Astronauts and an army of engineers and technicians on the ground are working hard to ready themselves and the space shuttle Atlantis for the final servicing mission to NASA's Hubble Space Telescope (HST).

If it goes as planned, the risky shuttle flight will leave the 17-year-old telescope at "the apex" of its capability, able to probe the Universe with a full suite of instruments for the first time.

But in an extremely remote worst-case, Atlantis could wind up flying backward into the atmosphere with its payload bay doors open, the better to burn up rapidly over the Pacific, its seven-member crew crammed into the shuttle Endeavour after an unprecedented rescue in orbit.

Both possibilities have been planned and replanned in the 6.5 years since the last servicing mission in 2002. As the telescope approaches its 100,000th orbit on the morning of Aug. 11, excitement over the prospects of another upgrade is growing.

"It is characteristic of every Hubble servicing mission that when we leave Hubble, we leave it a better observatory than it was before," says David Leckrone, the senior project scientist in the Hubble Program Office here.

"When the astronauts leave Hubble for the last time, we believe it will be at the apex of its capabilities," he added. "There will never have been a time in the past that Hubble was more capable scientifically than it will be after this mission."

If the mission goes as planned, the telescope will have a full set of scientific instruments in its light path for the first time since spacewalkers Thomas Akers and Kathryn Thornton pulled out the Hubble's High Speed Photometer in December 1993 to make room for the Corrective Optics Space Telescope Axial Replacement (Costar). Costar was built to fix a spherical aberration, ground into the telescope's main light-gathering mirror, by putting corrective pickoff mirrors in the light path to the instruments. But one instrument slot had to be sacrificed to make room for it.

Now the instruments themselves have corrected optics, and Costar is no longer needed. In its place, the servicing crew will install the Cosmic Origins Spectrograph (COS), an extremely sensitive ultraviolet spectrograph that will enable astronomers to begin mapping the structure and composition of the "cosmic web" of galaxies and intergalactic gas shaped in part by mysterious dark matter.

Also on the agenda will be delivery of the Wide Field Camera 3 (WFC3), which will allow the Hubble to probe deeper into the Universe than ever before in wavelengths ranging from ultraviolet through visible into the near-infrared. The crew will return the old Wide Field Planetary Camera 2 (WFPC2) to Earth to make room for WFC3.

In other work during the mission's five back-to-back extravehicular activities (EVAs), the four spacewalkers on Atlantis will attempt to repair failed electronics in the Space Telescope Imaging Spectrograph (STIS) and the Advanced Camera for Surveys (ACS). The Near Infrared Camera and Multi-Object Spectrometer (Nicmos) continues to function after a repair during the 2002 servicing mission.

If the new repairs work, "we will have a full toolbox," Leckrone says. "This will be the first time that Hubble has been fully instrumented, every instrument slot filled with an active scientific instrument, since December 1993."

Right now, Oct. 8 is the target launch date for what is known as Servicing Mission 4 (even though, because one mission was split, it actually will be the fifth shuttle return to the orbiting observatory). That date may even advance by three days, if flight preparations continue to go well.

But the mission is years late, compared with the 2-3-year intervals between previous flights to the telescope. The last orbiter to service the Hubble was Columbia, in March 2002. The tragic loss of that orbiter and its crew less than a year later threw the whole Hubble-servicing effort into disarray.

For a time it looked like Columbia's visit to the Hubble would be the last by humans, as NASA scrambled to develop a robotic way to service the telescope. Without the "safe haven" offered by the International Space Station to the crew of a damaged orbiter, former Administrator Sean O'Keefe ruled that human servicing missions were too risky in the post-Columbia environment.

Michael Griffin, O'Keefe's successor, restored human servicing after a reanalysis of the risk. The decision was as popular as O'Keefe's had been unpopular, both with the public and on Capitol Hill, but the safety questions remained. Development of the STS-125 mission included elaborate planning for a four-man crew to rescue their colleagues on Atlantis with the shuttle Endeavour, which will be waiting on the other pad at Kennedy Space Center when Atlantis lifts off (see p. 50).

The time spent preparing a robotic servicing mission wasn't a complete waste. One of the trickiest tasks the STS-125 spacewalkers will attempt is replacing a low-voltage power supply board in the STIS instrument that contains a failed power converter. The board lies behind a cover secured with 111 screws that could become that many pieces of dangerous orbital debris if they are not caught in a "fastener capture plate" that grew out of an attempt to develop a robotic way to do the job.

"Some of that technology that was envisioned for a fully robotic repair actually came to fruition in the form of the fastener capture plate," says Preston Burch, the Hubble program manager at Goddard.

