Design and analysis of the 2.5-m telescope instrument
rotator
Sloan Digital Sky Survey Telescope Technical Note
19980413
Walter
Siegmund, R. French Leger, Michelle
Pillers and Patrick
Waddell
Contents
Introduction
Virtually the entire rear surface of the 2.5-m telescope primary
support structure is covered by the Cassegrain instrument rotator. It
supports the SDSS camera and the spectrographs. For spectroscopy, a
plug-plate cartridge is mounted on the rotator in lieu of the camera.
The cartridge mates to the same pseudo kinematic mounting surfaces
that locate the camera and is coupled directly to the instrument
rotator bearing that is mounted to the primary mirror cell.
Consequently, the deformation of its mounting surface during
operation is expected to be inconsequential.
The spectrograph mounting is another matter. A fiber slit head
assembly, on either side of each cartridge, is inserted into a socket
in the corresponding spectrograph during the mounting process. Each
fiber slit head is spring loaded against kinematic mounting surfaces
within the spectrograph socket. As a result, some motion of the
spectrograph with respect to the cartridge is acceptable during
operation, e.g., due to varying gravity loads. We have specified no
more than +/-0.030" of relative motion at the spectrograph socket
opening between the cartridge and the spectrograph due to the
deformation of the instrument rotator. This applies for elevation
angles from zenith to horizon and for any rotator angle.
The existing instrument rotator has several shortcomings that are
corrected in the current design.
- In the current design, the moments due to the spectrograph
weights is not applied to the instrument rotator bearing. This
minimizes the deformation of the camera pseudo kinematic mounting
surfaces that might otherwise cause a focus gradient across the
camera.
- The new design provides more space for the rotator cable
wrap.
- The outer cylindrical drive surface must be ground very round
and smooth. It is hardened to allow this and to increase its
resistance to wear and surface deformation. The current design has
greater out-of-plane stiffness that is intended to facilitate the
heat treatment of outer cylindrical surface to increase its
hardness without causing excessive deformation of the part.
The instrument rotator consists of three modules.
- The hub bolts to the central instrument rotator bearing that
has a bore of 0.851 m and an od of 1.016 m. This bearing
constrains the rotator in x and y but allows it to rotate freely
about z. The hub supports the camera and the camera saddle.
- The rotator ring girder is concentric with the hub and
supports the spectrographs. It is constrained in z at three equal
intervals by cam rollers at its OD. It is driven by another roller
that bears on its outer precision ground cylindrical surface.
Large perforations in the ring girder allow access for
installation of equipment and maintenance.
- The membrane connects the hub to the ring girder. It keeps the
ring girder centered but allows it to be guided by the cam
rollers. It is intended to have very little out of plane stiffness
so as not to apply moments to the instrument rotator bearing.
Finite Element Analysis (FEA) is used to determine the amount of
distortion the ring girder will experience from the weight of the
spectrographs as it and they move about both the altitude and rotator
axes.
Description
All components are steel. The drawings show a front,
cross-sectional and back view of the rotator (Figure 1, Figure 2 and
Figure 3 respectively). The drawing is accompanied by a DXF file that
is dimensionally accurate. The major components appear on different
layers of the CAD drawing in DXF
format. Dimensions may be taken from this file.
- Hub (blue) - anchored to the rotator bearing with 12, 3/8 X 16
SS bolts. It is 3/8" plate welded to a 2.3 x 3.75 ring. It bolts
to the rotator bearing assembly. This interface is very rigid and
can be assumed to be fixed in x, y, and z.
- Front plate (black) - Part of the ring girder, it is a plate
constructed of 11 gauge (.125") steel sheet with six - 12" dia. -
access holes, two oblong access ports and six 4" dia. mounting
pads. The mounting pads incorporate two 3/8 X 16 tapped holes each
to provide attachment for the spectrographs.
- Back plate(black) - Part of the ring girder, this is a replica
of the front plate.
- Radial webs (cyan) - Part of the ring girder, this is a series
of internal webs made from 11 gage sheet steel (.125") and
perforated with 3" dia. oblong holes. All webs are to be welded to
both the front and back plates, inside web and drive ring. Welds
are to be 2" long spaced on 2" intervals.
- Inside web(black) - This is the inside surface of the ring
girder. It is made from 11 gage sheet steel (.125") and perforated
with 3" dia. oblong holes. This web is to be welded to both the
front and back plates and radial webs. Welds are to be 2" long
spaced on 2" intervals.
- Drive Ring (black) - Part of the ring girder, this is an outer
ring constructed from 2" x 5" bar stock and machined to the
dimensions shown. Continuously welded to front and back
plates.
- Membranes (green) - Four trapezoids of 1/8" sheet continuously
welded to the hub and front plate at four locations. The only
connection between the hub and the ring girder is through the four
membranes.
