2.5-m telescope test plan
Sloan Digital Sky Survey Telescope Technical Note
19961009
Walter
Siegmund
Contents
Introduction
The systems that comprise the 2.5-m Sloan Digital Sky Survey
(SDSS) telescope must be tested as installation and integration
occurs so that problems are discovered at the component and subsystem
level rather than at the system level where debugging is more
expensive and time consuming. Component testing allows problems to be
discovered earlier when they are less likely to cause major project
delays.
This document complements the telescope tests
dependency chart that is available as a postscript file. The
chart provides guidance as to the order of the tests and the state of
telescope assembly required for each test, whereas this document
describes each test.
Definitions
These are definitions of common parameters that are used in the
description of the tests.
a Telescope altitude angle
A Telescope Azimuth angle (south = 0°, east = 90°)
W Mean wind speed
Z Mean wind direction (south = 0°, east = 90°,
prevailing wind is 215° < Z < 360°)
X1,Y1, Telescope enclosure coordinate system.
Z1 +X1 == Z = 60°, i.e., toward east side of the building.
+Y1 == Z = 330°, i.e., toward north side of the building.
+Z1 == azimuth axis (zenith direction).
+Z1 = 0 at floor level.
X2,Y2, Telescope azimuth coordinate system.
Z2 +X2 == right altitude axis (wind baffle drive side).
+Y2 == telescope boresight at a = 0°.
+Z2 == azimuth axis (zenith direction).
X3,Y3, Telescope altitude coordinate system.
Z3 +X3 == right altitude axis (wind baffle drive side).
+Y3 == front of primary support structure.
+Z3 == telescope boresight at a = 0°.
Measure natural frequencies: I
Stability of pier
Purpose of test
- Measure the rocking stiffness of the pier so that (hopefully)
it can be eliminated as a significant source of wind-induced
tracking error.
- These parameters are inputs to the design of the axis
servos.
- Note that the torsional stiffness is not measured. It is not
likely to be important, can be estimated, and is hard to
measure.
Criterion
- The pier should contribute negligibly to wind-induced tracking
error, i.e., 5 mas 2D RMS at W = 10 m/s.
Conditions
- W > 5 m/s; 215° < Z < 360° a = 0°; A =
240°, 150° enclosure open and closed Special tools
- 3 ea accelerometers 1 ea LabVIEW 1 ea A/D card 1 ea cup
anemometer 6 ea wind dynamic pressure sensors
Record
- +x1 and +y1 acceleration at the top of the pier dynamic
pressure near the primary mirror platform (PMP)
Calculate
- Power spectral densities
- Transfer function
Mount stiffness
Purpose of test
- Measure the response of the mount to the wind.
- Determine the locked rotor resonance frequencies and other low
frequency resonances that are needed for the design of the axis
servos.
- Verify that no excessively compliant components are
present.
Criteria
- Natural frequencies are to be above 5 Hz.
- The secondary should contribute negligibly to wind-induced
tracking error, i.e., 5 mas 2D RMS at W = 10 m/s, with the
wind-baffle present. Since the wind-baffle is expected to
attenuate the dynamic pressure of the wind by a factor of 7, 35
mas 2D RMS is allowable at W = 10 m/s for this test.
Conditions
- W > 5 m/s; 215° < Z < 360° a = 30°,
80°; Z - A = 0°, 90°
- enclosure open and closed
- motor shafts locked Special tools
- 3 ea accelerometers 1 ea LabVIEW 1 ea A/D card 1 ea cup
anemometer 6 ea wind dynamic pressure sensors
Record
- +x2 and +y2 acceleration of the fork base
- angular acceleration of the fork base about the azimuth
axis
- angular acceleration of the PMP about the altitude axis
- angular acceleration of the 2ry about the x3 axis
- angular acceleration of the 2ry about the y3 axis
- dynamic pressure near the PMP
- dynamic pressure near the 2ry
Calculate
- Power spectral densities
- Transfer function
Measure internal baffles
Purpose of test
- The criteria are for the worst case tolerances and flexure
over 30° < a < 90°. They refer to departures from
the pupil obstruction in the baffle design.
