Previous Article Table of Contents
The SBRC 256 x 256 InSb Evaluation Results (1Dec92) (from ETS, NOAO Newsletter No. 32, 1 December 1992) The IR program has completed testing on a SBRC engineering and science grade device. The science grade device has been installed in the Cryogenic Optical Bench (COB), and observations have been made at the 1.3-m telescope. It can be said without reservations that these new SBRC devices are great. The telescope tests confirm the lab testing and did not uncover any new problems. The higher quantum efficiency of these devices has made the COB a truly powerful astronomical instrument. The dark current in these new devices is less than 1 electron per second with median results as low as 0.15 electrons per second. There is a picture frame effect (around the four sides) in the dark current which we believe is due to LED effects and which we have not yet eliminated. This effect is low and subtracts out of the reduced data. The output drain voltage and the voltage controlling the source follower current source must be lowered during integration to achieve these low values. Another problem is that the temperature sensor provided by SBRC in the package is also a LED source. It is necessary to paint over the diode or switch it off for best results. The quantum efficiency (QE) is very high, and we will have to improve our calibration to say how high for sure. We measured QEs of 100%+ in the J, H, and K bands, which is consistent with what SBRC reports. The University of Rochester reports a QE of ~80% at 3.3 um. We have not yet measured QEs in the thermal IR as our system saturates before we can read out the array. Our new data gathering, processing and reduction system on a SPARC station will not have this problem. There is no residual image retention problem in these devices. We imaged a very strong signal on the focal plane assembly and immediately did a series of 100 second darks and saw NO residual effects. The new passivation being used by SBRC has eliminated the need for a gate to control dark current and has also eliminated image retention effects. We confirmed this result at the telescope with the COB by setting up on a 0 magnitude star and then going to something very faint, and no residual effects of the bright star were seen. The read noise using our reset-read-read technique is 26 electrons rms and is very well behaved. We have not tried our multiple correlated read technique, but the University of Rochester reports it works, and they have seen read noise results in the 10 electron rms range. When we get the system running under the SPARC station, multiple correlated reading will be provided. The one problem area is the low full well capacity. With the small detectors (30 um) the capacity is around 350 K electrons per volt. Since the bias is low for InSb, we are getting around 70 K electrons full well. We will be working on this problem more in the future. With the faster readout system this will prove to be less of a problem, but it does provide a bound on the dynamic range. All in all we can say that SBRC has done a magnificent job of providing a new device which will be as important to astronomers as the older 58 x 62 InSb was in 1987. Al Fowler
Previous Article Table of Contents