[Top] [Prev] [Next] [Bottom]
Image Quality and Focus
The Pupil Alignment Mechanism
The Pupil Alignment Mechanism, or PAM, consists of an adjustable mirror in the NICMOS optical train that can be moved to make small corrections to the NICMOS focus and serves to properly position the pupil image of the telescope primary mirror onto the corrective optic. The motion of the PAM is limited to +/- 10mm in focus travel from its zero position. The NICMOS cameras were designed to share a common focus with the PAM close to its zero position. In the current state of the dewar, cameras 1 and 2 (NIC1 and NIC2) can each be focussed within the range of the PAM. Camera 3 (NIC3), however, cannot be focussed by motions of the PAM alone. Throughout this document we will use the PAM mirror position as the measure of focus position of the three NICMOS cameras.
The focus positions of all three NICMOS cameras have changed since launch with motion in the dewar. The positions are measured by observations of stars over a range of focus settings on a frequent basis. Phase retrieval is used, especially for NIC3, to determine the optimal focus positions. The focus history since shortly after launch is shown in Figure 4.5.
Figure 4.5: NICMOS Focus History as of June 9, 1997.
Cameras 1 and 2
The changes in dewar geometry leading to the lack of focus in NIC3 have also affected NIC1 and NIC2. By measuring the PSFs of stars at a series of PAM positions it has been determined that the optimal focus for NIC1 occurs for a PAM position of ~+2 mm and the optimal focus for NIC2 at ~0.5 mm. This difference is significant enough that NIC1 and NIC2 are no longer considered to be parfocal. A separate PAM position at the optimal focus is defined and maintained for each camera. The main effects of this decision are to similarly degrade the image quality of either NIC1 or NIC2 when used in parallel and to add overhead associated with changes from NIC1 to NIC2.
The encircled energy profiles for NIC1 and NIC2 at representative wavelengths are shown in Figure 4.6 through Figure 4.9.
Figure 4.6: Encircled Energy for Camera 1, F110W.
Figure 4.7: Encircled Energy for Camera 1, F160W.
Figure 4.8: Encircled Energy for Camera 2, F110W.
Figure 4.9: Encircled Energy for Camera 2, F160W.
Figure 4.10: Encircled Energy for Camera 2, F222M.
The NIC1 and NIC2 foci are sufficiently close that reasonably good quality images will be obtained by each when the PAM is positioned at the optimal focus of the other camera. The slightly defocussed images in NIC1 or NIC2 in parallel will be sufficiently good that parallel observations are strongly encouraged, as discussed below and in Chapter 3.
The PAM mirror must be moved when switching between NIC1 and NIC2 resulting in a 240 second instrument overhead. This may result in slightly less time available for science exposures, particularly if frequent shifts between NIC1 and NIC2 are made during an orbit. Efforts are now underway to reduce this additional overhead, but observers can minimize the impact of any such overheads by reducing the number of switches between NIC1 and NIC2 to a minimum. A compromise focus that shares the wavefront error equally between NIC1 and NIC2 is also supported.
Vignetting in NIC1 and NIC2
The lateral shifts of the NICMOS dewar have resulted in vignetting in cameras 1 and 2 in addition to NIC3. In the case of NIC1 and NIC2, the source of the vignetting is most likely the Field Divider Assembly (FDA) mask. Relatively small losses in throughput are observed near the edges of both NIC1 and NIC2 as shown in Figure 4.11 and Figure 4.12.
Figure 4.11: The curve below shows one column of a ratio between a recent NIC1 flat field at 1.1 µm and a flat field taken during thermal vacuum testing before launch. The overall normalization of the y-axis ratio scale is arbitrary. The approximately 6% decrease seen in rows near the bottom of the detector (labelled Line on this figure) shows the region of the NIC1 detector where vignetting has affected the throughput. The step-like edge near y=128 may be related to a quadrant boundary.
Figure 4.12: Similar to 4.11, this curve shows an approximately 10% decrease in throughput in rows near the bottom of NIC2 from vignetting by the FDA mask.
