A Scientific Assessment of a New Technology Orbital Telescope (1995)

The task group identified several areas of concern regarding the astronomical utility of the ATD/NTOT, some arising from the fundamental differences between the needs of space astronomy and the requirements of the experiments for which the telescope was designed, and others that are more generic and are based on experience with other missions. The concerns based on special astronomical needs include requirements for:

• Stabilizing images using faint guide stars;

• Minimizing background due to stray and scattered light in the optics and in the infrared by passive cooling;

• Understanding the impact of refiguring the mirror on observing efficiency; and

• Performing long integrations in a high-radiation environment;

The more generic concerns include the issues of:

• Analyzing the ATD/NTOT as a complete system;

• Ensuring the reliability of the software; and

• Controlling cost and schedule.

In subsequent sections the task group discusses at length aspects of these important issues that require further study during the final definition of the ATD/NTOT mission. The discussions are arranged roughly in a priority order reflecting the task group’s interests and expertise.

In the course of its briefings from BMDO, Lockheed, Itek, and Charles Stark Draper Laboratories, the task group identified areas where there are clearly open or, as yet, undefined issues that could significantly affect the ATD/NTOT’s performance of astronomical observations or valid demonstration of technology for future scientific applications. At present the ATD/NTOT is a collection of components that exist in one form or another in various parts of the relevant industries. However, little systems analysis has been done to validate the concept that the various components can be combined into a telescope that is useful for astronomical observations. The task group has no reason to think that they cannot be so combined, but a more complete systems analysis will be needed before launch.

Astronomical operations differ markedly from the proposed DOD experiments in the need for stellar guidance input. It is worth noting that although the ATD/NTOT’s collecting area is several times larger than the HST ’s, it has only about 1% of the available guidance field of view. Consequently, the ATD/NTOT must be able to produce a fine-error signal using very faint stars. The fine-guidance system as designed requires one guide star for controlling pitch and yaw and a second for stabilizing roll. The guidance analysis in the Lockheed/Itek briefing concluded that at 19th magnitude there is sufficient guide star availability and that the ATD/NTOT will achieve sufficient signal-to-noise to determine the centroid of a star’s image to 0.003 arc sec, assuming a detector temperature of 260 K with a 10-Hz sampling rate. The reliability of the estimates of signal-to-noise for centroiding was not clear to the task group. More importantly, however, guide stars will be significantly less abundant than assumed, especially at higher galactic latitudes.

The source of the average stellar density data presented in the Lockheed/Itek briefing is not known to the task group, but the briefing assumed an average guide-star density of about 25 per square milliradian at a limiting magnitude of 19. The task group’s calculations, based on the Bahcall-Soneira model (Figure 6.1), show that for latitudes greater than 40 degrees, the density falls below this number, even when looking toward the galactic center. The situation is considerably worse when looking away from the galactic center, with densities almost five times lower. The stellar density remains low over the entire hemisphere facing away from the center of our galaxy and rises by less than 50% in going from longitudes of 180 degrees to 90 degrees. Since the original estimate assumed success in finding guide stars only 80% of the time, it follows from Figure 6.1 that the success rate will be closer to 15% over most of the galaxy.

FIGURE 6.1 The number density of stars brighter than 19th magnitude, in the Bahcall-Soneira model of the galaxy, is shown as a function of galactic latitude for three different galactic longitudes.

For many of the surveys considered as prime possible uses of the ATD/NTOT, the sky will be only sparsely sampled and the fields can be chosen so as to include suitable guide stars. For observations requiring specific fields, the lack of available guide stars can be accommodated by requiring only a single guide star for controlling pitch and yaw, while using an external star tracker for controlling roll. Clearly these issues must be explored in greater depth.

The issue of telescope background is another area where astronomical operations are likely to differ significantly from DOD’s technological demonstrations. There is clearly a need in astronomy for a dark (i.e., sky-limited) background. The ultimate utility of a large aperture is the ability to concentrate a great deal of energy in a small region of the focal plane, limiting the background contribution. As the background is elevated, faint objects are detected at lower confidence levels and, thus, require longer integration times. In other words, the background reduces the effective collecting area and limiting magnitude of the telescope.

