Instrument Science Report ACS 2005-07
Two-Gyro Pointing Stability of HST
measured with ACS
Anton M. Koekemoer, Vera Kozhurina-Platais, Adam Riess, Marco
Sirianni, John Biretta, Cheryl Pavlovsky
June 29, 2005
ABSTRACT
We present the results of the pointing stability tests for HST, as measured with the ACS/
HRC during the Two-Gyro test program conducted in February 2005. We measure the
shifts of 185 exposures of the globular clusters NGC6341 and Omega Centauri, obtained
over a total of 13 orbits, and compare the measured pointings to those that were commanded in the observing program. We find in all cases that the measured shifts and rotations have the same level of accuracy as those that were commanded in three-gyro mode.
Specifically, the pointing offsets during an orbit relative to the first exposure can be characterized with distributions having a dispersion of 2.3 milliarcseconds for shifts and
0.00097 degrees for rotations, thus less than 0.1 HRC pixels, and agree extremely well
with similar values measured for comparable exposures obtained in three-gyro mode. In
addition, we successfully processed these two-gyro test data through the MultiDrizzle software which is used in the HST pipeline to perform automated registration, cosmic ray
rejection and image combination for multiple exposure sequences, and we find excellent
agreement with similar exposures obtained in three-gyro mode. In summary, we find no
significant difference between the quality of HST pointing as measured from these twogyro test data, relative to the nominal behavior of HST in regular three-gyro operations.
Introduction
As part of the evaluation of options for extending the lifetime of HST, a test was conducted in February 2005 to examine the operational implications of using only two
gyroscopes instead of the three gyroscopes nominally required for HST operations. The
Operated by the Association of Universities for Research in Astronomy, Inc., for the National Aeronautics and Space Administration
Instrument Science Report ACS 2005-07
February two-gyro test (F2G) consisted of a range of different observations carried out
with HST while using only two gyroscopes. These tests included an examination of the
pointing stability of HST, since one of the possible concerns with operating under twogyro mode (TGM) is to determine whether the pointing stability is significantly worse
than with three gyroscopes. If so, this could potentially impact the tracking accuracy of
HST during an exposure, as well as the accuracy of dither offset maneuvers from one
exposure to the next, and the accuracy of guidestar re-acquisitions from one orbit to the
next. In addition to degrading the resolution of the final image, a decrease in pointing stability can also affect the degree of accuracy with which sub-pixel sampling is achieved
with dithering. Therefore, an extensive set of tests was carried out to verify the pointing
stability of HST under TGM. This report presents the results from HST program 10443
(PI: J. Biretta) which used the ACS/HRC camera to conduct the pointing stability tests.
Observational Design and Methodology
This study addresses three critical questions concerning the pointing stability of HST
observations obtained in two-gyro mode:
1. Maintaining stability within an orbit, so that successive exposures during
an orbit are either located at the same position (if no dithering is used), or
are dithered accurately according to the commanded offsets, particularly if
sub-pixel shifts are required.
2. Maintaining sufficient re-acquisition accuracy from one orbit to the next to
enable multi-orbit observations to be successfully obtained, particularly if
sequences of exposures or dither patterns are spread across multiple orbits.
3. Ensuring that the pipeline processing system, specifically the MultiDrizzle
image registration, cosmic ray rejection and combination software, are able
to successfully process data obtained in two-gyro mode.
The observing program 10443 contained a wide variety of exposures, using a range of
exposure times and dither offset strategies as well as different types of guidestars, to quantify in detail the behavior of the telescope in two-gyro mode. The proposal used the HRC
camera (1024x1024 pixels), which has small pixels (26 milliarcseconds) and can thus provide a very accurate measure of the HST pointing accuracy. The observations were spread
over a total of 13 orbits from Feb 21 - 23, 2005, obtaining 155 exposures of the globular
cluster NGC6341 and an additional 32 exposures of the globular cluster Omega Cen at a
10o off-nominal roll. Several different guidestars were used, with magnitudes V=11, 13,
and 14. The exposure times were 10 seconds, 100 seconds and 500 seconds, and were
obtained in a variety of configurations including CRSPLIT sequences of 2, 4 and 5 exposures (with no dithering) as well as 2-point and 4-point dither patterns, with some of the 2point patterns containing a 2-exposure CRSPLIT pair at each location. Most of the observations used the F555W filter, with the remainder using the F330W filter. A summary of
2
Instrument Science Report ACS 2005-07
the observations for this program are provided in Table 1. These datasets allow the HST
pointing stability to be examined over a wide range of different observing modes.
