MIRROR MIRROR The James Webb Space Telescope--an infrared follow-on to the Hubble Space Telescope--won't launch until 2013, but Ball Aerospace & Technologies is accelerating critical work related to focusing the satellite's huge 18-segment, beryllium primary mirror in orbit.

Actively controlling JWST's 6.6-meter-dia. (21.6-ft.) mirror--roughly three times as big as Hubble's--to tenths-of-nanometer tolerances is a daunting technical challenge. Although massive, multi-segment mirrors are used in a number of terrestrial telescopes, such as the W. M. Keck facility in Hawaii, JWST will be the first in-space application (AW&ST July 7, 2003, p. 42). Additionally, the mirror for Webb, named for the Apollo-era NASA administrator, will require an unprecedented level of in-orbit control to ensure a consistently sharp image of deep-space objects.

The program has ballooned to more than $1 billion over its $3.5-billion budget, prompting NASA to slip the launch two years. Similar woes beset Hubble in its early days, and no one wants a repeat of the colossal embarrassment suffered after it was launched in 1990. That now-storied observatory had a 2.4-meter, single-element mirror built and tested by Perkin-Elmer--but with a test device that had been improperly assembled. Ball developed a corrective optics system that was installed during a shuttle servicing mission in December 1993.

BALL NOW IS ACCELERATING the development of a one-sixth-scale optical testbed designed to reduce technical risks associated with controlling the primary mirror. Those risks must be mitigated prior to a technical review scheduled for January 2007. To that end, a full end-to-end "phasing" of the telescope will be demonstrated via Ball's optical testbed by next September.

The device was built to validate software algorithms and sophisticated hardware on the ground, elevating confidence in the entire optical package prior to launch of the full-scale telescope. Testbed cost is about $6 million--not counting algorithm development--"a relatively small, but incredibly important piece" of Ball's $212-million subcontract, says Mark Bergeland, Ball's JWST program manager.

"One risk area for JWST is demonstrating an ability to 'phase' this large, segmented telescope [after] it's deployed in orbit," he adds. With an onboard camera, telescope operators "will take images of stars and use image-based wavefront sensing and control [WFSC] to move the mirrors into proper alignment.

"When we first deploy the 18 [primary mirror] segments, they won't be in phase. We'll get 18 different spots, when looking at a star, as opposed to a nice, sharp image. We'll go through a series of [WFSC] processes by first moving the mirrors to identify which spot corresponds to which mirror [segment]. The next step is to put . . . those 18 different spots into a well-patterned array, then bring the 18 into one spot. They won't be phased, but they'll be coaligned. We then go through coarse- and fine-phasing [to] align each of the 18 mirrors and get them all cophased to a fraction of a wavelength of light."

Each hexagonal mirror segment nominally measures 198 mm. (7.8 in.) across and 10 mm. thick. The testbed segments are made of glass, but the flight version's are constructed of very lightweight beryllium. Structurally strong and thermally stable, beryllium has a coefficient of expansion close to zero at 40 Kelvin, where the JWST will operate. Each flight segment also is being manufactured to "92% light-weighting," Bergeland says. A blank of beryllium material starts at about 540 lb., but is machined to a mirror segment weighing about 46 lb.

However, controlling movements of either the glass or beryllium segments involve the same mechanics. "Each of the [mirror] segments--on both the flight [model] and the testbed--has seven actuators on the back," he explains. "Six control rigid-body motion in six degrees of freedom, and the seventh actuator controls the segment's radius of curvature," which is required to focus the mirror.

In a fine-phasing process, these actuators can control movement "to a nominal step size on the order of 8-10 nanometers. That's the kind of resolution we need to phase the full telescope," he notes.

A number of tests over the next 6-9 months will ensure WFSC algorithms developed by Ball and subcontractor Adaptive Optics Associates are mature enough to function properly in orbit. All the algorithms will be "exhaustively tested on the one-sixth-scale telescope testbed," says Duncan Shields, Ball Aerospace's wavefront sensing and control manager.

Sensing a light wavefront is accomplished through a complex process known as "focus-diverse phase retrieval." That means, "when the [light] points are stacked, and all the mirror [segments] are fairly close to the right [position], we put a series of defocus lenses in the beam and look at how that perturbs the wavefront," Shields explains. "In so doing, we can extract exactly what final corrections need to be made in tilt, tip . . . and so forth to make the best possible wavefront [through] fine-phasing over a wide field of view."

Still, the system is designed to enable rephasing about every 30 days in orbit, tweaking the wavefront periodically to ensure a sharp image.

Ball engineers' primary long-term concern is "understanding and maintaining the stability of the telescope" in space. Because the JWST will be positioned at the L2 Lagrangian point, about 930,000 mi. from Earth, it will be subjected to cryogenic temperatures throughout its 5-10-year mission.

"The telescope will nominally operate at about 40 [Kelvin]," Bergeland says. "Just pointing to different portions of the sky will change its temperature slightly, and at the level of nanometer resolution, that's a significant concern. [The structure] has to be thermally stable enough to maintain the telescope's phasing. This is a passively cooled telescope, and any heat that gets from the spacecraft to the mirrors potentially can cause deformations" that will distort images. "Because our tolerances are tenths of nanometers, that really gives [our] mechanical engineers headaches," Shields adds.

DESPITE THE THOROUGH testing now planned, Ball's mirror-control system will ensure unanticipated in-orbit problems can be corrected. Every engineer and manager working on JWST recalls Hubble's initial fuzzy images, and measures are being taken to avoid that nightmare.

"We've incorporated design [measures] to give us the flexibility to [correct errors]," Bergeland says. "For instance, the six actuators in the system give us full six-degrees-of-freedom motion. That's more than we believe we'll really need, but it gives us a backup capability to [adjust mirror segments] in six degrees of motion rather than three. By design, we've incorporated the capability to fix issues if we run into [problems] on orbit."