Figure 1: Instrument rotator front
view.
Figure 2: Instrument rotator
section.
Figure 3: Instrument rotator back
view.
Dimensional stability of the spectrographs is a high priority,
i.e., the deformation of the instrument rotator due to changing loads
and the differential thermal expansion of the spectrographs and
instrument rotator should not apply significant forces to the
spectrographs. Consequently, attachments of the spectrographs to the
instrument rotator are made with linear bearings (dry
polymer-lubricated journal bearings) that decouple one force
component and all three moment components over small angles. They
allow differential expansion of each spectrograph with respect to the
instrument rotator, e.g., due to the difference of the thermal
expansion coefficients of the aluminum spectrograph and the steel
instrument rotator, carry large loads and are robust.
Each spectrograph is attached at three points indicated by a pair
of bolt holes (1.5" separation) circumscribed by the 4" diameter
circle that indicates a raised mounting pad that is machined coplanar
with the other mounting points (Figure 1). The linear bearings at the
extreme y locations allow translation in y; those at the extreme x
locations allow translation in x. This is indicated by the
orientation of the bolt hole pairs in the figure. Each bearing allows
translation in the direction normal to the line connecting the bolt
hole pair.
In the front and section AA views (Figure 1 and Figure 2), the
spectrographs are represented by two point masses attached by
weightless struts to the mounting pads. This model was used to
determine the forces applied to the spectrograph mounting points for
each load case. In-plane forces calculated from this model are larger
than are likely to be present at the actual attachments and are
conservative. The results are not much affected since the rotator has
high in-plane stiffness.
The deflection of the model and stresses were calculated for three
load cases.
- The three point supports are as shown. Gravity is in the +Z
direction (out of the page in the front view).
- The three point supports are rotated 30°
(counterclockwise in front view). Gravity is in the +Z direction
(out of the page in the front view).
- The three point supports are as shown. Gravity is in the +X
direction.
To summarize the constraints:
- Nodes on the interface of the hub to the telescope are
constrained in x, y, and z.
- Nodes at the three point supports are constrained in z.
- No moments at individual nodes are constrained.
Results
The displacements are shown in Figure 4, Figure 5 and Figure 6.
These figures are linked to animations of the deflections that are
exaggerated so as to be apparent (to view, click on the figure).
Stresses in the model are low as compared to the yield strength of
low carbon steel, i.e., 30000 psi. For example, the maximum stress
for case 1 was 2390 psi. Low stresses are to be expected in a design
where stiffness rather than strength is the more stringent
criterion.
Figure 4: Load case 1 (axial
gravity). An animation of the deformation is available.
Displacements in z, the axial direction, are in inches.
Figure 5: Load case 2 (axial
gravity).
Figure 6: Load case 3 (transverse
gravity).
Verification
A similar model was used to analyze a flat circular annular plate,
fixed on the outer edge and guided axially on the inner edge (Table
24, Case 1f, Warren C. Young, Roark's Formulas for Stress and
Strain, 6th Edition, McGraw-Hill, New York, 1989, p. 402.) The
material properties and overall geometry were the same as in the
rotator model. The details were different. The annular plate was
0.125 inches thick whereas the rotator is a hollow weldment with
stiffeners and holes.
The deflection of the inner edge of the annular plate model was
-0.093 inches. That calculated with formulas in Young was -0.096
inches. This agreement is more than adequate for our purposes.
Conclusions
The maximum axial displacement range of the instrument rotator is
a bit under 0.016" (case 1). The other cases give similar but smaller
values (the maximum for case 3 should be doubled to get the range
since the rotator has an angular range of more than 360°).
Obviously, this is an upper limit on the differential axial
displacement of the three mounting points for each spectrograph. The
spectrograph socket opening is 19.00" below the spectrograph mounting
points. On the instrument rotator, the mounting points have a maximum
extent of 15.95" by 62.97". Consequently, a differential axial
displacement of the mounting points causes a lateral displacement of
the socket opening of similar or smaller magnitude. This implies that
all motions of the spectrograph socket opening due to variations in
gravity loads are expected to have a range of less than 0.016", well
within the value specified in the introduction. The calculated
stresses in the model are low compared to the yield strength of low
carbon steel.
Yet to be added to the design is a cylindrical mounting surface
for an optical encoding system. This addition is not expected to
affect the deflection of the rotator or change its geometry
significantly. The surface will likely be added near the R35.6" inner
surface of the ring girder. About 224" of tape will be required at
this radius.
Date created: 4/14/98
Last modified: 4/14/98
Copyright © 1998, Walter A. Siegmund
Walter A. Siegmund
siegmund@astro.washington.edu