Chapter 4

Chapter 4, The Point Design

Section 4.4: Telescope Structure

NOAO Logo    Gemini Logo

The telescope structure, along with the bearings and drives, provides the bench for the optics of the telescope; the platform from which the control system, with its actuators, is able to track an object across the sky; and the means to correct image motion caused by the atmosphere and motion of the structure itself.

The primary requirements for the telescope structure are: (1) a minimum natural frequency higher than that of the bandwidth of the control system for the primary mirror so that corrections to maintain continuity of the segments of the primary mirror will not drive resonances of the structure; (2) sufficiently small quasi-static deflections under gravity and temperature so that the deflections will fall within the dynamic range of the sensors and actuators designed to detect and correct the deflections; and (3) a dynamic response to wind that minimizes the amplitude and frequency of the corrections that must be provided by the systems of adaptive control of the integrated telescope system.

This section presents: (1) the results from a preliminary structural analysis of the point design structure described in Section 4.1, (2) results of preliminary investigations seeking to improve on the performance of the point design, and (3) a description of planned steps for investigating the telescope structure.

4.4.1 POINT DESIGN TELESCOPE STRUCTURE

Figure 1 Point design telescope structure developed by Simpson, Gumpertz, and Heger.

The initial structure developed for the point design telescope is shown in Figure 1. This design was developed by the company of Simpson, Gumpertz, and Heger, Inc. (SGH) and is based on a structural configuration commonly used in radio telescopes. SGH prepared the conceptual structural design and performed finite element analyses to evaluate the expected structural behavior needed in developing other aspects of the telescope design. See Appendix 4.4 A for the complete SGH report.

The mount portion of the telescope structure consists of built-up rectangular tube sections. The preliminary FEA (finite element analysis) used square tubular sections with dimensions of 1.50 m by 1.50 m with a 25 mm wall thickness. The mount is supported by four vertical hydrostatic bearings that ride over a single circular track to provide azimuth motion. Radial restraint is provided at the center of the bottom level framing of the mount.

Figure 2 Finite element model of a raft structure.

The optical support structure (OSS) pivots in elevation on hydrostatic bearings located on the mount. A hexagonal framework consisting of built-up rectangular steel tubes, with dimensions of 1.93 m by 1.37 m by 32 mm wall thickness, connects the remainder of the OSS to the elevation axis trunnions. Behind the elevation axis is the combined counterweight, elevation drive, and control surface. The primary mirror back structure space truss is mounted directly to the front of the hexagonal framework. In the preliminary FEA, the truss members were considered to carry axial load only and were treated as rod members with a cross-sectional area only. These sections are equivalent to nominal 6 in. (6-5/8 in. (168 mm) actual diameter) standard pipe sections with a cross-sectional area of 5.58 sq in. (32.3 sq. cm).

Primary mirror segment rafts, which are space trusses in themselves, are fitted into the openings in the front face of the primary mirror back structure, and are mounted to the back face with a hexapod structure of actuators. Structural members are typically 1 in. (1.315 in. (33 mm) outside diameter by 0.133 in. (3.4 mm) wall thickness) and in. (1.050 in. (27 mm) outside diameter by 0.113 in. (2.9 mm) wall thickness) standard round pipes. A raft with mirror segments above and the hexapod actuators below is shown in Figure 2. Structural members of the mirror back structure space truss, which support the hexapod actuators, are also shown.

The secondary mirror support structure (or tripod) consisting of three trusses is also mounted to the primary mirror back structure (see Figure 1.). The truss members are steel rectangular tubes with dimensions of 6 in. by 12 in. with 3/16 in wall thickness (152 mm by 305 mm by 8 mm). Steel cables with an effective cross-sectional area of 1.00 sq. in (6.45 sq. cm.) between the primary mirror back structure and points along the secondary mirror support trusses provide out-of-plane stiffness to those trusses.

