Next: The Gold Spectrograph
Previous: System Overview

2 System Performance

Figure 2 shows the measured throughput for the system using the old Texas Instrument chip. The throughput with the Ford chip should be significantly better at all wavelengths, ( 20-50%).

The ordinate gives the fraction of photons striking the primary mirror that are detected by the CCD. These measurements were made with the old TI chip by observing Gunn & Oke standards through a very wide slit ( 11"). The falloff in the UV is mostly due to transmission losses in the camera, while the loss of response towards 1 micron reflects the decreasing sensitivity of the TI CCD and the difference in blaze of the gratings used. A practical working wavelength limit is 9700Å.

Figure 3 shows the count rate expected at the zenith for a 10th magnitude star observed with several gratings using the old TI5 chip.

Again, these refer to measurements taken with a very wide slit ( 11"). Observed flux through a narrow slit will depend critically on seeing.

The expected count rate can be computed from the above curves and the number of photons per Angstrom per second incident on the 2.1-meter primary mirror. The equation below gives the number of photons/sec/Å, for a star of magnitude and an extinction loss of magnitudes for the 2.1-meter telescope. in Å.

A star of = 15.0 will illuminate the primary with 37 photons/second/Å at 4500 Å at the zenith (extinction at 4500 Å 0.2). For more discussion see the end of the paper by Massey, Strobel, Barnes, and Anderson Ap.J. 328 ,315,1988.

Finally, differential atmospheric refraction is non-negligible when observing in the UV with a narrow slit. For example, when working at an airmass of 2.0, there can be a displacement of 2 arcseconds between images at 4000 Å and 8500 Å. The spectrograph slit can be rotated to the parallactic angle to reduce refractive effects, but there is non-negligible overhead in spectrograph rotation (done manually).