On-orbit HRC and WFC Internal Tungsten Lamp Count Rates

AUTHORS : A.R. Martel, G. Hartig, and M. Sirianni

INTRODUCTION :

After the installation of ACS in HST on March 7, 2002, it was found from the Servicing Mission Functional Test (SMFT) and Servicing Mission Observatory Verification (SMOV) programs that the count rates of the internal tungsten lamps were significantly higher than the rates measured in the ground campaigns. These higher flux levels, as well as the changes in spectral distribution, can probably be attributed to "hotter" lamps because of reduced convective cooling of the filament by the gas inside the lamp envelope, due to the zero-g environment, in addition to the reduced cooling of the envelope itself due to the vacuum environment (E. Johnson, private communication, see also Johnson 1997, SER CAL-001; 1998, SER CAL-002). Model predictions indicate that a 2% change in filament temperature can result in an increase of 12-31% in radiant output with a strong wavelength dependence.

In this report, we present the revised on-orbit count rates and compare them to the pre-launch measurements. These new rates should be used for all future ACS programs requiring the internal tungsten lamps (CTE measurements, flat field stability and repeatability, etc...). We note that the nominal gains and amplifiers for the HRC are amp C and gain=2 e-/DN and for the WFC, amps ABCD and gain=1 e-/DN. The default lamps for internal calibration are tungsten 4 (TUNG4) for HRC and tungsten 2 (TUNG2) for WFC and the deuterium lamp for the HRC F220W and F250W filters.

DATA :

1. Orbit

The first on-orbit dataset was taken on 7 Mar 2002 for the SMFT campaign. The images have been archived in the ACS database at Johns Hopkins University as entries 35336 to 35350. The observation sequence is essentially identical to the ground sequence taken at the Kennedy Space Center in Dec 2001 (see KSC SMFT) except that the read-out amplifiers for the HRC bias frames were interchanged between amplifiers B and C. Entry 35337 (full WFC frame with TUNG1 illumination) suffers from data dropout so entry 35350 should be used instead.

The second on-orbit dataset with useful internal flat fields was observed as part of the SMOV 9005 program ("CCD Functional", M. Clampin, M. Sirianni). All the HRC flats were illuminated with TUNG3 and the WFC frames with TUNG1. For both detectors, pairs of full-frame flat fields were acquired through F555W with amp C for the HRC and ABCD for WFC at gains of 1, 2, 4, and 8 e-/DN. These were followed by pairs in F435W and F814W (gain=2, amp C for HRC and gain=1, amps ABCD for WFC). Subarrays in F555W were also acquired with WFC with the four different amplifiers.

The results from the SMFT and SMOV programs permitted a rapid update of the flat field exposure times in the interim calibration program 9562 ("ACS Internal Flat Field Stability", R. Bohlin, J. Mack) which was executed on Apr 22, 2002 with TUNG2/WFC and TUNG4/HRC through the F435W, F625W, and F814W filters in the standard gain and amp configurations (HRC : gain=2, amp C; WFC : gain=1, amps ABCD). The results from this program are considered below.

2. Ground

Comparison images acquired during ground calibration campaigns at BATC, GSFC, and KSC can be downloaded from the ACS pre-flight database at Johns Hopkins University. We give preference to the data presented in HRC and WFC Internal Tungsten and Deuterium Lamp Count Rates at nominal gains, in particular the ambient measurements since these are significantly more comprehensive than the thermal vacuum measurements (TB/TV 3). We make no distinction between MEB sides 1 and 2 (all the on-orbit images were acquired on side 1) and we do not make any corrections for detector temperature (the SMFT dataset, for example, was acquired at -67 C for both CCDs and most of the ground data were obtained at -81 C for HRC and -77 C for WFC).

RESULTS :

The statistics were performed on the imaging regions of the HRC and WFC frames, including the HRC Fastie finger and the WFC central gap, using the 'iterstat' task in IRAF (with binwidth=0.001 in 'imstat'). For the SMFT and SMOV flats, the median bias levels in the leading physical overscans over the same regions defined in 'calacs' were first measured and subtracted from the imaging area. For the CAL 9562 flats, the medians were measured directly on the imaging area of the _sfl frames, the sum of the two flat fields of a given filter, processed through 'calacs' (bias and dark subtraction, etc...). For consistency, the ground images were measured in the same manner as the on-orbit data. If more than one frame is available for one filter, the average is listed. The counts were then divided by the integration times from the RPS2 file (since 'calacs' version 3.5h (10-Apr-2002) and earlier populates the file headers with truncated exposure times) and converted to electron rates with the appropriate gains tabulated in HRC#1 : Gain, Linearity, Saturation, Noise and WFC#4 : Gain, Linearity, Saturation, Noise.

The results are tabulated in Tables 1 (HRC) and 2 (WFC). In columns 2 and 7, the ID number for the ground and SMFT programs or the abbreviated FITS file name are listed for SMOV 9005 (j8cr*) and CAL 9562 (j8ef*). The median count rates are tabulated in e-/sec/pixel and converted to DN/sec/pixel (in parentheses) for the standard configurations of gain=2, amp C read-out for the HRC and gain=1, amps ABCD read-out for the WFC. Digital saturation is reached at 65535 DN/pixel for both detectors. The percentage change between the ground and on-orbit count rates are tabulated in columns 6 and 11. The change in output of the HRC lamps, TUNG3 and TUNG4, is very similar at all wavelengths but considerably different between the WFC lamps, TUNG1 and TUNG2. But the percentage changes of TUNG2, TUNG3, and TUNG4 are nearly all identical (25% to 46%). Unfortunately, because of the unexpected higher count rates in orbit, several frames of SMOV 9005 are saturated, in particular all the gain=1 images in F435W, F555W, and F814W. These images are ignored.

