The Advanced Camera for Surveys
  instrument > calibration > results > by item > detector > sbc >



Bias, Overscan
Gain, Noise


Bias, Overscan
Gain, Noise




Shading, accuracy


Encircled energy

Internal Count Rates






Stray Light

Light leak
Filter dust

ACS Home

Last updated
04 January 2002 17:33:40

Maintained by

Printable version

SBC Dark Frames

AUTHORS : Gerhardt R. Meurer



Measure the global dark rate of the Solar Blind Channel and determine what quantities govern it (e.g. MAMA tube temperature). Monitor temporal evolution of the dark rate. Present an accumulated long exposure dark image and discuss its features. Locate and measure hot pixels.


Table 1 lists the available dark frame images that have been obtained with the SBC as of June 1999. The data were obtained for a variety of purposes. These include (1) as part of functional and aliveness tests, (2) as background comparison images for other calibartion goals, and (3) explicitly for determining the dark rate and dark structure.

I. A "Super Dark" Image


During the first thermal vacuum campaign (Feb-Mar 1999), ten one hour long dark images were obtained (see Table 1). The data were taken during the flat fielding phase of the campaign during which the thermal environment of the instrument was constant. All data were obtained after the SBC had been reached an equilibrium temperature of 31°C. An IDL program was used to combine the images and to locate hot pixels.


Figure 1 : SBC "Super Dark" image

Superdark thumbnail
FITS GIF GIF thumbnail: 2x rebin GIF thumbnail: 4x rebin

Dark structure. Figure 1 shows the "Super Dark" image displayed with square root scaling and 2x rebinning. A gzipped fits version of the full resolution image can be obtained by clicking on the fits button below the image. The image is normalized by the exposure time to give counts/pixel/hour. The following features are noticeable in the Super Dark:

Overall shape :
Averaged over the whole frame the dark rate is 0.34 counts/pixel/hour (hot pixels excluded). In the central region the dark rate is ~ 0.9 counts/pixel/hour. The dark rate is considerabley lower in the corners and edges, particularly on the left side (low column number), where the dark rate is as low as 0.05 - 0.1 counts/pixel/hour.
Broken anode:
Rows 599 to 604 are affected by the broken anode of the SBC MAMA. Rows 600 to 603 are effectively dead, while row 604 shows an enhancement in flux. The broken anode has a similar appearance in the SBC flats.
Repeller wire brightening :
A bright vertical line at column 501 (~2 pixels wide) is produced by the repeller wire. This ~20% enhancement is caused by the repeller wire forcing more electrons down the MCP in the region were the wire is closest to the detector. In direct images such as SBC flats the repeller wire also produces a shadow that is displaced from the bright column.
Hot pixels :
Hot pixels are scattered throughout the image. In the above display they are most noticeable at the edges of the frame, where the background is darker. In bad pixels are defined as those having count rates >3 counts/pixel/hour in at least four out of the ten dark images used to make the frame. A list of hot pixels and their count rates can be found here.

II. Global dark rate trends


MAMA Tube Temperature and MCP current. The global dark rates from Table 1 are plotted against MAMA tube Temperature TSBC and MCP current IMCP in Fig. 2.

Figure 2 : Dark rate vs Temperature and Current

rate vs. temp dark
rate vs. current

Strong dependencies are seen in both cases. The lines in Fig. 2 correspond to a factor of ten change in the dark rate for a change in TSBC of 16.6°C (solid line) or 6.4°C (dashed line). In terms of IMCP, a factor of ten change in the dark rate corresponds to a change of 7.5 mA (solid line) or 3.38 mA (dashed line). The plots are very similar in appearance because IMCP and TSBC are strongly correlated. However, the relationship between dark rate and IMCP appears to be somewhat cleaner than the relationship with TSBC. In both plots, there are strong deviations from the (eye-fitted) lines. These amount to deviations on the of ~0.8 dex in some cases. The cause of these deviations is being investigated and is not yet known.

Temporal variations The strong dependence of the dark rate on TSBC and IMCP can easily mask any second order effects. Therefore, Fig. 3 plots dark rate over only a very limited range in time (measured realtive to the first exposure taken on 6 Mar 1999, UT 5:11). This interval of time covers most of the flat field phase of the first thermal vacuum campaign. During this interval, there was no cycling of the SBC power, and TSBC = 30.8 ± 0.2 °C and IMCP = 64.47 ± 0.15 mA were fairly steady.

At first blush, as in Fig. 3a, there is an apparent 10% decline in the global count rate over 100 hours. However, much of this may be due to the IMCP dependency. When this is removed, using the steeper of the two slopes shown in Fig. 2b, then the temporal correlation is much less obvious - maybe a 5% effect with about 5% intrinsic scatter.

Figure 3 : Dark rate versus time

raw global count rate Global count rate corrected for IMCP dependency
a. Raw global count rate. b. Global count rate corrected for IMCP dependency.