Previous Article Next Article Table of Contents
Engineering and Technical Services: (1Mar95) ALADDIN - 1024 X 1024 InSb FPA Development Program Status (from Director's Office, NOAO Newsletter No. 41, 1 March 1995) The collaboration between USNO and NOAO to develop a low background 1024 X 1024 Focal Plane Array (FPA) with SBRC has recently passed a major milestone. We now have a SCA (sensor chip assembly) which has over 900K good pixels. We have met all the design specifications goals with the exception of read noise. A readnoise of less than 25e rms is achieved using Fowler Sampling and 8 sample pairs. We still have a few engineering problems related to hybridization that need to be resolved. As one can see from the picture earlier in the newsletter on ALADDIN, the corners are not bonding as well as we want, and there are voids in the epoxy bonding the detector material to the readout. We expect these problems to be resolved in the first quarter of FY 1995. This is only the second SCA to be made so we are quite happy with its performance. The design specifications are given below for those not familiar with the project. ALADDIN 1024 X 1024 InSb Focal Plane Characteristics SCA Specifications Number of Pixels: 1024 (H) X 1024 (V) 1,048,576 elements Architecture: 4-Independent 512 X 512 Quadrants Pixel Size: 27 um square Effective Fill Factor: 100% Readout Type: PMOS SFD Unit Cell CMOS Shift-registers Control Logic PMOS or NMOS Output Drivers Number of Outputs: 32 (8 per quadrant) Frame Rate: 50 ms Reset Options: Destructive and non-destructive by rows IR Detector: Thinned InSb Full Well: 250 X 10^3 e- at 1.0v bias Wavelength Range: 0.6 -5.5 um with special AR coatings Operating Temperature: 35 deg K Dark Current: < 0.1 e-/sec Noise: < 25 e rms with Fowler Sampling Quantum Efficiency: > 80% 0.9 to 5 um Defective Pixels: < 0.5% No Bad Rows or Columns An unretouched quasi-flat-field image taken with this SCA is available in the ALADDIN anonymous ftp area via instructions elsewhere in the Newsletter. We have measured the read noise, using a narrow band filter (~60KHz) and Fowler 1 sampling, at 46 e- rms. The read noise is improved by using more pairs, and we have gotten less than 20e- rms using 16 sample pairs. Design changes and another lot run of readouts are in the works to further improve the noise. The dark current was less than we could measure due to lab temperature controller errors but is certainly much less than 0.2 e- per second at 400 mv reverse bias. We are able to get a full well capacity equivalent to the applied bias in the ALADDIN design unlike its 256 X 256 predecessor (CRC463), where several hundred millivolts of the applied bias were lost. So the full well capacity (i.e. where it saturates) was around 170K electrons at 400 mv bias. Although we have not as yet measured the QE, it looks to be equivalent to the CRC463 device. Although the design allows for either PMOS or NMOS output drivers, all data taken so far is with the PMOS output drivers. There is no significant evidence of any LED effects, but it is quite sensitive to thermal changes. At this time we have not attempted to operate the device at the 20 Hz frame rate, but a study of the output waveforms indicates that it should meet this specification. Further details and testing information can be obtained by contacting the author. We only received the device in late December 1994 and have not completed all the testing one would like at this time, but from the data we have so far it is safe to say that the Aladdin device is real and will be a very successful array. A.M. Fowler (afowler@noao.edu)
Previous Article Next Article Table of Contents