The mass of the mount and the optical support structure for the structural model used by SGH, broken down into subsystems, is as follows:

Mount647,000 kg

Optical support structure694,000 kg
   Rafts (with mirrors, mirror supports, etc.)144,000 kg
   Secondary3,000 kg
   Back structure121,000 kg
   Tripod structure18,000 kg
   Counterweight with structure301,000 kg
   Main OSS frame107,000 kg

Total mass of structural model1,341,000 kg

4.4.2 LOWEST MODE NATURAL FREQUENCIES

With this point design finite element model, SGH calculated the first four natural frequencies of the entire telescope structure. Table 1, which is repeated from the SGH report, lists the frequencies and a general description of the mode shape.

The first mode shapes for zenith pointing and horizon pointing are shown in Figure 3. These mode shapes and the next three higher mode shapes are shown in the SGH report.

Table 1 First four natural frequencies and
mode shapes for the entire telescope structure.
Zenith Pointing
ModeFrequency (Hz)Mode Shape
12.17Nodding (Elevation)
22.24Side sway
32.70Torsional (Azimuth)
43.09Tripod
Horizon Pointing
ModeFrequency (Hz)Mode Shape
12.16Nodding (Elevation)
22.16Torsional (Azimuth)
32.27Side sway
43.09Tripod

Figure 3 First mode shape of the telescope structure for zenith pointing (left)
and horizon pointing (right).

4.4.3 DEFLECTIONS UNDER OPERATING WIND LOADS

SGH ran a set of finite element analyses on the telescope structure using representative spectral wind loading obtained at the 8-m telescope at the Gemini South Observatory. The average outside wind speed was between 11.5 and 13 m/s when the wind measurements were taken. The RMS (root mean square) translations (TX, TY, and TZ) and rotations (RX, RY, and RZ) at the rafts and the secondary mirror (M2) as calculated by SGH are repeated in Table 2.

Table 2 RMS deflections of the point design telescope structure using representative spectral wind loads obtained at the Gemini South Observatory. The headings listed refer to the vent gate, wind screen, telescope direction with respect to the wind, and zenith angle at Gemini South at the time the data were taken.
Vent Gates Closed, Wind Screen Closed
Telescope Pointing into Wind, and Zenith Angle = 30 deg.

TX, RMS
(µm)
TY, RMS
(µm)
TZ, RMS
(µm)
RX, RMS
(µrad)
RY, RMS
(µrad)
RZ, RMS
(µrad)
Minimum among rafts0.325.350.400.3680.0180.114
Average among rafts0.886.002.700.4470.1010.153
Maximum among rafts1.756.866.420.6300.2610.264
M216.8039.350.660.6331.54321.850
Vent Gates Open, Wind Screen Open
Telescope Pointing into Wind, and Zenith Angle = 30 deg.

TX, RMS
(µm)
TY, RMS
(µm)
TZ, RMS
(µm)
RX, RMS
(µrad)
RY, RMS
(µrad)
RZ, RMS
(µrad)
Minimum among rafts0.325.350.400.3680.0180.114
Minimum among rafts1.6613.751.270.9450.0740.297
Average among rafts2.7915.417.031.1580.3080.387
Maximum among rafts4.7117.5416.551.6150.7770.664
M242.6099.912.081.6643.89954.638

4.4.4 PRIMARY MIRROR SEGMENT RAFTS

Table 3 Lowest natural frequencies and
mode shapes for a raft structure.
ModeFrequency (Hz)Mode Shape
1 & 25.8Side sway
312.0Torsional
4 & 521.5Clamshell
621.8Vertical

Figure 4 First mode shape of a typical raft. The raft is the upper truss. Selected
members of the back truss are shown as the lower truss. The hexapod actuators connect the two trusses.

In the analyses performed by SGH, the overall telescope structure was modeled using a simplified representation of the primary mirror segment rafts by using lumped masses located at the position of the center of gravity of each raft. A separate finite element model, which is shown in Figure 2, was prepared and analyzed to obtain more detailed performance at the rafts. The SGH point design has seven mirror segments mounted on a single raft structure. The use of a raft structure has two advantages: (1) several segments can be removed or installed onto the telescope in one operation, and (2) the raft provides a low order method of correction to the segments, leaving the higher order corrections to the individual segments.

The SGH analysis of the raft calculated the natural frequencies and mode shapes of an isolated raft structure as shown in Table 3.

The first mode shape for the raft structure is shown in Figure 4. This mode shape and the next three higher mode shapes are shown in the SGH report.