A companion program, 10455 (PI: J.Biretta), obtained exposures in three-gyro mode
of the same target (NGC6341), prior to the two-gyro test. This program contained a subset
of the exposures in 10443, using the same observing configuration, and therefore provided
useful exposures that could be directly compared with those obtained in two-gyro mode.
Analysis and Results
Initial Processing and Distortion Correction
All the exposures were first processed through standard ACS calibration, including
gain correction, bias and dark current removal, and flat field correction. The resulting calibrated FLT files were then transformed onto an undistorted output frame using the
MultiDrizzle software (Koekemoer et al. 2002), which makes use of Drizzle (Fruchter &
Hook 2002) to remove the ACS geometric distortion using the most up-to-date distortion
files (IDCTAB) and distortion residual images (DGEOFILES), as specified in the image
headers. This step also accounts for slightly different distortion terms in the different filters (F555W and F330W), and furthermore removes small additional scale changes due to
velocity aberration resulting from changes in the motion of HST along the line of sight to
the target during an orbit. Thus, exposures from different times and with different filters
could be directly compared with one another. The resulting set of drizzled images, one for
each of the 185 exposures that were obtained, were then each examined in detail to verify
that there were no problems with any of the exposures, before continuing with the pointing
measurements.
Catalog Generation
The pointing accuracy was measured for all the exposures, regardless of whether they
had been obtained as part of a CRSPLIT or NUMEXP sequence with no dithering, or
whether a dither pattern had been used. The goal was to measure how well each exposure
aligned with the commanded pointing of the telescope. This was measured by first creating a catalog for each exposure using the IRAF DAOFIND software (Stetson 1987), with
parameters optimized for centroiding on unresolved stellar sources. For NGC6341, the
F555W images typically contained about 2,000 - 2,600 stars that could be matched
between all the exposures (10, 100 and 500 seconds) while the F330W images contained
about 800 - 1,100 stars that could be matched. For Omega Cen, only about 50 - 60 stars
were matched, so this dataset served more as a consistency check. Cosmic rays were generally not a problem in these exposures; most of the exposures were short enough to have
a low number of cosmic rays, and the number of stars was generally large enough that
occasional stars affected by cosmic rays would show up as significant outliers and could
be easily rejected.
3
Instrument Science Report ACS 2005-07
Table 1. Summary of Visits and Exposure Specifications for HST Program 10443.
Orbit
Visit / Exposure Specifications
1
1
1
*
Visit 01: [V=13 guidestar, 18 exposures]
Exp 1
NGC6341
F555W GAIN=2 CRSPLIT=NO
Exp 2
NGC6341
F555W GAIN=2 CRSPLIT=4
Exp 6
NGC6341
F555W GAIN=2 CRSPLIT=2
Exp 8
NGC6341
F555W GAIN=4 CRSPLIT=2
NUMEXP=10
NUMEXP=1
NUMEXP=1
NUMEXP=1
10x 10s
4x 100s
2x 10s