Camera 3
NIC3 has suffered the largest shift in focus, now estimated to be at a location equivalent to a ~-14 mm motion of the PAM from its nominal position after reaching a maximum of -17.7 mm in late March. Because the PAM is limited to motions of +/- 10 mm, NIC3 cannot now be focussed with the PAM alone.
Figure 4.13 shows a series of measured star images over a range of PAM focus positions. The images were taken in steps of 1 mm of PAM motion from +8 mm on the left to -8 mm on the right. The current (June 30, 1997) best focus positions are +1.68 +/- 0.49 mm (NIC1), -0.32 +/- 0.43 mm (NIC2), and -14.00 +/- 0.06 mm (NIC3). Consequently, at either the NIC1 or NIC2 focus positions, the NIC3 image quality is very poor.
Using the TinyTIM software package it is possible to calculate model PSFs for any PAM position, including positions not reachable by the actual PAM mechanism. The model PSFs produced in this way agree well with the observed PSFs as shown in Figure 4.14, produced by R. Fosbury, R. Hook, and W. Freudling of the ST-ECF.
Figure 4.13: A series of NIC3 images of two stars at different PAM positions ranging from +8 mm to -8 mm (left to right). The upper star is located near the top of NIC 3 (y=225). The lower star is located near the bottom of NIC3 (y=15) in the vignetted region of the detector. Each subimage is centered in a 20 by 20 pixel box (i.e. 4 by 4 arcsec). These images were obtained during the coarse alignment test before the peak expansion of the dewar. The NIC3 focus was estimated to be at a PAM mirror offset of -14.8 mm at that time, similar to the focus at the time of this writing. An image of a point source in NIC3 when NIC2 is prime would be similar to the star images in the sixth or seventh subimage from the left. When NIC1 is prime a NIC3 point source image will be similar to the image in the eighth or ninth subimage from the left. The full range of the PAM mirror extends to -10 mm or two steps farther to the right than covered by this focus sweep.
Figure 4.14: Model PSFs computed for NIC3 and several positions of the PAM mirror are shown in the ST-ECF figure reproduced below. Since the figure was produced, the camera 3 focus was moved to ~-14 mm.
Zoom to enlarged view
Using the TinyTim models (provided by Richard Hook of the ST-ECF), Figure 4.15 through Figure 4.17 show the encircled energy for Camera 3 in three passbands the infocus, -2mm, and -4 mm defocus positions. The -4 defocus curve corresponds approximately to what can presently be achieved with Camera 3 at the limit of the PAM focus range.
Figure 4.15: Encircled Energy at Three Focii for Camera 3, F110W.
Figure 4.16: Encircled Energy at Three Focii for Camera 3, F160W.
Figure 4.17: Encircled Energy at Three Focii for Camera 3, F222M.
To illustrate the impact on parallel observations, the predicted encircled energy for NIC3 observations of a point source at the three different PAM mirror positions are shown in Figure 4.18. These curves have been derived from model PSFs generated by TinyTim. Observed PSFs are in good agreement with the model PSFs generated by TinyTim. Although Figure 4.18 overstates the present situation (-17.15 vs. the current -14mm focus for NIC3), it shows the limitations of NIC3 parallels when at the NIC1 or NIC2 focus positions.
Figure 4.18: Three curves showing the encircled energy in NIC3 for PAM mirror positions corresponding to -9.5 mm (dashed curve) and the NIC2 (dotted curve) and NIC1 best focus positions (solid curve). The TinyTim program was used to generate the PSFs for an assumed best NIC3 focus corresponding to a PAM mirror offset of -17.5 mm.
NIC3 Vignetting
In addition to the loss of focus, there is evidence of significant vignetting of the NIC3 field of view. As the PAM mirror is moved to shift the focus forwards or backwards, it simultaneously translates the field of view laterally, moving one or more obstructions into the field of view. The observed vignetting in NIC3 is most likely a combination of a warm bulkhead edge, far from focus, which affects the lower ~1/4 of the detector (in y) and a portion of a mask on the NICMOS field divider assembly (FDA) that results in a decrease in throughput over a smaller portion of the detector. The approximate extent of the warm component of the vignetting is shown in Figure 4.19 and Figure 4.20. This also delineates the portion of the detector where degraded PSFs are observed. Figure 4.21 shows the decrease in throughput over a smaller area caused by the FDA mask.