In Figs 1 and 2, linear fits to the percentage change between orbit and ground rates are shown for the HRC/TUNG3-WFC/TUNG1 and HRC/TUNG4-WFC/TUNG2 lamps, respectively. From these, the on-orbit rates can be estimated by using the results tabulated in Tables 1 and 2 of HRC and WFC Internal Tungsten and Deuterium Lamp Count Rates.

• Fig. 1 : Linear fits to the orbit-to-ground percentage change in the count rates of HRC/TUNG3 and WFC/TUNG1.
• Fig. 2 : Linear fits to the orbit-to-ground percentage change in the count rates of HRC/TUNG4 and WFC/TUNG2.

Table 1 : HRC Tungsten 3 and 4 Count Rates

Note : The count rates in DN/sec are listed for the standard Amp C, gain=2 read-out. A gain of 2.22 e-/DN is used for the conversion.

Table 2 : WFC Tungsten 1 and 2 Count Rates

Note : The count rates in DN/sec are listed for the standard Amp ABCD, gain=1 read-out. A gain of 1.0 e-/DN is used for the conversion.

DISCUSSION :

The count rates measured in orbit are significantly higher than those measured on the ground in ambient environment. Can these differences be explained solely by the vacuum environment ? Comparison of count rates measured at GSFC in ambient (SSDIF) and vacuum (TV3) environments for HRC/TUNG3/F555W flat fields show an increase in flux of only ~5% and in HRC/TUNG4/F625W, roughly 7%, too low to account for the values listed in Table 1. We conclude then that the zero-g environment is a bigger contributing factor than the vacuum environment to the higher count rates and steep spectral distribution observed on orbit.

We can also verify if the the lamp illumination pattern is different on orbit than on the ground and if any such change results in an increase in observed flux. ACS was launched with the fold mechanism at the HRC position and with the caldoor retracted i.e. away from the light path. Before the SMFT, only the caldoor was initialized when it was moved into place. Thus, any slight change in the position of the fold mechanism, possibly from forces exerted on the instrument during launch and/or from gravity release, could affect the SMFT HRC images. The fold mechanism was re-initialized a few days later, prior to execution of SMOV 9005 (in Vehicle Disturbance Test 1 on 12 Mar 2002) so these images are preferable for comparison with ground data.

In Figs 3 to 5, we show the ratio of HRC SMOV-over-ambient flat fields in the F435W, F555W, and F814W filters. All three show a dark vertical band at pixel column 700 and a smooth rise of ~5% from the amp A-C edge to the amp B-D edge. The apparent location of the Fastie finger shadow has also shifted horizontally by 1.2 pix towards the amp B-D edge, as a result of gravity release in the light path elements (M1, M2, IM1, IM2, etc...). Although ACS was launched with the corrector mechanisms in nominal positions for optimal focus on the ground, it was found in orbit (SMOV 9013, 9014) that large tip-tilt corrections are necessary to counteract the coma introduced by gravity release in the HRC and WFC channels. The unfortunate secondary effect of the tip-tilt corrections is a slight shift in the location of the Fastie finger shadow.

Similarly, in Figs 6 and 7, we show the ratio of WFC SMOV/SMFT-over-ambient flat fields in the F555W and F814W filters. The F555W image suffers from a gain offset in the amp D quadrant, a new dust particle in the amp C quadrant, and streaking from physical saturation in the amps C and D quadrants. But except for these features, the illumination pattern appears relatively unchanged in this filter, A similar result is found for the F814W image (Fig 7), except perhaps for a slight increase in flux along the amps C-D edge and in the amp D corner.

• Fig. 3 : Ratio of SMOV/ambient HRC F435W flat fields for TUNG3 (amp A is in the top-left corner and Amp D in the bottom-right corner). The display scale is linear from 0.95 to 1.05. The illumination pattern shows a smooth rise of ~5% towards the amp B-D edge.
• Fig. 4 : Ratio of SMOV/ambient HRC F555W flat fields for TUNG3.
• Fig. 5 : Ratio of SMOV/ambient HRC F814W flat fields for TUNG3.

• Fig. 6 : Ratio of SMOV/ambient WFC F555W flat fields for TUNG1 (amp A is in the top-left corner and Amp D in the bottom-right corner). The display scale is linear from 0.97 to 1.03. The streaking is due to near physical saturation in the SMOV gain=2 flat field. The "half-moon" signature of a new dust particle on the F555W filter is apparent in the amp C quadrant.
• Fig. 7 : Ratio of SMFT/ambient WFC F814W flat fields for TUNG1. Some weak stray illumination is observed along the amps C-D edge and in the amp D corner.

CONCLUSION :

The on-orbit count rates for the internal tungsten lamps were measured for the F435W, F555W, F625W, and F814W filters for the HRC and WFC using SMFT, SMOV and interim calibration data. Comparison with the ground-based measurements in ambient environment shows a significant increase in fluxes and a strong wavelength dependence, approximately 25-46% (F814W to F435W) for the HRC and 16-43% (F814W to F435W) for the WFC. Count rates in ground vacuum campaigns predict an increase of only a few percent so an additional factor, possibly the zero-g environment, must contribute to the large increase observed on orbit.

The illumination pattern on the HRC detector (TUNG3) is significantly different from that observed on the ground. It shows a smooth rise of about 5% towards the amps B-D edge and a small translation of the shadow of the Fastie finger is apparent. By comparison, the illumination pattern for the WFC (TUNG1) shows only a slight enhancement in flux along the amps C-D edge.

The tables will need to be updated as more internal calibration flat fields with different detector and filter combinations become available.