Concerns about background arise from two design aspects of the ATD/NTOT. The first is the lack of a significant forward baffle. This deficiency increases the solar and lunar avoidance angles and allows more of the sky to contribute to stray light in the telescope. Stray light is a major issue at both optical and infrared wavelengths. The second issue is the large, rectangular cross section, tripodal truss supporting the secondary mirror. These elements provide flat specular surfaces facing the primary obscuration that permit photons to enter the focal plane area by direct reflection or by other paths through the optics.

Because the dynamic range of brightness is so large between the Sun and Moon, on the one hand, and deep-space objects, on the other, scattered light can be a tremendous problem for any studies of faint objects, that is, for the prime astronomical targets of the ATD/NTOT. Like the problem of a low density of guide stars, the problem of scattered light from specific, bright sources can be minimized by a careful choice of the fields of a survey. Avoiding bright objects is not, however, a general solution for all fields or for all types of scattered light. Further analysis of the scattered light is essential, and options should be explored for installing baffles and surface treatments to minimize specular reflections.

The ATD/NTOT derives much of its cost and weight leverage from its ability to fly a light, meniscus mirror with many actuators that can be reconfigured on orbit to compensate for low-spatial-frequency errors of figure. The input data for the optical correction is intended to come from an on-board, wavefront sensor. Although this sensor has been identified as slope measuring (and having a mass of 35 kg), no information was available about its speed, dynamic range, noise, and calibration levels—all factors with a significant impact on astronomy operations.

It is envisioned that the influence functions of each of the primary mirror’s actuators will be determined after launch. In all likelihood for astronomical operations, as with the HST, phase-retrieval analysis of images will provide better input to wavefront correction than the input provided by the on-board sensor. Thus it is reasonably possible that for astronomy operations, both in manipulating the influence functions and in sensing the wavefront, the reconfiguring of the optical system will not be autonomous.

Over the course of the acquisition, tracking, and pointing experiments for which the ATD/NTOT has been designed, it is unlikely that wavefront stability will be an issue. The experiments are of short duration and impose only modest limits on stability. Astronomical observations, however, characteristically require longer on-target times and therefore more demanding requirements on the telescope that arise in many different ways. At the very least, the requirement to reconfigure the primary mirror together with the possible need for interaction with the ground is a time overhead factor on operations. The Molniya orbit envisioned for the astronomy phase of the operation involves a dramatically varying thermal environment whose effect on the telescope point response function has not been quantified.

The undefined character of thermal and other figure drivers means that their impact on the required rate of figure correction is not known. Moreover, between corrections, the telescope point-spread function will degrade

from a just-corrected to a needs-to-be-corrected state. The amplitude of this variation can be controlled at the expense of observing overhead as long as the intercorrection period is long compared to a typical integration time. Should the required correction period become too short, then the nature of the science that can be accomplished will be seriously compromised.

One option for exercising a more deterministic control of image quality is to close a slow feedback control loop around the figure actuators driven by output from the wavefront sensors. This option assumes that autonomous figure correction is adequate for astronomy operations. In effect, the figure control loop becomes an analog of the pointing control loop. Assuming a low signal-processing overhead, the speed of the loop is driven by the rate of photon arrival into the wavefront sensor. Assuming that a 19th-magnitude star is used for figure sensing, Lockheed and Itek have estimated that the figure could be updated at intervals of a few minutes. The viability of this approach will depend on the quality of the correction data supplied by the wavefront sensor, the ability of the system to correct the wavefront without disturbing the field, and the availability of a suitable star for a phase reference. In light of the guide-star problem discussed above, this issue needs careful study. Since this issue is inherent to the ATD/NTOT’s technology, perhaps the optimum approach is to implement the necessary control software and investigate it during the technology-evaluation phase of the mission.

Since the Clementine project was carried out under constrained schedule and cost, the lessons learned from it can be applied to the proposed ATD/NTOT. For a variety of reasons, software development has proven more difficult to plan and complete on schedule than has hardware development. The software to run the Clementine mission had not been completed by the time of launch and was still being written during the early stages of the mission. Lessons learned from Clementine about software development include the following:¹

• Take care to understand the impact of the software development environment on test and schedule;

• Realize that “non-commercial” processors are not supported as well as commercial ones;

• Assure that undefined processor states cannot enable critical circuitry; and

• Make maximum use of software verification tools prior to new uploads to the spacecraft.