The displacements and rotations for the mirror segments under gravity loadings were calculated by SGH and are repeated in Table 4.

Table 4 Displacements and rotations of the mirror segments under gravity loads.
1 g Gravity Load in +X Direction
Mirror
Segment
TX
(µm)
TY
(µm)
TZ
(µm)
RX
(µrad)
RY
(µrad)
RZ
(µrad)
S16549-20326644563452-21
S26736-28234136033831-28
S362511455-86275371
S463809-5-5632142
S569189-59-51434444
S66537215-2597-5673450-29
S76732294-3464-7103832-40
1 g Gravity Load in +Y Direction
Mirror
Segment
TX
(µm)
TY
(µm)
TZ
(µm)
RX
(µrad)
RY
(µrad)
RZ
(µrad)
S1-20867771470-4029-55130
S2-2576396-2032-3096-624-72
S3-968763971-4170-79-4
S4-36370-6-3231-70-2
S526393-3026-3132-453
S622267791568-4027472-29
S72716403-1937-310454574
1 g Gravity Load in +Z Direction
Mirror
Segment
TX
(µm)
TY
(µm)
TZ
(µm)
RX
(µrad)
RY
(µrad)
RZ
(µrad)
S17139451-601033
S284-4856185154-1
S32102561-18312
S4-2-541221-100
S5-1-7744910802
S6-6641450-58-970
S7-80-4456083-1473

4.4.5 PRELIMINARY INVESTIGATIONS TO IMPROVE PERFORMANCE

Further FEA studies were performed on the model developed by SGH in two particular areas of interest. The first study investigated alterations to the mount with two goals: (1) the possibility of increasing the first mode frequency so that the effective bandwidth of the control system can be increased, and (2) removing the interior braces to provide an envelope for a fixed gravity instrument. A rendering of the modified mount is shown in Figure 5.

Figure 5 Rendering of the telescope structure with the modified mount structure,
viewed from the back side of the telescope. The mount rides on two azimuth tracks, with the large
outriggers providing additional stiffness at the altitude axis bearings. Framing has been removed
between the two mount columns to provide space for gravity invariant laboratories. The optical
support structure has also been modified from a centered, single counterweight and altitude drive
disk to the two disks shown.

The goal of a second study was to eliminate the cables that brace the secondary support tripod trusses. These cables could (1) introduce unwanted vibration into the telescope structure caused by the aerodynamic effects of wind on the cables, (2) increase scattering of light by diffraction, and (3) increase the fluctuating radiation background during IR observations. This study also illustrated the possible benefit of using advanced materials in the telescope structure.

4.4.5.1 Stiffening of the Mount

The SGH conceptual point design was calculated to have a minimum natural frequency of 2.16 Hz. A significant gain in the lowest natural frequency is possible by using two azimuth journals, along with additional structural outriggers, to provide a wider base and improved structural load paths. For example, the European Southern Observatory VLT (very large telescope) unit telescope design uses two azimuth journals in this manner.

The two-track mount design also offers the opportunity to provide a volume for fixed-gravity instruments in the middle space below the optical path. By placing structural outriggers to the exterior of the altitude axis bearings, the structural bracing required on the interior between mount columns can be removed. The OSS is modified so that the single counterbalance and drive disk located at the centerline of the OSS is replaced with two such disks, each located adjacent to a mount column.

Two optimization runs were performed on the modified mount configuration, using the cross- sectional areas of the structural members as the optimization parameter. The first run was set to optimize the first mode frequency at 2.2 Hz to match the lowest mode frequency of the SGH model. The second model was optimized to 3.3 Hz, which is near the highest frequency that can be achieved with the mount configuration as shown. The results of the two optimization runs for the modified mount configuration are compared with the results obtained by SGH on the original mount configuration in Table 5.

The SGH model is 50% lighter than the modified mount when the modified mount is optimized to approximately the same first mode frequency obtained for the SGH model (see Table 5). The modified mount model can be optimized to a first mode frequency of 3.3 Hz, or 50% higher than for the SGH model. The penalty, however, is a mass of almost three times that of the SGH model.

