Software Talk (11:20 am)
Click on pictures for PS versions.
Atmospheric Dispersion (arcsec) at Keck, relative to reference position
of 5000 A light. Note how badly the dispersion increases after about 60-deg
zenith distance, nearly doubling at Z = 72-deg.
The refraction values in this figure were calculated using the Starlink library refraction routines, with parameters for Mauna Kea.
|Atmospheric Dispersion (arcsec) at Keck, relative to reference position of 5000 A light. Now plotted vs. airmass. Past Z = 60, the airmass changes very rapidly with increasing zenith distance, so that the airmass is more than doubled by Z = 72-deg.|
|Atmospheric Dispersion (arcsec) with the ADC correction. The current ADC design has only 700mm of travel, corresponding to full correction up to Z = 55-deg. (The correction was calcuated using the Phillips-developed IRAF software for producing ADC spot diagrams.)|
|Different ADC design, rejected due to difficulties. On the left is the crossed prism pair (Risley prism), which produces large and variable tilt to the input beam. An Amici prism is a pair of prisms of different glasses (eg crown and flint), combined to give zero-deviation at some design wavelength but still disperse light. A pair of Amici prisms can be crossed as with the Risley prism set (right) to produce zero-deviation of the input beam. However, it is difficult or impossible to find glasses with both the needed dispersive properties and high tranmission over the range 0.3 to 1.1 micron.|
|Schematic showing conceptually the effects of a tilted beam entering a slit (or arbitrary point in the telescope focal plane). The effects are twofold: (1) the pupil is displaced on the grating; and (2) the beam now runs through the instrument at non-design angles, leading to greater aberrations. Thus, any angular deviation of incoming beams must be small with respect to the beam opening angle.|
|Conceptual diagram of Linear ADC design. The second prism exactly cancels the effects of the first prism, leaving only an offset in the positions of the rays (and thus the images in each color). Note that while this design is zero-deviation in angle, there is a large offset in the position of the reference ray (green).|
|Diagram showing the free parameters for exploring improvements in image quality. In the Nelson and Mast study (general Keck ADC), the prism angle of 5-degrees was assumed, and values of alpha, beta and epsilon coarsely tested. Best results were found with the inner surfaces parallel and perpendicular to the optical axis.|
Spot diagram for current ADC design, full 700mm extension, at 7 arcmin
off-axis. For scale, the box is 1-arcsec in Y. This figure is for PA of the
prisms at 0-degree wrt the field.
Other diagrams are available:
|Plate-scale changes introduced by the ADC. As a quick test, the distance between chief rays and centroids for spots located at 4 arcmin and 10 arcmin off-axis were calculated using spot diagrams (above). As seen in this figure, the maximum deviation from "No ADC" is about +/- 0.1 arcsec over 360 arcmin. NB: This does not include any differential-refraction effects from the atmosphere!|
Andrew C. Phillips / Lick Observatoryphillips@ucolick.org