HOBBY EBERLY TELESCOPE    

SPECTRAL CALIBRATION SYSTEM     

LRS Grism 1 (300 l/mm, ~4.5 /pix, 4000-10500 )

REGION PS GIF  4000-6000  6000-8000  8000-10500  Bright lines

Residuals

Nonlinear component of the standard 5th order fit Residuals after fitting a cubic polynomial Residuals after fitting a 5th order polynomial

LRS Grism 2 (600 l/mm, ~2.0 /pix, 4300-7300 )

REGION PS GIF 4300-5500  5500-7300  Bright lines

Residuals

Nonlinear component of the standard 5th order fit Residuals after fitting a cubic polynomial Residuals after fitting a 5th order polynomial

The previous line lists and plots were generated in an iterative manner. Night sky lines were first used to identify the bright lines in the HgCdZn and Ne spectra independently. The HgCdZn and Ne plots were then added and the previously identified lines were used to make further identifications. A comparison with Keck LRIS line lists was then performed to improve confidence in the identifications. We now believe all plotted lines to be correct identifications, but some small uncertainty remains in the far red end (>9000 A), so use these identifications at your own risk. Feedback is always welcome.

We have determined the standard fit to be a 5th order Chebyshev polynomial (6 coefficients). Though basically a standard polynomial it is rumored that the Chebyshev form, in the presence of noise, fits the underlying function more precisely. Though most spectrographs do not show the red turn over we observe, we believe that this is a consequence of the unusually large range of angles of light exiting the grating.

Finally, all data represented below was acquired with the LRS set to observe the highest resolution possible with a given grism. We used the 0.5 arc-second pinholes illuminated by the comparison lamps. This should give the user and idea of which lines are likely to be blended. Only additive error effects were taken into account in the data reduction. No flat fielding, illumination correction, or photometric calibration was performed.