It is anticipated that reduced masses with first mode frequencies of 3.0 Hz or higher are possible with further studies of mount configurations, plus more detailed optimization procedures. However, this preliminary study indicates that higher first mode frequencies can be reasonably obtained, along with the added benefit of the ability to provide the volume for a gravity invariant laboratory.

Table 5 Comparison of mount masses and lowest mode frequencies for the SGH model, and the modified mount model optimized at two different lowest mode frequencies.
SGH modelModified Mount Optimized to 2.2 HzModified Mount Optimized to 3.3 Hz
Mount Mass647 tonnes947 tonnes1840 tonnes
1st mode2.17 Hz2.203.30 Hz
2nd mode2.24 Hz2.253.31 Hz
3rd mode2.70 Hz3.424.77 Hz
4th mode3.09 Hz5.738.28 Hz

4.4.5.2 Use of Advanced Materials

A number of materials are available on the market today that may be beneficial to the performance of the telescope structure. Some of these include composite materials that allow higher stiffness properties compared to steel and aluminum, and at substantially lighter unit weights; smart materials that can be used to control deflections; materials that increase the damping of selected structural vibrations; and materials with significantly smaller coefficients of thermal expansion to reduce thermal deflections.

One key area in which composite materials may be used is in the secondary support tripod trusses. There are two immediate goals to be achieved: (1) elimination of the cables used in the SGH model to provide out-of-plane stiffness to the trusses, and (2) lightening of the trusses to both increase the first mode frequency of the telescope and to reduce the amount of the OSS counterweight.

The SGH point design used standard rectangular structural tubes with dimensions of 6 in. by 12 in. with a 3/16-in. wall thickness, with the 12-in. dimension perpendicular to the plane of the truss. Simply removing the cables reduced the first mode frequency to 0.64 Hz, down from 2.17 Hz. The motion in the telescope structure in the first mode was almost completely in the trusses. Using a composite material with a modulus of elasticity of 44 x 106 psi (steel: 29 x 106 psi) and a density of 103.6 lb/cu. ft. (steel: 490 lb/cu. ft.), and maintaining the structural dimensions of the 6 in.x 12 in. x 3/16 in. tube resulted in a first mode frequency of 1.78 Hz, with no cables. Increasing the dimensions of the rectangular structural tube to the equivalent of a rectangular tube with dimensions of 6 in. x 18 in. x 5/16 in. led to a first mode frequency of 2.31 Hz. With the tube having the dimensions of 6 in. x 18 in. x 5/16 in., the movement in the telescope was in overall rocking of the telescope, with very little strain energy in the trusses.

4.4.6 THE NEXT STEPS

The point design developed by Simpson, Gumpertz, and Heger has provided the starting point to evaluate the expected structural behavior needed in developing other aspects of the telescope design. The next step is to further refine the SGH finite element model so that a better understanding of the static and dynamic behavior of the primary mirror segments and secondary mirror under changing gravity and representative wind loads can be obtained.

The current finite element analyses use a separate finite element model to predict the behavior of the individual mirror segments within a primary mirror raft structure. A second finite element model for the overall telescope structure uses a single lumped mass element with the overall mass and inertia properties of a raft to represent each primary mirror raft. Although this describes overall telescope structure behavior, it does not provide sufficient information at the individual mirror segment level. The individual mirror segments, on their separate support systems, need to be included in the telescope structure model to provide this detail.

The secondary mirror, with its supports, is also represented by a single lumped mass element. A more detailed model of the mirror, along with its supports, is required to better describe the deflections and motions at the secondary mirror.

As a concurrent effort, the fidelity of wind loadings used in the finite element analyses will be improved. The dynamic wind loadings measured at the Gemini South Observatory, which are described in Section 5.5 of this document, have provided valuable information about the actual loadings that occur on a telescope and within an enclosure, from wind with a measured velocity and energy spectrum flowing over the site. Those measurements have provided information that will allow us to (1) extend the bandwidth of the frequencies of wind forces loading the telescope, (2) determine more accurately the forces acting on individual elements of the telescope arising from wind moving into the enclosure and over the telescope structure from a representative site wind, and (3) analyze the behavior of the telescope considering the correlation of the dynamic forces of the wind and the distance over which those forces remain correlated.


March 2002