2x 500s
2
2
2
2
2
3
3
3
3
3
Visit 02: [V=13 guidestar, 28 exposures]
Exp 1
NGC6341
F555W GAIN=2 CRSPLIT=4
Exp 2
NGC6341
F555W GAIN=2 CRSPLIT=2
Exp 3
NGC6341
F555W GAIN=2 CRSPLIT=NO
Exp 4
NGC6341
F555W GAIN=2 CRSPLIT=2
Exp 5
NGC6341
F555W GAIN=4 CRSPLIT=NO
Exp 6
NGC6341
F555W GAIN=2 CRSPLIT=4
Exp 7
NGC6341
F555W GAIN=2 CRSPLIT=2
Exp 8
NGC6341
F555W GAIN=2 CRSPLIT=NO
Exp 9
NGC6341
F555W GAIN=2 CRSPLIT=NO
Exp 10 NGC6341
F555W GAIN=4 CRSPLIT=NO
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
4x
2x
2x
4x
2x
4x
2x
2x
4x
2x
4
4
4
5
5
5
5
5
5
5
6
6
6
Visit 03: [V=13 guidestar, 45 exposures]
Exp 1
NGC6341
F555W GAIN=2 CRSPLIT=NO
Exp 3
NGC6341
F555W GAIN=2 CRSPLIT=5
Exp 5
NGC6341
F555W GAIN=4 CRSPLIT=2
Exp 6
NGC6341
F330W GAIN=2 CRSPLIT=4
Exp 8
NGC6341
F330W GAIN=2 CRSPLIT=2
Exp 9
NGC6341
F330W GAIN=2 CRSPLIT=2
Exp 11 NGC6341
F330W GAIN=2 CRSPLIT=NO
Exp 12 NGC6341
F330W GAIN=4 CRSPLIT=2
Exp 14 NGC6341
F330W GAIN=2 CRSPLIT=4
Exp 16 NGC6341
F330W GAIN=2 CRSPLIT=NO
Exp 17 NGC6341
F330W GAIN=2 CRSPLIT=2
Exp 19 NGC6341
F330W GAIN=2 CRSPLIT=NO
Exp 20 NGC6341
F330W GAIN=4 CRSPLIT=NO
NUMEXP=10
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
10x 10s
5x 100s
2x 500s
4x 10s
4x 10s
2x 100s
2x 100s
2x 500s
4x 10s
4x 10s
2x 100s
2x 100s
2x 500s
7
7
7
7
Visit 04:[V=14 guidestar, 18 exposures]
Exp 1
NGC6341
F555W GAIN=2 CRSPLIT=NO
Exp 2
NGC6341
F555W GAIN=2 CRSPLIT=4
Exp 6
NGC6341
F555W GAIN=2 CRSPLIT=2
Exp 8
NGC6341
F555W GAIN=4 CRSPLIT=2
NUMEXP=10
NUMEXP=1
NUMEXP=1
NUMEXP=1
10x 10s
4x 100s
2x 10s
2x 500s
8
8
8
8
8
Visit 05:[V=14 guidestar, 14 exposures]
Exp 1
NGC6341
F555W GAIN=2 CRSPLIT=4
Exp 2
NGC6341
F555W GAIN=2 CRSPLIT=2
Exp 3
NGC6341
F555W GAIN=2 CRSPLIT=NO
Exp 4
NGC6341
F555W GAIN=2 CRSPLIT=2
Exp 5
NGC6341
F555W GAIN=4 CRSPLIT=NO
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
4x
2x
2x
4x
2x
10s
100s
100s
10s
500s
2-PT DTH
2-PT DTH
2-PT DTH
9
9
9
9
10
10
10
10
10
Visit 06:[V=11 guidestar, 32 exposures]
Exp 1
NGC6341
F555W GAIN=2 CRSPLIT=NO
Exp 2
NGC6341
F555W GAIN=2 CRSPLIT=4
Exp 3
NGC6341
F555W GAIN=2 CRSPLIT=2
Exp 4
NGC6341
F555W GAIN=4 CRSPLIT=2
Exp 5
NGC6341
F555W GAIN=2 CRSPLIT=4
Exp 6
NGC6341
F555W GAIN=2 CRSPLIT=2
Exp 7
NGC6341
F555W GAIN=2 CRSPLIT=NO
Exp 8
NGC6341
F555W GAIN=2 CRSPLIT=2
Exp 9
NGC6341
F555W GAIN=4 CRSPLIT=NO
NUMEXP=10
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
NUMEXP=1
10x 10s
4x 100s
2x 10s
2x 500s
4x 10s
2x 100s
2x 100s
4x 10s
2x 500s
2-PT DTH
2-PT DTH
2-PT DTH
11
12
12
12
13
13
13
13
13
Visit 07:[V=13 guidestar, 10o off-nominal roll, 32 exposures]
Exp 1
OMEGA-CEN F555W GAIN=2 CRSPLIT=NO NUMEXP=10 10x 10s
Exp 2
OMEGA-CEN F555W GAIN=2 CRSPLIT=4
NUMEXP=1
4x 100s
Exp 6
OMEGA-CEN F555W GAIN=2 CRSPLIT=2
NUMEXP=1
2x 10s
Exp 8
OMEGA-CEN F555W GAIN=4 CRSPLIT=2
NUMEXP=1
2x 500s
Exp 10 OMEGA-CEN F555W GAIN=2 CRSPLIT=4
NUMEXP=1
4x 10s
Exp 11 OMEGA-CEN F555W GAIN=2 CRSPLIT=2
NUMEXP=1
2x 100s
Exp 12 OMEGA-CEN F555W GAIN=2 CRSPLIT=NO NUMEXP=1
2x 100s
Exp 13 OMEGA-CEN F555W GAIN=2 CRSPLIT=2
NUMEXP=1
4x 10s
Exp 14 OMEGA-CEN F555W GAIN=4 CRSPLIT=NO NUMEXP=1
2x 500s
2-PT DTH
2-PT DTH
2-PT DTH
* Exposure 8 of Visit 01 was not executed, and is not included in this analysis.
