Building the Hubble Telescope: Earthly Solutions to Galactic Challenges

Recall a project that still inspires you, years later. We’ve got a few of those. And, to paraphrase Neil Armstrong, one of our favorites sure was one giant leap for welders.

Customer and Industry

Working with Perkin-Elmer, Inc. (now Goodrich Optical and Space Systems), Lockheed Missiles and Space Corp (the prime contractor), and the United States government, we undertook a project which was critical to the country's mandate to explore the universe and captured the nations imagination from the very start. The job was part of one of the most ambitious projects ever attempted in the aerospace industry: Build the Hubble Telescope, an astronomical observatory free from the distorting effects of Earth’s atmosphere. It used the largest primary optical mirror ever made at the time, and advanced scientific data gathering hardware and software to let scientists see 14 billion light years away.

The Challenge

Our part of the project was to build the Main Ring, the backbone of the Hubble Space Telescope (HST) structure, out of 6Al-4V titanium alloy. All the carbon fiber structures for the mirror and instrument packages anchor on the Main Ring, which provides the stiffness to keep all the structures in line. It’s also the attachment point for the primary mirror and the solar panels. Not only was the design not complete at the start of construction, which in itself complicated things, but the plans called for blind riveting, a technique with which we had no experience at the time. We were also tasked with building the rotating assembly fixture for adding all the structures onto the Main Ring, along with a series of alignment tools for smaller carbon fiber assemblies.

One major challenge was the metal itself.

Titanium is very springy, which makes forming difficult. It is harder to machine, so it requires more specialized cutting tools. It also distorts as material is removed so machining sequences must be carefully planned to account for the distortion. Titanium is a reactive metal that quickly forms an oxide skin when newly cut metal is exposed to air. The oxide interferes with welding, so great care must be taken to remove the oxide and prevent it from reforming by welding in an inert atmosphere.

The other difficulty with welding titanium is that the weld shrinks laterally and longitudinally as it cools at a much greater magnitude than most other metals. This distortion must be accounted for at every step in fabrication.

From the manufacturing point of view, the extreme precision of every component for assembly and final alignment of all the optical systems was complex and challenging.

Because the challenges pushed the limits of the available technology of the time, the mechanical tolerances were so tight, and with the intense glare of public and government scrutiny, this was not a job for an ordinary fabricator.

The Journey

The project would require all of our capabilities. While it was essentially a machining and assembly project, welding was used to piece together arc segments to make the blanks for the large top and bottom channel sections and inner and outer skins and in fabricating the internal structures.

We approached the job as a cooperative effort with Perkin-Elmer designers, who knew what their finished size envelope was, their target weight and stiffness and the requirements for the interfaces with the various structural attachments. We provided overall design for manufacturability and development of the interfaces and supporting structures as they evolved over the course of the project.

The Challenges

The Main Ring is a riveted assembly of welded and machined 6AL-4V titanium components, consisting of formed and welded inner and outer shells, welded and machined upper and lower channel section rings and welded and machined internal ribs.

The first challenge dealt with fabricating the titanium components. Because of the size of the parts, all of the welding had to be performed in the shop environment using special ENI designed shielding to maintain weld quality. For this alloy, the welds need to be stress relieved. Since the inner and outer skins were at finish thickness, the stress relief needs to be performed in a vacuum or an inert atmosphere.

The second challenge dealt with the riveting process. It takes a lot of work to prepare the parts for riveting and not much time to do the riveting itself. The over 4000 rivets were all blind so they could be installed from the outside of the assembly. There was no access to the inside to allow conventional riveting tools to be used. In addition, assembly joints all had to be tight fitting to ensure that we could maintain final assembly tolerances.

The Solutions

At the time, titanium welding was still in its infancy and was considered an art form. The few firms that worked with titanium welded it in vacuum chambers or dedicated inert gas chambers, which we knew weren’t practical because of the size of the major components. All of the welding was performed in the shop environment using special ENI designed local shielding. This, coupled with rigidly controlled weld processes, allowed us to maintain X-ray weld quality levels. Most of the parts were welded in an unfinished state so they could be stress relieved in air, then finished machined to remove the oxide surfaces. The big inner and outer shells were necessarily welded at finish thickness, so we built a special nitrogen filled furnace to do the stress relief of these parts in house.

The solution for the riveting challenge was even trickier.

Our team was concerned about distorting the assembly when the gaps between parts closed as the riveting progressed. We resolved that problem by building it so tight that there was no room for significant movement.

The parts were progressively assembled for match drilling the holes for over 4000 blind rivets. Precision fixturing was designed and built to accurately locate the match drilled holes. Temporary blind fasteners were used to maintain the position of the mating parts as drilling progressed.

This solution created challenges of its own. Namely, the assembly was so tight that there was not enough room to slide the one piece outer and inner shells off to clean the 4,000 burrs between the mating parts that were generated during drilling. That problem was solved by heating the outside shell to expand it enough to slip past the top and bottom plates. Once the outer shell was off, everything else came apart easily and the team was able to clean the burrs.

After all the burrs were removed, the parts were precision cleaned. The mating surfaces were masked and the remainder of the internal surfaces was painted with a special non-reflective coating.

Final assembly was done in a clean area to prevent trapping debris inside the assembly. The parts were all assembled using temporary blind fasteners and the assembly inspected against final requirements. The temporary fasteners were then replaced with blind rivets, one at a time, to maintain the overall alignment. The assembly was inspected several times in this process to verify the alignment.

But that wasn't the last riveting-related solution we would have to develop.

The high tech, aircraft blind rivets, like pop rivets, had a pull pin that locked in place at assembly. Of the 4,000 rivets, 25 pins didn’t lock and instead fell inside nearly blind spaces in the assembly. We later found that typically, 10% of the pins fail this way. Since we couldn't leave the bits inside, we strapped the ring to a weld positioner and rotated it, tapping with rubber hammers to shake the bits through the small openings in the radial stiffeners towards the few openings in the outer shell.

We tracked them by sound. After a couple of days we had removed all but six, which we eventually located by X-ray in truly blind spaces that were acceptable to our customer. To keep any more from falling out later, we welded every rivet pin in place using the same gas-tungsten arc weld process we used for fabricating the parts.

After the riveting was completed, critical interface surfaces were precision machined and inspected.

The Results

The ring was extremely flat, right at the low end of the target weight range and a magnitude stiffer than the design requirement. The customer was thrilled and we were honored with the Perkin-Elmer Supplier of the Year award, significant publicity and, equally important, immense job satisfaction. After nearly 10 years of work by all concerned and a disappointing start in space, the HST went on to take brilliant images from billions of light years away, dazzling Earthlings and keeping scientists busy for decades.