Previous Article Next Article Table of Contents
CCD News (1Mar93) (from CTIO, NOAO Newsletter No. 33, 1 March 1993) As predicted in Newsletter No. 32, Reticon #3 CCD has replaced Reticon #2. This CCD is installed in a special dewar for use with the Air Schmidt cameras. Quantum efficiency performance surpasses expectations: we find 52% at 3200 A, 85% at 4000 A, 90% at 4700 A, then slowly falling to 83% at 6000 A, 56% at 8000 A, 33% at 9000 A and 5% at 10000 A. The QE improvement over Reticon #2 is typically a factor of two from 3500-5000 A and a factor of 1.3 in the red. At wavelengths longer than 6000 A the QE of Reticon #3 is very similar to our Tektronix CCDs, and there is now no QE advantage to be gained by specifying the Folded Schmidt camera + Tek 1024 option with the R-C Spectrograph (however the Tek + Folded Schmidt has smaller pixels, rather better image quality and less fringing). Read noise is 4 e- rms at a gain of 1 e-/adu. This CCD also has good cosmetics and charge transfer. The Tektronix 2048 CCD sees much use for direct imaging and on the 4-m Echelle long camera. Until this CCD is installed on an ARCON controller (early 1993), we can only utilize one amplifier of the four available, thus readout times are long. Recent work by Ricardo Schmidt has been very successful in reducing readout times without impairing read noise and charge transfer. Examples are given below. Some of the improvement results from the installation of fast 16-bit Analog-to-Digital convertors (Analogic 4342) in the VEB controllers. Gain Read e-/ RON Old Read Time ADU (e-) Time (min:sec) (min:sec) 5 1:53 8.2 10.0 3:25 10 2:35 3.9 5.6 4:07 15 3:17 2.7 4.2 4:47 20 3:59 2.0 3.5 5:30 30 5:22 1.2 2.7 6:52 As it has been 18 months (September 1991) since a general summary of the CCDs in use at CTIO has been given, I tabulate the information below. Readers should also consult recent Newsletters for details of specific applications. CCDs in use now: 1) TI #3. This CCD sees use for direct imaging in circumstances where high UV QE is important, and occasionally with the Echelle long cameras where the small pixel size gives maximum resolution. The CCD surface is wrinkled, and TI #3 should only be used in slow beams. Cosmetics are poor by modern standards. TI #3 is UV-flooded to give high QE in the blue and UV; the figures achieved are somewhat variable for reasons which are not totally understood. 2) Tek #4. This Tek 512 CCD belongs to Rutgers University, and therefore top priority is for use with the Fabry Perot Interferometer on the 4-m and 1.5-m telescopes. It sees some use for direct imaging, but for most applications its use has been superseded by larger CCDs. It is cosmetically excellent, the only defect being a single pixel with low full well (about 200000 e-). 3) Tek 1024. We have two of these CCDs. Tek 1024 #2 has four working amplifiers and at present is being converted to operate with an ARCON controller, while #1 has a single operative amplifier. Both have perfect cosmetics, even at the very lowest light levels. QE falls precipitously below 4000 . They see most use for direct imaging, but on occasions are used for spectroscopy with the Folded Schmidt camera. 4) Tek 2048. We have one, with another on order. This CCD is scheduled to be converted to ARCON operation in early 1993. UV response is much better than the other Teks; however there are several column defects, and there are some additional columns which have poor charge transfer at very low light levels. The 2048 sees most use for direct imaging and with the Echelle long cameras. 5) EEV (ex-GEC). Now our oldest CCDs. They are front-illuminated and have been coated with laser dyes to provide some blue and UV response. One is dedicated to the 1.5-m spectrograph (#10). A second (#11) is installed in an Air Schmidt dewar. It sees niche use for programs where fringing cannot be tolerated, and since it is flat, the images are rather better than those obtained with the Reticon. 6) Thomson. We have two of these front-illuminated CCDs, each coated with Metachrome in order to provide some UV and blue response. They are cosmetically excellent with four working amplifiers and very low noise, but their use has declined since we received the big Teks which have much better QE. After-images can be a problem. One Thomson is dedicated for use at the Schmidt telescope, the other is a lab test device. Both are operated by ARCON controllers. 7) Reticon. See above. Future plans: These are difficult to predict in any detail since we do not know the quantity or quality of the Loral CCDs we might expect from the two wafer runs in which we participated, nor when our second Tek 2048 will be delivered. We hope to be able to improve our spectroscopic capabilities at the 4-m, 1.5-m and 1.0-m telescopes, replacing present CCDs with thinned, coated Loral 3072 x 1024 or 1200 x 800 CCDs. The Loral CCDs all have 15 um pixels which are a better match to the spectrographs than are the larger pixels of the Reticon, EEV and Tektronix CCDs. Other CCDs slated for early retirement are the TI and the Tek 512, while eventually we hope to provide a larger CCD at the Schmidt. At the same time, we will be converting all our CCDs to ARCON controllers. This is a daunting task, but the performance and reliability gains should transform directly into higher-quality data and more efficient use of telescope time. Brief Summary of CCD Characteristics TI Tek512 Tek1024 Tek2048 Reticon Thomson EEV Pixels 800x800 512x512 1024x1024 2048x2048 1200x400 1024x1024 384x576 Pixel size(um) 15 27 24 24 27 19 22 QE 3000 A 50 4 2 20 40 19 20 QE 4000 A 50 35 30 60 85 18 17 QE 5000 A 60 65 62 75 89 28 22 QE 6000 A 60 80 74 76 83 40 35 QE 7000 A 50 80 76 73 75 38 45 QE 8000 A 37 67 63 63 56 32 30 QE 9000 A 20 37 34 35 33 13 14 QE 10000 A 10 10 8 8 5 3 Read noise(e-) 6-8 8-10 3-6 3-6 4-6 3-5 6-8 Preflash (e-) 25 0 0 0 0 0 0 Full well (Ke-) 40 600 > 300 > 300 > 250 250 150 Note: The lower read noise figures correspond to a gain of about 1 e-/adu, the higher figures are for a gain of 3-4 e-/adu. Alistair Walker
Previous Article Next Article Table of Contents