4
10s
100s
100s
10s
500s
10s
100s
100s
10s
500s
2-PT DTH
2-PT DTH
2-PT DTH
2-PT DTH
4-PT DTH
2-PT DTH
2-PT DTH
2-PT DTH
4-PT DTH
2-PT DTH
2-PT DTH
Instrument Science Report ACS 2005-07
Source Matching and Shift Measurement Results
Once catalogs had been generated, the process of matching sources between exposures
was carried out iteratively, solving for shifts and rotations simultaneously. As a consistency check, scale changes were also initially solved for. It was found that all the scale
changes due to velocity aberration were successfully removed, and no additional timedependent scale changes were present (that might arise, for example, from breathingrelated effects in the telescope). The only remaining scale change was between short and
long exposures, primarily along the y-axis to the level of a few times 10-5, when the x and
y scales were separately solved for. This is accounted for by a known effect related to CTE
(charge transfer efficiency) which produces slight changes in the centroids of stars. However, this scale change is not time-dependent, and comparisons between exposures of the
same length showed no significant scale changes at all during this program. Possible
changes in skew were also searched for but were not detected at any significant level (and
are also not expected, based on our understanding of the ACS camera optics). Therefore,
since no other scale changes were found, the final analysis was conducted by solving only
for a shift and a rotation, keeping other terms fixed.
The resulting measurements for each exposure represent the difference between the
commanded HST pointing (irrespective of whether or not a dither pattern was being executed), and the actual pointing obtained. All of these can be presented relative to the first
exposure in each orbit, and are shown in Figure 1.
∆ Roll angle (deg)
∆Y (mas)
∆X (mas) Tot. Shift (mas)
Two-Gyro Pointing Stability Within Each Orbit: Program 10443 (ACS/HRC)
Orbit 1 Orbit 2 Orbit 3 Orbit 4 Orbit 5 Orbit 6 Orbit 7 Orbit 8 Orbit 9 Orbit 10 Orbit 11 Orbit 12 Orbit 13
20
15
10
5
0
10
5
0
-5
-10
10
5
0
-5
-10
0.005
0.000
-0.005
0 20 40 0 20 40 0 20 40 0 20 40 0 20 40 0 20 40 0 20 40 0 20 40 0 20 40 0 20 40 0 20 40 0 20 40 0 20 40
minutes minutes minutes minutes minutes minutes minutes minutes minutes minutes minutes minutes minutes
Figure 1: Relative pointing accuracy of HST for each of the 185 exposures obtained in
program 10443 with the ACS/HRC. All exposures are shown, irrespective of whether they
were obtained as part of a dither pattern or whether no dithering was used. The difference
between the commanded and actual pointing for each exposure is shown relative to the
first exposure in an orbit, and therefore provides an indication of the degree of stability of
HST during an orbit. It can be seen that the stability is excellent, generally remaining constant to the level of a few milliarcseconds, comparable to three-gyro performance.
5
Instrument Science Report ACS 2005-07
The results plotted in Figure 1 show that the stability of the telescope is generally
excellent, with individual exposures aligning to within a few milliarcseconds of the commanded offsets during an entire orbit, for all the orbits in this test. Some orbits do show a
slight gradual change with time (for example, orbits 7 and 9) of up to 4 - 5 milliarcseconds. This is generally found to occur even in three-gyro mode, and is interpreted as the
result of thermal changes taking place in the telescope. It should be noted that these 13
orbits were not contiguous but instead were obtained over a 3-day period, during which
other targets at different sun angles were also observed, therefore slight changes in tracking due to thermal effects are to be expected. However, the principal result from Figure 1
is that both the translational and rotational stability of the telescope remains excellent and
shows no drastically different behavior to what is already known from three-gyro mode.
A more quantitative comparison between two-gyro and three-gyro stability can be
obtained by examining the relative offsets in similar data obtained under three-gyro mode.
In addition to program 10455 obtained in three-gyro mode prior to the two-gyro tests, we
also examined the Cycle 12 program 9750 (PI: K. Sahu), which made use of ACS/WFC
over 105 consecutive orbits to obtain a large number of exposures of a single location in
the galactic bulge. The results from this study are presented in detail by Gilliland et al.