The vignetting is a function of the PAM mirror position. The figures shown are for a PAM position of -9.5 mm, the closest to focus position available for NIC3. At this focus the vignetting is substantial. However, when NIC3 is observed with the PAM positioned for either NIC1 or NIC2, we do not detect any evidence of vignetting. This effect can be seen in Figure 4.13 where the star images in the lower row show the effects of vignetting at PAM positions of -6, -7 and -8 mm. Repositioning the Field Offset Mirror (FOM) would allow observations in NIC3 with reduced vignetting. Tests sufficient to enable general use of this option are planned for August and September 1997. Should this prove successful, then the new FOM position would be adopted as the default for all Camera 3 observations.
Figure 4.19: A NIC3 flat field at 2.4 µm measured on-orbit divided by a flat field measured during thermal vacum testing shows enhanced thermal background in the lower quarter of the NIC3 detector. This emission is due to vignetting by a warm bulkhead edge that is far from focus. A line plot of a row average is shown in Figure 4.20. Over the same portion of the detector, the PSF is degraded, as shown by the lower set of images in Figure 4.13.
Figure 4.20: The average countrate as a function of row for the NIC3 2.4 µm flat field ratio shown in Figure 4.19 is plotted showing the range of the detector where increased thermal background and degraded PSFs occur. This component of the vignetting is thought to be produced by a warm, out-of-focus bulkhead edge.
Figure 4.21: A second component of the vignetting in NIC3 is shown in this ratio of a recently measured NIC3 flat to a flat measured during thermal vacuum testing. At 1.1 µm no enhanced thermal emission is detected. However, at rows near the bottom of the NIC3 detector a loss of throughput as large as 60% is evident.
Paths to Recovery of NIC3 Capability
It is possible that at some point the ongoing, normal loss of cryogen from NICMOS will result in a relaxation of the forces on the dewar that have led to the loss of focus for NIC3. If the present trend in the motion of the NIC3 focus position continues, NIC3 may return to focus by late 1997 or early 1998. Even if this does not occur, it will still be possible to obtain optimal focus for NIC3, by refocusing the HST itself. Since motions of the HST secondary mirror impact the operation of other science instruments (e.g., the WFPC2 does not have an internal focusing capability), if an HST refocus is required then NIC3 observations will be performed in one or two "campaigns" during Cycle 7-NICMOS.
Observers are advised to seriously consider the necessity for NIC3 observations
and to plan programs with NIC1 or NIC2 whenever possible. It is particularly
important to minimize scheduling and orientation restrictions for NIC3 observations.
Field Dependence
The PSF is at least to some extent a function of position in the OTA field of view. Preliminary data indicate that this effect is extremely small and that only a small degradation will be observed. Movement of the FOM, on the other hand, is expected to have a greater effect on the PSF quality. Since use of FOM motions is an "available" mode (i.e., not supported), an extensive calibration of this is not planned. Initial tests indicate that the PSFs are not strongly effected by FOM motion and some further testing is presently planned.
Transient Bad Pixels
Flat fields taken on orbit show a population of pixels with very low count rates. In some cases, pixels with low count rates in one flat field will be normal in the subsequent one. A working hypothesis is that the bad pixels are caused by debris lying on top of the detectors. Paint flecks from the optical baffles are one possible source of this debris. The largest of these areas of bad pixels occurs in NIC1 and is shown in Figure 4.22 below. Approximately 100 pixels in each of NIC1 and NIC2 are affected by this debris and a similar number are expected to be affected in NIC3. As described below, dithering is recommended for observers who believe that these and other pixel defects could adversely affect their science goals.
Figure 4.22: A portion of a NIC1 flat field image shows the largest of the groups of pixels affected by debris. This bit of "grot" is roughly 5 by 9 pixels and is located in the upper left quadrant of NIC1. This particular group of bad pixels has remained constant for several weeks. Other features are transient on week time scales.
[Top] [Prev] [Next] [Bottom]
stevens@stsci.edu
Copyright © 1997, Association of Universities for Research in Astronomy. All rights
reserved.
Last updated: 07/24/97 15:31:29
