate Scale: Images taken of the TV grid-of-holes mask indicate the plate scale is 0.205 arcsec/pix. The spacing of the grid marks on this mask is 12 arcsec.
B 100s V 30s R 10s I 5s Z 8s
During the October 2002 DEIMOS Commissioning run, autoguider centroids were monitored during some exposures in order to characterize typical guiding errors. The results were less than encouraging: guider centroids were often off at a level of 0.1--0.2 arcseconds (compared with the spec, which is 0.05 arcsec for tracking). When added in quadrature with the typical seeing, these do not represent a huge degradation in image quality (although the degradation is significant). However, they are worrisome with respect to alignment, particularly if the guide star is faint and therefore integration times are large (ie not small compared to the alignment image integration time).
Listed below are the 4 sets of data (links lead to PS plots):
Perhaps of interest in these plots is
TV alignment notes: Empirical leveling of detector: 1. Piston edge of pickoff mirror to (....) above the mask form along the edge where the slitmask goes. NB: This is 0.8 mm above the correct position! 2. Tilt PXL to bring the inner edge of mirror in focus while slitmask is in focus at 70mm distance from the inner edge (ie where mirror abuts). -- to adjust tilt in TV-Y, rotated the mounting plate wrt z-axis (two "chicago'd" bolts on nose plate; there's now an access hole to the second one). -- to adjust tilt in TV-X, invent something because there doesn't seem to be an adjustment for this one. -- The easiest way to see features on slitmask is with side illumination, bringing the surface into high relief. Normal Steps: 1. Piston edge of mirror to correct height above mask form (.....) 2. Adjust fold mirror so that TV sees the inner edge of pickoff mirror at Y=675 px. 3. Align DEIMOS star lasers (one center, one edge) to fall on slitmask at 4.5 arcmin level (67mm from edge) at center of TV field. PLace paper over Canon lens cover. Rotate DEIMOS star to trace out edges of pupil, and confirm that the camera is receiving all the light. (NB yckoff mirror should be covered over here. The laser "spot" is highly elongated in the axis where there's no focusing by the cylindrical slitmask, but there seems to be a bright spot toward the middle.) 4. Remove slitmask. Point laser at center of pickoff mirror. Tilt outer edge of slitmask to place bean into camera (rotate DEIMOS star to trace out edges as above -- these spots are well focused). 5. Rotate PXL head to remove field rotation.
DEIMOS BACKUP PROGRAMS (draft):
Hello Fellow Deepsters, Following up on Tuesday's telecon, I'm gathering ideas for backup programs for conditions of: * Poor transparency * Bright sky/moon * Poor seeing (extended objects less severely affected) * Twilight (ie bright sky, short exposures) I envision these programs falling into 3 broad catagories: 1. DEEP support programs -- ie programs that directly support the DEEP 1HS and 3HS observations and/or science; 2. DEEP supplemental programs -- ie programs that provide supplemental data (eg imaging in additional passbands) or complementary data (eg redshift survey at lower than target z's) 3. Non-DEEP programs -- ie, programs by individuals that have no direct relevance to DEEP. (Whether we allow and how we decide which of these programs to be run on DEEP observing time is TBD). I offer some suggestions here, and solicit others: * Emission line velocity studies (eg use wide and narrow slits on extended, lower-z disks to help understand integrated line-widths at higher z) * Redshifts of brighter/larger galaxies (we can use short exposures and slitmasks with "sloppy" alignment -- might be a twilight program) * Try to identify "missed" redshifts using lower-dispersion grating for increased spectral coverage. * (Brighter) QSO targets, different spectral range than survey observations. * Search for / survey of low-luminosity, low-z star-forming galaxies (may be difficult to select via photo-Z with our passbands). * Survey of CNELG candidates. If it appeals to people, I might suggest that some portion (10%? 20%?) of the "back-up" time be set aside for Category 3 projects of interest to "junior" team members (eg. postdocs) and/or people who actually go observing. Such programs should probably be run through an internal TAC. Please send feedback, and I'll do what I can to collate and distill ideas for the Team meeting. Thanks! drew
Andrew C. Phillips / Lick Observatory
phillips@ucolick.org