(2005), and a subset of these data are examined here. Since the thermal characteristics of
the telescope become very stable after about one day if the pointing remains unchanged,
most of the exposures in this program are not representative of HST observations in general, therefore we chose the first 20 orbits of this program as a more representative
comparison with the data from program 10443.
In Figure 2 we show a histogram summarizing the relative offsets that were presented
in Figure 1 for two-gyro mode, along with the similar measurements obtained for the comparable data from program 9750. In both cases, the shifts and rotations represent the
differences from the commanded pointings for all exposures within an orbit, relative to the
first exposure in the orbit. From the figure it is apparent that the distributions are not significantly different, and this is also borne out by the quantitative comparison between the
two distributions when we characterize each distributions in terms of its r.m.s. dispersion.
The measured dispersion for the total shifts is 2.29 milliarcseconds for two-gyro and 2.19
milliarcseconds for the three-gyro observations, and the rotational dispersions are 0.00097
degrees and 0.00093 degrees for the two-gyro and three-gyro observations respectively.
These are not different in any significant way.
An independent check of the comparison of relative offsets between two-gyro and
three-gyro data was carried out by A. Riess, using the “starmatch” software that has been
in use for the Type 1a Supernova Search program. This is presented in Figure 3 and also
shows excellent agreement between the two-gyro and three-gyro data, as well as demonstrating that the offsets are typically below 0.1 HRC pixels, or < 2 - 3 milliarcseconds.
6
Instrument Science Report ACS 2005-07
Program 10443: Stability within each orbit (NGC6341)
120
Total Shift
Exposures
100
2-Gyro: r.m.s. = 2.29 milliarcsec
3-Gyro: r.m.s. = 2.19 milliarcsec
80
60
40
20
0
0
2
4
6
8
10
Total Shift (milliarcseconds)
12
14
120
Roll Angle Change
2-Gyro: r.m.s. = 0.00097 degrees
3-Gyro: r.m.s. = 0.00093 degrees
Exposures
100
80
60
40
20
0
0.0000
0.0010
0.0020
0.0030
Roll Angle Change (degrees)
0.0040
0.0050
Figure 2: A comparison of the distribution of shifts and rotational offsets between the
two-gyro test data and comparable three-gyro data. These offsets represent the differences
between commanded and measured pointings, as in Figure 1. There is no significant difference between two-gyro and three-gyro data for either shifts or rotations; in fact, it is
highly encouraging to see that the dispersion of the two-gyro shifts is only 2.29 milliarcseconds (less than 0.1 HRC pixels), and the width of the distribution of rotations is similarly small at 0.00097 degrees, which is well below any level that would impact the data.
Next, we compare the repeatability with which guidestars are re-acquired from one
orbit to the next. We examined the first 5 orbits of program 10443 and found that the average uncertainty of re-acquisitions is 16.8 +/- 4.8 milliarcseconds; given the small number
of measurements this is not significantly different from the value of 13.5 +/- 3.8 milliarcseconds derived from the comparable three-gyro orbits from program 9750. Generally, reacquisitions at the start of a new orbit involve a higher level of uncertainty than offsets that
are executed within an orbit, therefore the larger values are not unexpected. It is also worth
noting that the 9750 data were obtained in Cycle 12 and on a different target, therefore this
check simply verifies that there is no dramatic difference in reacquisition accuracy for this
two-gyro test as compared with known three-gyro behavior.
7
Instrument Science Report ACS 2005-07
Pointing Stability within orbit, Visit 4, GS=14
10
-0.0025"
0.0025"
number of exposures
8
6
2G-xshift
2G-yshift
3G-xshift
3G-yshift
4
2
0
-0.20
-0.10
0.00
HRC pixel shifts
0.10
0.20
Figure 3: An independent check by A. Riess on an early visit of program 10443, showing
the comparison of relative offsets between two-gyro and three-gyro data, measured with
the “starmatch” software that is used for the Type 1a Supernova Search program. This also
shows excellent agreement between the two-gyro and three-gyro data, as well as demonstrating that the offsets are typically below 0.1 HRC pixels, or < 2 - 3 milliarcseconds.
