Loral 3K X 1K CCD in the Blue Air Schmidt Camera (1Dec94)

Previous Article Next Article Table of Contents

Loral 3K X 1K CCD in the Blue Air Schmidt Camera (1Dec94) (from CTIO, NOAO Newsletter No. 40, 1 December 1994) As mentioned previously (Newsletter No. 39, p.21) we are in the process of installing a Loral 3K X 1K CCD in the Blue Air Schmidt camera, with the goal of bringing it into service towards the end of the present year. The Arcon controller that will be used with this device has been fabricated and was thoroughly tested using an engineering grade CCD. However, when the science grade detector was installed we had a most unwelcome surprise. This chip had been selected as an excellent device based on the initial evaluation carried out at NOAO Tucson. However, when operated for the first time at CTIO, the CCD exhibited some 30 hot columns, and more seriously had read-out noise of 30e^- RMS, compared to the 8-12 e^- RMS typical of similar Loral CCDs. It is speculated that the CCD may have suffered damage in transit between KPNO and CTIO. Subsequent investigation has revealed that the high RON is due to hot pixels (LEDs) in both serial registers. These inject substantial excess dark current during readout, and hence add an extra shot noise contribution to the effective RON. By reducing the serial clock voltages we have been able to eliminate the LEDs from one of the two serial registers and achieve a RON of ~9e^- RMS at the cost of substantially reduced full well capacity. The other serial register has not responded to such treatment, so this CCD will have to be used in single readout mode, doubling readout times. At the time of writing we are still determining whether there exists an operating regime in which acceptable full well capacity, linearity, serial charge transfer, and low dark current are all obtained, while still maintaining RON low enough to make this device scientifically interesting. We also need to find out if the required operating parameters remain sufficiently stable over time. If further tests show that the present Loral CCD can be made to give scientifically useful performance, then the benefits this chip offers relative to the present Reticon CCD will likely outweigh its remaining disadvantages, and we will bring it into service as originally planned. In brief, the Loral has somewhat higher QE than the Reticon at all wavelengths from 3000-9000 Ź- -up by a factor of 1.4 or more from 4500-8000 Ź. Its much larger format and smaller pixel size result in a factor of 1.4 more coverage and a factor 1.8 higher dispersion (fewer Ź per pixel) when used with a given grating, as well as finer spatial sampling (0.5"/pix compared to 0.8"/pix) and a matching slit width (1"-1.5") more closely in accord with typical seeing conditions. Since the Loral CCD is flat it should be possible to achieve more uniform image quality than was allowed by the roller coaster surface of the Reticon. Mechanical problems during our first, and so far only, engineering run meant that we had to operate the spectrograph far from auto-collimation in order to reach focus, which introduced substantial astigmatism. Despite this, we were able to achieve better than 3 pixel FWHM image quality over the entire chip and saw images of 2 pixels FWHM at best. Certainly, there was no evidence that the resolution of the CCD was a limiting factor in the image quality, a problem which has been experienced with the Loral CCDs at KPNO. One additional concern with the Loral is that it fringes at wavelengths longward of 7000 Ź. However, the fringe amplitude was found to be only 5% (compared to 3% with the Reticon) so that conventional flat fielding techniques should reduce residual fringing to 0.5% or below. If the Loral does not prove to be useable, then we will continue to use the Reticon CCD with the blue Air Schmidt. Unfortunately, we can only offer one or the other, since exchanging the CCD installed in the camera is not something that can be done on a routine basis. We must therefore select the one that offers the best mix of characteristics for most users, even though it may be less than optimum for some programs. In the meantime we are trying to obtain a replacement Loral CCD, although the remaining stock of potential good devices obtained during the foundry run is not very large. A final decision on which chip to leave installed in the Blue Air Schmidt will be taken following two more engineering runs at the end of October and November. The Folded Schmidt plus Tektronix 1K CCD will in any case continue to be available. This provides useful sensitivity from 4000- 10000 Ź, low 3e^- RON, and does not fringe. The wavelength coverage and dispersion obtained with a given grating are similar to those with the Reticon while image quality is somewhat better although variable (see Newsletters No. 25 and No. 26 for full details). For programs scheduled to use the R-C Spectrograph from December onwards, we will decide which of the two supported Camera/CCD combinations to assign based on information in the proposal and inform the PIs accordingly. Users needing further information on the characteristics of the different Camera/CCD combinations should contact Steve Heathcote or Bob Schommer. Steve Heathcote, Ricardo Schmidt, Bob Schommer, Alistair Walker

Previous Article Next Article Table of Contents