Another piece of hardware on the STS-125 mission that evolved from the planning for robotic servicing is a low-impact docking system (LIDS) plate that eventually will be used to attach a propulsion module that will deorbit the telescope after its useful life has ended, now expected to be at least five years off (and probably many more).

When Atlantis robotic arm operator Megan McArthur grapples the telescope and settles it on the flight support system at the back of Atlantis's payload bay that will serve as the spacewalkers' workbench, the LIDS plate will automatically attach to the telescope via its soft-capture mechanism, which also includes the targets that an unpiloted propulsion module will need to find its docking position.

The robotic-mission-planning "provides a good technical baseline for future work for developing a propulsion module for deorbiting HST," Burch says.

At the present state of technology, it is clear human repairmen will be much better able to service the telescope than robots, but they'll have their hands full. EVA planners here have laid out a tightly packed series of five spacewalks, with elaborate contingency plans for completing as much work as possible if problems arise.

Two of the four EVA mission specialists - John Grunsfeld and Mike Massimino - have performed servicing missions to the telescope before (as has Scott Altman, the STS-125 commander). Grunsfeld, an astronomer by training, has been there twice, and will get to use the telescope for his own research during the post-mission checkout period (see p. 50).

Grunsfeld and Massimino will be joined outside by first-time space travelers Michael Good and Andrew Feustel, switching off in teams of two for five days straight starting on Flight Day 4. As in the past, one of the spacewalkers will ride the shuttle robotic arm to haul instruments and other hardware to and from the payload bay, while his partner will clamber along handholds on the telescope.

Given post-Columbia concerns over damage to the orbiter's delicate thermal protection system (TPS), the first order of business after reaching orbit will be the painstaking inspection of the ceramic tiles and reinforced carbon-carbon panels that protect the aluminum structure from the blazing heat of reentry. The crew will use the camera on the shuttle's robotic arm and the laser imagers on the orbiter boom sensor system (OBSS) to go over the TPS inch by inch.

After that, the OBSS will be stowed so the arm can grapple the telescope and set it up on the soft capture mechanism for the five spacewalks.

Here, by flight day, is the nominal EVA schedule for the mission:

* Flight Day 4 (EVA 1). The first job of the first spacewalk of the mission will be to remove the WFPC2 and install WFC3 in its place. Next, the EVA team will install one of two battery modules, each containing three of the 125-lb. nickel-hydrogen batteries that power the telescope when its solar arrays are in eclipse. After 18 years of operation, battery performance is starting to degrade, and the replacements should last as long as the telescope. At the end of the EVA, the spacewalkers will spend about 45 min. on final setup of the soft capture mechanism at the bottom of the Hubble.

* Flight Day 5 (EVA 2). The highest-priority job of the servicing mission is replacing all six of the gyros that keep the Hubble locked onto its distant targets during the long exposures needed to collect as much light as possible from the distant reaches of the Universe. Only three are working at the moment, and since 2005 only two have been used. The EVA team will spend about 3 hr., 20 min. replacing all three of the two-gyro boxes called rate sensor units. The rest of the EVA will be devoted to replacing the other battery module.

* Flight Day 6 (EVA 3). In the first half of this spacewalk, the EVA crew will remove the Costar device, stow it in the payload bay, and install the COS in its place, sliding it in on the rails that hold the large instrument units inside the telescope behind its exterior doors. After that they will start the two-step process of repairing the ACS. In January 2007, all three channels of the instrument went down, and after a month of work controllers here were able to restore only the Solar Blind Channel, which blocks visible light to focus on ultraviolet wavelengths. To restore the rest of the instrument's functionality, the astronauts will replace a charge-coupled device electronics box and power it up with a replacement low-voltage power supply.

* Flight Day 7 (EVA 4). First up on this spacewalk will be the STIS repair, which would be a simple circuit-board changeýýout were it not for the 111 screws in the way. The EVA team will install the fastener capture plate over the access panel, and then use a power tool to unscrew all of the fasteners. As the name implies, the screws will remain in the capture plate when the spacewalkers remove it. Once inside the instrument, the crew will remove a failed power supply card and insert a replacement. The access panel will be replaced with a simpler cover that can be latched in place with two levers designed to be manipulated with spacesuit gloves. The rest of the spacewalk will be devoted to installing New Outer Blanket Layers (NOBLs) over two pieces of original-equipment thermal insulation that have degraded in the harsh space environment over time. The NOBLs, coated stainless steel foil panels, are attached over the damaged insulation with expanding plugs that fit into door-vent holes.