A problem inherent in imaging arrays is their susceptibility to radiation events, often called cosmic rays. These events leave behind an enhanced signal that can span one or more pixels. Two factors contribute to the severity of these events. That is, the number of detected events is proportional to:

• The thickness of the imaging chip, with thicker chips trapping many more events and having longer detection paths; and

• The intensity of the radiation field.

Thus, thin, backside illuminated CCDs (such as the Texas Instrument chips in the HST’s old WFPC) have fewer cosmic-ray events detected per unit time than thick, front-side illuminated chips (such as the Loral-Fairchild chips in the HST’s new WFPC2). Similarly, detectors on ground-based telescopes are protected by the Earth’s magnetic field, while the HST sees the enhanced radiation field in low Earth orbit. A spacecraft in a geostationary or Molniya orbit would spend much of its time above the protective Van Allen Belts. Thus, its CCDs would have a substantially higher event rate than corresponding chips in a ground-based telescope or the HST.

Experience with the HST has shown that cosmic-ray events can render less useful a significant fraction of the pixels on a chip. They also necessitate breaking total exposure time up into many 20-minute subintegrations in order to be able to statistically eliminate cosmic-ray detections from the final images. There is a cost associated with such subintegrations, since reading out a CCD chip can take 1 to 2 minutes (though new low-noise output amplifier designs already suggest that much shorter readout times will become the norm)—time during which observations cannot be obtained. Additionally, for studies of faint moving objects, such as a Kuiper Disk survey,

removal of cosmic-ray events can be problematic since the target object is moving relative to the background stars and can thus be mistaken for spurious events by many cosmic-ray detection algorithms.

As mentioned in Chapter 4, exploiting the ATD/NTOT’s eccentric orbit to obtain very long exposures is an important technology demonstration of this mission. The feasibility of this test will depend on additional study of the extent to which cosmic-ray events can be handled, either by reducing them through shielding of the chip (although at least one side cannot ever be shielded), by finding intelligent ways to analyze data contaminated with such events, or by implementing continuing improvements in CCD technology.

The planned ATD/NTOT mission is defined as a build-to-cost mission and not a build-to-specification mission. This constraint has important implications. If such a plan is continued, great care must be taken to ensure that the mission schedule does not slip, since time slips always result in higher costs and, if the costs are capped, a loss of capability. However, it is important that early in the mission, the capabilities be well defined so that all of the necessary pieces can be completed. A plan should also be assembled that dictates which specifications are most important in the event of a schedule slip, thus providing a mission with the most value for the budget.

Some of the astronomical projects discussed in this report can be accomplished with the baseline telescope and instruments. However, some important scientific projects and technology demonstrations will require enhancements, for example, an optical framing camera. Great care must be taken in deciding to add such instruments, since their addition has implications for the mission’s cost and schedule. In today’s cost-constrained NASA, it is not reasonable to perturb the budget greatly in order to add an instrument of the cost of the WFPC2. NASA and/or DOD should investigate alternatives, such as a university instrument of lower than class A or an instrument procured in the mode of Clementine or an instrument derived very directly from a camera currently being developed. It is also important that NASA and the astronomical community understand the short time scale needed for the delivery of such instruments and that the reasonableness of achieving timely production of such an instrument be considered.

If scientific investigations are to be an important part of the ATD/NTOT mission, then scientists will need to be included early so that appropriate decisions can be made regarding details such as filters. Such early inclusion means that the selection process for scientists will have to come early in the mission and be expedited.

Traditionally in NASA programs, the interval is long from formulation of a Research Announcement to the point when the personnel start work. The short development schedule anticipated for the ATD/NTOT does not allow a lengthy selection procedure if the goal of including scientists early is to be met. Similarly, the finite lifetime of the mission requires that personnel be used efficiently throughout the mission. Thus, consensus will have to be reached on how to achieve early involvement of scientists, in order to minimize any impacts on the delivery schedule and maximize the flexibility of the team to react to problems. Since such an approach is consistent with many of the goals being espoused within NASA today, the ATD/NTOT mission would provide an excellent opportunity to evaluate this new paradigm.

1. Space Studies Board, National Research Council, Lessons Learned from the Clementine Mission, National Academy Press, Washington, D.C., in preparation.