We installed SITe science-grade CCDs in KPNO's CCD Mosaic Imager in June, replacing the frontside Loral chips that have been used since the commissioning of the instrument. To test out the new detectors, we had an engineering run in mid-July. The new CCDs represent a vast improvement over the old ones. The new chips are thinned, back-side illuminated devices and are coated with SITe's standard AR coating, much like the several T2K chips that are used as imagers and spectrographic detectors around the mountain. They have 15 µm pixels, however.
We have had limited time to reduce and analyze the data we obtained from the engineering run, but we can offer the following information. The table below lists the gain, read noise, and bias values of the chips as we are currently operating them:
Read
Gain Noise Bias
Chip (e-/ADU) (e-) (ADU)
1 3.7 4.6 346
2 3.2 5.0 391
3 3.7 6.1 254
4 3.5 5.2 434
5 2.9 4.2 175
6 3.6 4.8 253
7 3.9 4.8 404
8 3.5 5.4 393
This is with a long (8 µs) dwell time needed to avoid excessive read noise
but that results in a readout time of almost 4 minutes. We regret the
inefficiency of this longer readout time, and we expect to shorten it in the
near future. The chips have excellent charge transfer efficiency and minimal
bad columns (2 or 3 per chip)
and charge traps. The one deficiency is fairly significant variations in
their flat fields. There is an overall "mottling" of ~ 3% and several of the
chips show a series of rings with amplitudes of up to 15% peak-to-valley.
These variations are somewhat wavelength-dependent. Preliminary tests
indicate that, after flat fielding using dome flats, residuals from these
features remain at the ~ 0.5% level. We are working further to understand
what limitation they will place on the accuracy to which frames may be
flattened.
The CCDs are very flat physically. The whole array shows no significant degradation of image quality even out to the corners of the 36' × 36' field. We obtained 0.8" FWHM images in both B and R bands on nights of good seeing. We also obtained astrometry frames to compute geometric distortion maps in a number of bands. From these we measure that the gaps between the sides of chips are typically 0.75 mm (50 pixels) and the gaps between the ends of chips are typically 0.50 mm (37 pixels).
We have previously described the pupil ghost problem with the 4-m corrector; it is particularly severe with narrow-band filters. The Solgel coating on the surface that causes the problem has deteriorated further since we last wrote about this (see NOAO Newsletter No. 52). Consequently, we have carried out laboratory tests in collaboration with DAO and Cleveland Crystals. These tests showed that hardened spin-coated Solgel was much less vulnerable to deterioration. Therefore, we are recoating this surface with the more resilient coating at DAO this summer.
We continue to work on completing our planned complement of Mosaic filters. The U filter is our current highest priority and will be ready for usage in September. We also hope to have Gunn/SDSS g', r', and i' by some time in the fall (z' is currently available).
We are continuing to improve the acquisition and reduction software. The GUI that controls the automatic display of images as they are read out has been improved to allow more control over this process. We are in the process of adding the capability of performing flat fielding "on the fly" for the display that takes place during read out. This will greatly aid the observers' quick assessment of data quality.
Spurred by the planned installation of the science-grade CCDs, many proposers applied for time with the CCD Mosaic Imager. A total of 104 nights were awarded during the 1998B semester, 45 on the 4-m and 59 on the 0.9-m. While operation of the CCD Mosaic Imager is basically unchanged from before, a number of the important characteristics of the instrument are different from the information in the manual. We will be updating the manual this summer to reflect the current state of the instrument. Please watch the Mosaic Web site (http://www.noao.edu/kpno/mosaic/mosaic.html) for further information and for a link to the updated manual.
Todd Boroson, Taft Armandroff, George Jacoby, Rich Reed (for the Mosaic Team)