Transcript Conventional Mirror Technology - The University of Arizona College

Lightweight mirror technology
using a thin facesheet with active rigid support
J. H. Burge, J. R. P. Angel, B. Cuerden, H. Martin, S. Miller
University of Arizona
D. Sandler
ThermoTrex Corp.
Advanced lightweight mirror technology being
developed at University of Arizona
Motivated by Next Generation Space Telescope
Builds on UA developments for adaptive optics
NGST primary mirror requires a break from
conventional technologies
HST
NGST
Collecting area
4.5 m2
25-40 m2
Mass
800 kg
600 kg
180 kg/m2
15 kg/m2
Warm
40 K
F/2.4
F/1.2
Mass/area
Operating temperature
Focal ratio
Why?
Bigger for seeing weaker sources, also diffraction limit in IR
Faster focal ratio to limit size into launch vehicle
Passively cooled to 40 degrees to allow far IR operation
Lighter so it can be sent to a more distant orbit, far from earth background
Conventional Mirror Technology
• Use glass because of its stability. Once the mirror is
figured, it will maintain its shape.
• Make the mirror thick enough to have rigidity against
dynamic loads and parasitic forces.
• Make the mirror rigid using mass efficiently -attach facesheet to backsheet with ribs.
• Support the mirror by controlling the applied forces.
• HST Egg crate
• MMT
• Lightweight technologies
Membrane with Active Rigid Support
Ideal shape
Structure deforms,
taking membrane with it
Actuators are driven
to compensate
These mirrors depart from conventional thinking
• The mirror surface itself has little tendency to take the correct
shape on its own.
• Uses rigid position actuators
• Relies on active control with bandwidth defined by time scales
of instability or thermal drift of structure
• Actuator length is driven to accommodate errors in the support
structure (different from AO DM which drives surface to have
figure errors that compensate the atmosphere)
• All system rigidity comes from support structure and
connections to glass
Key advantages of active mirrors
• Achieves weight and figure goals of NGST
• Robust system, can correct unexpected problems
• Optimum use of materials
– Carbon fiber structure for light weight, stiffness
– Glass for stable, high-quality optical surface
• Facilities and techniques now exist to make 8-m NGST
– make the parent and cut into segments
• Actuators are key elements
– Mass produced and tested economically
– System is designed to tolerate failed actuators
Obvious questions
•
•
•
•
•
•
•
•
•
How can such a membrane be manufactured?
Will this really work?
Can it survive launch?
What are glass properties at 35K?
Are actuators available that have nm resolution at 35K?
Will such a complicated system be reliable?
How does one choose the number of actuators?
What is the next step in developing this technology
Can this technique provide the NGST primary mirror in
time for 2007 launch?
Fabrication of glass membrane
The concept is to work the glass while it is rigidly bonded in place
Demonstration of a 53-cm prototype
2 mm thick Zerodur membrane, f/1.4 sphere
Carbon fiber support made by Composite Optics, Inc
36 screw-type Picomotor actuators from New Focus
Total mass of 4.7 kg (21 kg/m2)
Figure 33 nm rms after backing out static gravity effects
Substrate and some funding provided by NASA Marshall
Actuator and glass attachment for 53-cm prototype
Figure of shell while it was blocked down
150 nm
-150 nm
48 nm rms
Optical measurements of 53-cm prototype
After manually adjusting actuators to optimize the figure
Raw measurement
Calculated figure after
subtracting self-weight deflection
150 nm
-150 nm
53 nm rms
33 nm rms
Demonstration of survival of 1-m glass membrane
• 2.2 mm shell, sagged to 4-m radius
• supported on 75 dummy actuators,
roughly 100/m2, giving ~400 Hz
