NICMOS DETECTORS (FOCAL PLANE ARRAYS)

Following the characterization of the candidate Focal Plane Arrays (FPAs), a preliminary selection has been made for the primary flight units with corresponding camera assignments.

GENERAL PERFORMANCE CHARACTERISTICS

The following table indicates some of the performance figures determined for the flight FPAs which have been tentatively selected from the characterization tests. These may not be the final FPAs that fly, and the specific values may be updated following additional pre-flight detector calibration. These values, however, provide a good representation of the performance which may be expected. 

 
                 ROCKWELL         FPA:    79T0016     79T0011   79T0017  
                 REQUIREMENT   CAMERA:       1           2         3

DARK CURRENT       <5                     0.18        -0.04     0.06
(e- per second)                           +/-0.09     +/-0.09   +/-0.10

Noise              <50                    30.16       28.41      33.36
(e-)

# BAD PIXELS     <1967 (3%)               86(0.13%)   74(0.11%)  37(0.056%)          
CONVERSION GAIN                           10.32       10.23       12.83
(e- per ADU)

SATURATION (e-)                           184000      173000      205000
(@ .46V bias)
(98% Linearity)        

50% CUTOFF                                2.55        2.525       2.52 
WAVELENGTH

QUANTUM EFFICIENCIES

 Preliminary measures of the wavelength dependent Quantum Efficiencies for these FPAs are shown in the accompanying figure. The fine details in these QE curves should not be interpreted as detector artifacts, but most arise from uncertainties in the characteristics of some elements in the test set-up. Like the above performance figures these values should not be taken as final, but are provided as an indication of the kind of QEs that should be expected.

SHADING

 The NICMOS FPAs exhibit a noiseless signal gradient orthogonal to the direction of primary clocking (horizontal in flight). The magnitude of the shading signal is dependent upon:

  • Time between a pixel read and the last reset in it's row
  • Pixel Clock Rate (the "readout duration")
  • Array Column Number (similar functional forms, but different scaling)

    Shading is absolutely repeatable for image frames clocked in exactly the same manner. In the calibration pipeline subtraction of dark frames of identical integration time and clock rate will remove the shading signature. Thus, for routine calibration purposes, exposure times used should match those maintained in standard dark reference frames. Interpolation (polynomial or spline fitting) may be fruitful for differing times.

    READOUT NOISE

    The readout noise characteristics of each of the three flight detectors are quite similar. In general, this noise component is approximately 30 electrons (specific peak values see General Performance Characteristics) with distribution function across all pixels with a FWHM of about 8 electrons. While the 4 quadrents of each FPA has it own set of readout electronics, the readout noise characeristics of all quads in an FPA are very similar. The accompanying figure demonstrates this for quads 1 and 4 of the flight array designated for Camera 1.

    LINEARITY & SATURATION

    The NICMOS FPAs exhibit essentially linear response over most of their useful dynamic range. In both the very low, and very high flux regimes (in the later case prior to saturation) a departure from linearity occurs. The non-linear portions of the response curves are highly repeatable and will be characterized on a pixel-by-pixel basis. Linearity corrections, based upon this characterization, will be performed in the data reduction pipeline for the non-saturated regions of detector performance. The accompanying figure illustrates the regimes of non-linearity and saturation for a representative small region the flight array designated for Camera 1. In this figure, the zero-crossing offset is an artifact of the charactizaton testing.

    FLAT FIELD RESPONSE

     Flat Field frames taken with the NICMOS flight FPAs exhibit both pixel-to-pixel variations, and global (large scale) structures which may be removed with the application of appropriate calibration reference files.

    These representative frames, in the J and K spectral bands, were imaged with the primary flight candidate arrays. The flats were taken with an incandescent Quartz-Tungsten-Halogen lamp illuminating a diffuser at the entrance pupil of a dewar optical system. The FPAs were operated at flight temperatures of 58-degrees K using a 12.5 microsecond pixel clock, and biased at 0.95 Volts. Pulsed resets were applied to the array elements, and integration times of 250 milliseconds were used.

    Each flat field image was averaged from 10 co-added frames, and dark current has been removed by subtracting reference dark frames. The resulting flat fields have been normalized to compensate for detector and wavelength dependent QE variations. Note not only the structural difference in the flat fields, but the color dependence as well. The intensity scale provided with the flats is in counts (ADUs), where 1 ADU is approximately 10 electrons.

    Return to NICMOS UofA Home Page home page.