Proceedings of the ESO conference on Progress in Telescope and Instrumentation Technologies, held in Garching, Germany, ed. M-H. Ulrich, p. 57 (1992.)
Abstract
Introduction
The Furnace
Mold Construction
Glass Loading
Casting and Annealing
Results
Conclusions
Acknowledgements
| Figure 1 | Figure 2 | Figure 3 |
| Figure 4 | Figure 5 | Figure 6 |
| Figure 7 | Figure 8 | Figure 9 |
Abstract
We describe the process of casting a 6.5 meter diameter borosilicate
honeycomb mirror. Descriptions and photographs cover the beginning of
mold building in June 1991 through the casting process in April 1992
and finally a look at the finished blank in June 1992. This mirror
will be used in the MMT Conversion project which will upgrade the
existing Multiple Mirror Telescope to a monolithic 6.5 meter aperture.
This casting also represents the culmination of a decade long
development project to produce affordable 8 meter mirrors which
deliver excellent images.
Introduction
As part of the continuing effort to produce high performance and
affordable mirrors for large telescopes, the Steward Observatory
Mirror Laboratory has cast a 6.5 meter diameter borosilicate honeycomb
mirror blank. The casting of this large mirror represents the final
step in developing the technology and techniques necessary to produce
8 meter lightweight mirror blanks. The new 6.5 meter F/1.25 parabola
will be installed in the Multiple Mirror Telescope operated by the
University of Arizona and the Smithsonian Institution.
The honeycomb structure provides these large mirrors with improved mechanical performance and improved thermal performance compared to more conventional solid blanks. The honeycomb structure provides an order of magnitude increase in mechanical stiffness compared to the same mass of glass in a solid meniscus. This stiffness allows the large mirror to better resist wind forces and allows us to build lighter telescope structures. By ventilating the honeycomb with ambient temperature air, the thermal time constant of the mirror can be reduced to less than one hour. This short thermal time constant allows the honeycomb to track the mountaintop air temperature and reduce the effects of mirror seeing.
A series of smaller mirrors for telescopes have been produced as part of the technology development effort. This series included the successful casting of three 3.5 meter diameter borosilicate honeycomb blanks in 1988 and 1989. These have been described by Goble et. al. (1989).
The Furnace
Since 1990, the furnace on the rotating turntable has been expanded to
allow for the casting of mirrors up to 8.4 meters in diameter. The
upgrades to the furnace control system have been described by Hill
et. al. (1990). Figure 1 shows a cross-section of the
furnace with the 6.5 meter honeycomb mold inside. Figure 2 shows the
upper section of the furnace being lowered into place. The entire
inside surface of the shell and the hearth are lined with 8 kilowatt
electric heaters. The temperature of the glass and mold is controlled
with 1 C ° by servoing these 269 heaters to obtain a uniform surface
temperature. The furnace control system contains 630 type N
thermocouples which are used for temperature control and monitoring.
Mold Construction
In June 1991, the assembly of the complex honeycomb mold began inside
the furnace. The borosilicate honeycomb is formed by melting chunks
of glass into a complex ceramic fiber mold. This mold is removed from
the honeycomb after the blank has been annealed. Figure 3 shows the
ceramic fiber hexagonal boxes being bolted to the floor of the mold.
The boxes are vacuum formed out of aluminasilica fiber into the rough
hexagonal shape by the Rex Roto Corporation. These boxes are then
fired to 1180 C ° before being machined to their final shape on a NC
milling machine. The hexagonal core boxes are held down against their
own buoyancy in the molten glass by silicon carbide bolts. These
custom bolts have been injection molded by the Ferro Corporation. The
bolts are installed before the tops of the core boxes are glued in
place with ceramic adhesive. The bolts attach to hexagonal tiles that
make up the bottom of the mold. These tiles are arrayed on the flat
furnace hearth and may be seen in Figure 2. Figure 4 shows the final
assembly of the 6.5 meter mold in January 1992. After the mold is
assembled and inspected, it is heated (empty) to 1180 C ° to drive
off volatiles and cure the ceramic adhesive bonds. This prefire step
also serves to test the furnace control system. Figure 5 shows the
geometry of the honeycomb structure and the 1020 hexagonal cores.