* Flight Day 8 (EVA 5). The crew will start this EVA by replacing one of three fine guidance sensors (FGSs) located at 90-deg. intervals around the telescope's circumference with one pulled from the telescope in 1999 and refurbished. The 478-lb. optical units provide extremely accurate pointing information for the telescope, but two of them are degrading. Only two are essential for the pointing task, so replacing one of them will ensure the telescope will be able to continue operating with an accuracy of only 0.007 arc-sec. of jitter. The devices also are used for astrometry. Once the FGS is replaced, the spacewalkers will spend the remainder of the final EVA finishing the ACS repair and closing out the telescope and payload bay before returning to the airlock.

Given the packed schedule, Hubble planners have set up an elaborate schedule of fallbacks and priorities if any part of the nominal timeline is lost to technical problems with the servicing or the need to reinspect the orbiter TPS. After gyro replacement, servicing priorities are, in order: installing WFC3 and COS, replacing the batteries, replacing the fine guidance sensor, repairing STIS and the ACS and installing the NOBL panels over the degrading insulation. If fuel levels permit, Atlantis also will reboost the telescope.

Just as the Hubble program and the astronaut corps have advanced what humans can do to maintain and upgrade the telescope in orbit, so have manufacturers on the ground improved their techniques for building the unique instruments that collect and analyze the photons Hubble pulls in from afar. Ball Aerospace and Technologies has built seven instruments for the Hubble over the years, including the Costar corrective-optics unit, and has chased the state of the art in optics through the various upgrades.

"We've come a long way since Costar days, and really probably the new challenges now are trying to get the state-of-the-art detectors into our instruments at the last possible moment," says Thomas Delker, principal optical engineer at Ball. "You start down the path of developing your sensors, and the first ones that come out are never as good as the last ones that come out. There's so much risk in building these sensors up and getting them to work and getting through all their qualifications . . . that you're often so far behind the curve of technology by the time something needs to go into the instrument that your sensors are getting a little stale."

By drawing on its experience over the past two decades, Ball has produced two new instruments that will advance what astronomers call the "discovery efficiency" - field of view multiplied by optical throughput - over older Ball instruments on the telescope. The WFC3, for example, will boost the discovery efficiency in near-ultraviolet and blue light by a factor of 35 over ACS, and by a factor of 15-20 over Nicmos in near-infrared.

With those new capabilities - designed to complement Hubble's older instruments - astronomers hope to continue their investigation of topics as diverse as galactic evolution and the formation of stars and planets; the sources of carbon, iron and other elements necessary for life; and the nature of the mysterious dark energy that Hubble found is causing the Universe to expand more rapidly.

"A large proportion - ballpark, 40-50% - of the most important science Hubble has done in the past has been unexpected," Leckrone says. "Questions have arisen we didn't even know how to ask when Hubble was first launched. The Universe has many more surprises in store for us, and I think with this incredibly powerful toolset of new instruments and restored instruments, we're going to stumble on the unexpected as we always traditionally have in the past."

NASA's experience operating and upgrading Hubble over the years also may have a payoff for future human exploration beyond low Earth orbit, when astronauts will have to service their spacecraft and live in them at the same time while far removed from the support available on Earth. Some of the tools future explorers are likely to use, like the "pistol grip tool" astronauts on the International Space Station use to drive bolts to the proper torque, were developed for Hubble servicing, and are likely to find their way to the Moon and beyond eventually.

And while the Hubble missions themselves have been very carefully choreographed - a luxury that probably won't be available on the way to Mars - some underlying principles that go into their planning, such as the need for redundant ways to do a given task, may serve the planners of those future Mars missions as well. The Hubble program hashes out lessons learned "in great gory detail" after every mission, but Burch, the Hubble program manager, worries that they won't be captured for future use.

"I don't think NASA does as well as it could do with capturing lessons learned from previous programs," he says. "It seems like every program that comes along there's a new set of people, a new set of managers, and they all think they know how they want it to get done and they go off and do it. There's been a lot of effort over the years to generate databases, and to pull together documents."

Although the docking mechanism the STS-125 crew will leave behind on the Hubble will work with NASA's planned Orion crew exploration vehicle as well as with a robotic propulsion module to deorbit the telescope safely, there is little chance those lessons learned will be used for an Orion mission to service the telescope. Ed Weiler, NASA's associate administrator for science and the original program scientist on Hubble, advocated retrieving the telescope for display in the Smithsonian's Air and Space museum before the Columbia accident. But now, he believes, it's time to prepare to deorbit the telescope by the mid-2020s, when it will begin to pose a reentry hazard.

"By the time the Orion's flying routine, you're talking eight more years, seven more years, the manufacturer of gyros, reaction wheels, fine guidance sensors, those will all be shut down," Weiler says. "So there won't be any replacement parts to fly [and] the Orion is designed to fly to the Moon. It's not designed to take up tons of cargo to Hubble. . . . I think it's time to move on."