Processing Two-Gyro Data through MultiDrizzle
As a final test, we processed the multiple two-gyro exposure sequences through the
MultiDrizzle software, which is used in the HST pipeline to automatically register images
based on their headers, perform cosmic ray cleaning and create a final combined image
using Drizzle. A variety of multiple exposure sequences were obtained in the two-gyro
test, including those specified using CRSPLIT, NUMEXP, and dither patterns. As a first
test, the images were processed through MultiDrizzle using only their header astrometry
as a basis for registration, thus ignoring the offsets of a few milliarcseconds that were
shown in Figure 1. This is the current behavior of the pipeline, in the sense that the images
are registered based on their commanded offsets. The results from this test are shown in
Figure 4, comparing a 10-exposure sequence of NGC6342 obtained in the two-gyro test
with a similar sequence obtained during the three-gyro test. It is readily apparent that there
8
Instrument Science Report ACS 2005-07
is no significant difference in the quality between the two images, and this is confirmed
quantitatively by measuring the PSF of stars in the images in each case. The average PSF
of stars in the images was found to be 2.42 pixels (0.0605”) and 2.43 pixels (0.0608”) for
the three-gyro and two-gyro data respectively, which is excellent agreement.
A similarly excellent agreement was found for all the other image sequences obtained
in the two-gyro test program 10443, which is in fact to be expected since the offsets of the
exposures as shown in Figure 1 were found to be only a few milliarcseconds, or less than
about 0.1 HRC pixels. Therefore, these test images can be combined directly as-is, using
the astrometric information in their headers, without any need to solve for shifts since the
actual offsets achieved agree so well with those that were commanded. If it is ever necessary to incorporate measured shifts, then this can be done by means of the “Tweakshifts”
script which uses the mechanism described in this report to solve for shifts and apply them
to the image headers, prior to image combination.
Figure 4: The results of using MultiDrizzle to process multi-exposure sequences of the
globular cluster NGC6341, which was the principal target for the two-gyro pointing test
program 10443. The image on the left was obtained from the precursor program 10455 in
three-gyro mode, while the image on the right was from the two-gyro program 10443.
These images are of the same general region of the cluster (except for a very small difference in the location of the detector because they were obtained a week apart). In both
cases the image headers were used to align the images, thus assuming that the commanded
shifts represent the true offsets between the images. Their overall appearance of these
images is essentially the same, and this is confirmed quantitatively by the fact that there is
no significant difference in the average measured PSF of stars on these images (Sirianni et
al. 2005, in preparation).
9
Instrument Science Report ACS 2005-07
Summary
We have presented the results of the pointing stability tests for HST, as measured with
the ACS/HRC during the Two-Gyro test program conducted in February 2005. We have
measured the shifts of 185 exposures of the globular clusters NGC6341 and Omega Centauri, obtained over a total of 13 orbits, and have compared the measured pointings to
those that were commanded in the observing program. We find in all cases that the measured offsets of shifts and rotations agree with those that were commanded to the same
level of accuracy as in three-gyro mode. Specifically, the differences between commanded
and actual pointings during an orbit relative to the first exposure can be characterized with
distributions having dispersions of 2.29 milliarcseconds for shifts and 0.00097 degrees for
rotations, thus less than 0.1 HRC pixels, and agree extremely well with similar values
measured for comparable exposures obtained in three-gyro mode. In addition, we successfully processed these two-gyro test data through the MultiDrizzle software which is used
in the HST pipeline to perform automated registration, cosmic ray rejection and image
combination for multiple exposure sequences, and we find excellent agreement with similar exposures obtained in three-gyro mode. In summary, we find no significant difference
between the quality of HST pointing as measured from these two-gyro test data, relative to
the nominal behavior of HST in regular three-gyro operations.
Acknowledgements
We are grateful to Ken Sembach and Ron Gilliland for a careful reading of an earlier draft
and for useful comments. We are also grateful to the ACS group and the scheduling team
who prepared the proposals for this two-gyro test, and to Ron Gilliland for kindly sharing
the data from his ISR in order to facilitate a comparison with three-gyro operations.
References
Fruchter, A. S. & Hook, R. N. 2002, “Drizzle: A Method for the Linear Reconstruction of
Undersampled Images”, PASP 114, 144
Gilliland, R. 2005, “Guiding Errors in 3-Gyro: Experience from WF/PC, WFPC2, STIS,
NICMOS, and ACS”, ISR-TEL-2005-02
Koekemoer, A. M., Fruchter, A. S., Hook, R. N., Hack, W. 2002, “MultiDrizzle: An Integrated Pyraf Script for Registering, Cleaning and Combining Images”, HST
Calibration Workshop, p. 337
Sirianni, M. et al 2005, ACS ISR, in preparation
Stetson, P.B. 1987, “DAOPHOT - A Computer Program for Crowded-Field Stellar Photometry”, PASP, 99, 191
10