fundamental frequency
• aluminum backing plate
• Survived 3 dB over Atlas IIAS load
in Lockheed Martin’s acoustic test
facility
• Membrane survived shipping
mishap as well as acoustic test
Cryogenic CTE for glasses
1.2
Borosilicate
1.0
0.8
0.6
CTE (ppm/°K)
0.4
0.2
0.0
-0.2
Li Aluminasilicate
-0.4
Zerodur
-0.6
Fused silica
-0.8
ULE
-1.0
-1.2
0
20
40
60
temperature (K)
80
Cryogenic actuators
• Early prototype designed and built by ThermoTrex and U of A (uses
proprietary ThermoTrex mechanism)
Concept demonstrated, now being optimized for production
Achieves 25 nm resolution at 77K
Requires zero hold power
5 mm total travel, F > 100 g
500
total mass of 72 grams
400
position (nm)
•
•
•
•
•
300
200
100
0
0
10
20
steps
30
40
Wavefront control system
• Wavefront sensors are under development (phase retrieval from
images and interferometry with star light appear feasible)
• Make correction at primary, rather than inducing opposite distortion
into a deformable mirror
• Close the loop using an on-board computer
• Adjust figure every few observations, or every few days, depending
on stability.
• These types of systems are mature for ground based systems
(a) initial state
(b) after adaptive correction.
Segmented mirror built by TTC showing interference fringes for
=351 nm.
Prototype for the thin shell adaptive secondary mirror. This optic
has a 2-mm membrane supported on 25 actuators with bandwidth >
100 Hz.
Effect of failed actuators
• Failed actuators will be retracted, leaving area
unsupported
• Coupled with membrane strain in a complicated way
• if 5% of the actuators fail (8 actuators out of 150 on the
NMSD), the cryo performance will degrade by ~3 nm
rms, depending on details of glass
System design and optimization
• Determine statistics of CTE variations within glass
• fixed weight -- optimum actuator density vs membrane
thickness found by differentiation
• Careful analysis of all launch loads
• Adjust actuator density and location from FEM
• Design coupling from actuator to glass
Actuator coupling to glass
University of Arizona 2-m NGST demonstration mirror
to be measured interferometrically at 35K mid-1999
Weight summary for 2-m NMSD
Item
Kg/m2
Glass membrane, 2 mm thick borosilicate, 2.2 gm/cm3
4.4
Actuators and cabling, 50/m2, 50 gm per actuator
2.5
Load spreaders, 50/m2, 14 gm per load spreader
0.7
Attachments to membrane
0.3
Launch restraint hardware, 6.2 gm per actuator
0.3
Carbon fiber support structure
4.0
Total weight per square meter
12.2
2-m NGST Demonstration Mirror goals and design values
Mirror development program from NASA Marshall Space Flight Center
Parameter
Requirement
Specified goal
Predicted value
>1.5 meters
2m
2m
Figure ( = 633 nm)



Mid-spatial errors



2.0 nm rms
1.0 nm rms
1.0 nm rms
15 kg/m2
< 15 kg/m2
12 kg/m2
Lowest structural resonance
Not specified
Not specified
~70 Hz
Lowest resonance of membrane
Not specified
Not specified
360 Hz
Diameter
Finish
Areal density
Next Generation Space Telescope
Error budget for primary mirror
Mirror Surface Error
12 nm rms
Mirror
fabrication
6.6 nm rms
Membrane
fabrication
5.2nm rms
Null corrector
Calibration
4 nm rms
Thermal effects
7.5 nm rms
Membrane
cryo distortion
7.3nm rms
(after correction using actuators)
10° variation
across segment
1.5nm rms
Control system
errors
6.7nm rms
Actuator
resolution
6 nm rms
Wavefront
sensor
3 nm rms
Scale up for flight mirror
• Same basic design, actuator density,
membrane thickness
• We can make the mirror in any
geometry op to 8.5-m f/1
• The difference from the 2-m is the
CF backing structure. We had
baselined a 6-m monolith for NGST,
but are now designing segments for
proposed deployed systems.
• Now looking at real fabrication issues
for this.
6-m monolith
System mass < 400 kg