The 6.5 meter mold is surrounded by a segmented wall built of castable SiC cement. These sections of a cylinder are held together against the hydrostatic pressure of the molten glass by 80 Inconel steel bands. These bands wrap 90 degrees around the mold cylinder or tub and then exit the furnace. Outside the furnace, pneumatic cylinders pull on the bands to contain the liquid glass while allowing the mold to expand and contract. As the mirror cools, the pressure in the cylinders is reduced to avoid crushing the fragile honeycomb structure. Each band has a 25 x 50 mm cross-section and exerts 2000 N of tension while the furnace is at 1180 C °. Thermal expansion causes each of these bands to grow by more than 100 mm as the furnace is heated.
Glass Loading
Chunks of E6 low expansion borosilicate glass are produced by the
Ohara Corporation in Japan. The 5 kg chunks with broken surfaces melt
together smoothly to form the continuous honeycomb. The expansion
coefficient of this glass is 2.9*10 -6 per C ° and Ohara has
produced over 25 tons with an expansion coefficient uniformity of
2*10 -8 per C °. The relatively low softening temperature of the borosilicate
glass allows us to form it into the complex geometry of the honeycomb blank
in a single casting operation. Figure 6 shows the chunks of glass being
hand-loaded into the honeycomb mold in March 1992.
Casting and Annealing
On March 29, 1992, we began the heating process for the casting of the
6.5 meter honeycomb. The mold was heated slowly to 500 C ° and held
there for 12 hours to allow a final test of the temperature control
system. At that point, heating continued at a rate of 20 degrees per
hour. As the glass began to soften at 700 C °, the rotation speed
of the furnace was ramped to 7.40 rpm (April 2, 1992). The rotation
causes the glass to form a parabolic surface 3 cm above the tops of
the core boxes. Rotating the furnace during casting saves the cost of
adding and then grinding off an additional 10 tons of glass. Figure 7
shows the rotating furnace. The mirror was allowed to cool after
spending several hours at the peak temperature of 1180 C °. Figure 8
shows the 3 month long annealing and cooling cycle. Some photos of
the melting chunks of glass appear in the July 1992 issue of Sky &
Telescope magazine.
Results
After three months of patient waiting and watching, the mirror finally
reached ambient temperature in late June. Initial inspection has
shown that there are several dozen ~1 cm bubbles on the surface. The
average faceplate thickness is 35 mm, thinner than the target but
still well over the 28 mm required for the finished blank. As of this
writing, no significant defects have been identified. In Mirror Lab
parlance, ``It's a keeper.''
The mirror will be lifted from the furnace in August 1992 after it has been inspected and the surrounding bands and tub walls have been removed. It will be turned to a horizon-pointing orientation where the soft refractory cores will be removed with a high-pressure water spray nozzle. After a final inspection the mirror will be transferred to the polishing section of the Mirror Lab. Diamond generation of the asphere is expected to begin in early 1993. The mirror will then be polished to a precise parabolic shape with the stressed lap. Martin et. al. (1992) describe recent work with the stressed lap on 3.5 meter mirrors.
Conclusions
The first attempt at casting a 6.5 meter diameter borosilicate
honeycomb mirror at the Steward Observatory Mirror Lab has been
successful. This success will pave the way for additional castings
between 6.5 and 8.4 meters. The stiffness and short thermal response time
of these mirror blanks will improve the achievable image quality of large
ground-based telescopes.
Acknowledgements
The credit for the successful casting of the 6.5m mirror truly
belongs to a group of dedicated Steward Observatory and MMTO
employees. We would specifically like to thank the following people
for their contributions to this casting:
Eric Anderson,
Curt Blair,
Karl Buckendahl,
Jan Collins,
Mitch Collum,
Richard Cromwell,
Warren Davison,
Scott DeRigne,
Ken Duffek,
Robert Esterline,
Elena Ewing,
Patty Freimark,
Greg Hagedon,
David Harvey,
Bruce Hille,
Stephen Hinman,
Mark Hunten,
Greg Johnson,
Karen Kenagy,
Alan Koski,
Richard Kraff,
Randy Lutz,
Sandy Mashburn,
John Mathews,
Mike Mazzola,
Barry McClendon,
Robert Meeks,
Vince Moreno,
Robert Nagel,
Blain Olbert,
Bill Omann,
Ted Parvu,
Bruce Phillips,
Jeff Rill,
Skip Schaller,
Erv Smith,
Russ Stenman,
Wally Stoss,
Tom Trebisky,
Doris Tucker,
Dan Watson,
Russ Warner,
Steve Warner
We would also like to thank Lori Stiles of the University of Arizona Office of Public Information for providing the photos in a timely manner. Eric Anderson and Dan Watson prepared the drawings of the furnace and mold.
The casting of this mirror and the associated technology development has been funded by the University of Arizona, the Smithsonian Institution and the National Science Foundation.