![[Eric Deutsch Home]](/deutsch/images/cluster.gif)
On June 16 the SP300i spectrograph and Mustang CCD camera saw first light through the MRO 30'' telescope. Although the weather was mostly cloudy, some brief episodes of clearing allowed some test exposures. In attendance was Jeff Morgan, David Trowbridge, Chris Stubbs, Michael Braunstein, ???, and Eric Deutsch. Many images and spectra of various lamps, the twilight sky, and Vega were obtained; only a few of the highlights are shown here.
In the afternoon we experimented with the spectrograph detector focus and alignment. The initial alignment allowed the top of the CCD to be in reasonably good focus while the bottom was in very poor focus as seen in the following image of a Neon arc spectrum.
The alignment was then adjusted so that the center of the CCD was in better focus with the top and bottom edges being is poorer focus as seen in the following Neon arc spectrum. Hopefully an even better focus and alignment can be acheived in the final configuration. It does seem likely, however, that when the central region and fibers are reasonably sharp, the outer edges and fibers will not be.
When we finally got on the sky in a cloud break, we placed Vega in the spectrograph pickoff mirror, as here imaged by Tbird. The diffraction spikes are from Vega itself. For various reasons the following image is not precisely the one that corresponds to later spectra. The FWHM here is about 3 arcsec and the tbird image was obtained in 2x2 bin mode at 1.093 arcsec per pixel.
With Vega on the fibers, we took a 5 second exposure with Mustang and obtained the resulting image with all the fiber spectra. Each of the fiber numbers are labeled at the bottom.
A cut across this image looks as follows. The dotted vertical lines represent the positions of each of the 24 fibers:
Here is a map of the relative brightnesses of these fiber spectra on the pattern of the fiber head. In the following figure, the fiber with the most mount of light is denoted as fully filled in with black. All the other fibers are partially filled with a spot whose area is proportional to the amount of light seen in that fiber relative to the brightest one. The fraction of light in the various fibers relative to the total amount of light collected in all fibers is as follows; light loss between the fibers is not included:
Fiber #: 19 9 21 16 15 8 18 12 10 11 7 17 23 1 % Light: 47.7 13.4 8.2 7.5 6.1 5.8 3.7 3.1 1.0 0.6 0.5 0.4 0.4 0.3
And finally, here's the spectrum of Vega as obtained from this 5 second exposure. This is an extraction of the brightest fiber (19) only; no attempt to coadd the other fibers has been made.
A back-of-the-envelope calculation provides some further insight into the expected performance of this system. An extraction of one of the weaker fibers provides the following spectrum:
The estimated signal to noise here is about 15 at 5000 Angstroms. In the best best fiber of this 5 second exposure, there were about 67,000 total counts at 5000 Angstroms, while in this S/N~15 spectrum, there are about 177 total counts at 5000 Angstroms. Therefore, one might expect to obtain a S/N~15 spectrum of a V=6.5 A0 star in 5 seconds, and thus such a spectrum of a V=11.5 A0 star in a 10 minute exposure. The spectra analyzed here were taken in Mustang's 2x2 bin mode, yielding 1.5 Angstroms per pixel. The FWHM was about 3.2 arcseconds and the telescope focus was clearly not optimal at the time. Guiding was on and apparently working well. This estimate does not take into account any dark current or charged particle hit effects. The spectrograph was also not optimally focused and aligned. Nor was any attempt made to apply a flat field to the CCD. Nonetheless, the approximate expected depth is evident; this estimate can easily be taken as a very conservative limit to the throughput of the system.