- The baffles must be installed so that they block all direct
rays to the focal surface.
- They must not vignette rays in the photometric field of
view.
- Obstruction of the edge of the entrance pupil at the edge of
the astrometric field of view must be less 2%. The RMS image
diameter at the edge of field 18 is 1.4 arc sec. If 2% obstruction
causes a 2% shift in centroid, this would result in a 28 mas
centroid error. This is small enough to be acceptable.
Criteria
- Parts must satisfy tolerances on shop drawings.
- The primary and secondary baffles are to be located to 2 mm
true position referred to their respective optical surfaces.
- The conical baffle is to be located to 3 mm true position at
30° < a < 90°.
Conditions
- W < 5 m/s; Z don't care
- a = 0°, 90°; A don't care
- enclosure open
Special tools
- piano wire
- protective cover for the primary mirror
- planks Technique
- Use a tape measure, piano wire and scale to measure
positions.
Record
- X3,Y3,Z3 of points on primary, secondary, and conical
baffles.
Comments
- Measurement at a = 90 degrees will require a protective cover
for the primary mirror and planks supported by the telescope to
walk on.
Measure spectrographic field distortion
Purpose of test
- The coordinate transformation from the sky to the focal plane
for each spectrographic target requires that the spectrographic
field distortion be measured.
Criterion
- In the fiber position error budget, the budget for errors in
the field distortion is 4 µm RMS. Note that this is a change
from the earlier value of 1 µm RMS (Table 4 of "Results
of the 3/93 Drilling Tests" Sloan Digital Sky Survey Telescope
Technical Note 19930430) that neglected systematic errors and
flexure. The residual error from independent measurements fit to
the same distortion model should satisfy this criterion.
Conditions
- W don't care; Z don't care
- Sky condition; clear, good seeing
- a = 30°, 90°; A don't care
- enclosure open
Special tools
- Spectrographic distortion plug-plates
- Spectrographic distortion coherent fiber bundle
- Guide camera with centroid software
Technique
- Feed guide camera with coherent fibers plugged into
spectrographic distortion plug-plate.
- Make multiple observations of an astrometric field of stars
using several rotator angles.
Record
- Use guide camera to measure centroids of stars.
- Fit centroids to distortion model to calculate xd, yd, a1, a3,
a5, a7, i.e., the center of the distortion pattern and the
coefficients of the first four terms of an odd power series
expansion.
- Verify that xd and yd are inconsequential as they should be
for collimated and aligned optics.
- Note that a1 is the scale which can be adjusted for each
plate.
- Compare measured coefficients to those from modeling the
as-built optical system.
- Verify stability of distortion with elevation, rotator angle
and time.
Comments
- Another means of performing this task is to use the science
fibers and to offset the telescope one fiber radius (1.5 arc
seconds) in four directions. This requires special purpose
software to analyze the resultant data from the spectrograph and
requires that the PSF be known.
Scattered light
Purpose of test
- Scattered light affects the photometric accuracy of the imager
and the spectrophotometric accuracy of the spectrographs. This
test will discover design errors, positioning errors, excessive
flexure, and flaws in the baffle coatings.
Criteria
- The point source normal incidence transmittance (PSNIT, the
ratio of flux in the focal surface to incident flux) should be
less than 2x10^-6 for sources more than 30° off axis.
- For sources less than 30° from the field center, the
PSNIT can be higher, but it is desirable that the focal surface
illumination be uniform. The wings of the point spread function
(PSF) should fall smoothly with field angle (I.
King, PASP 83, 199, 1971).
Need better wording and more quantitative
criteria.
Conditions
- W < 5 m/s; Z don't care
- a = 30°, >80°; A don't care
- lunar phase within 60 degrees of full at suitable elevation
angles
- Jupiter, Venus, Sirius, ..., at suitable elevation angles
- enclosure open
Special tools
- integrating CCD cameras and mounting
- script to move telescope and take baffle images
Technique
- Make finding chart. Take images of baffle system viewed from
focal plane both on-axis and at field edge. Use diffuse sky
illumination. Label all baffles.
- Visual inspection. From focal plane, especially field edge,
look at baffles. Start with diffuse illumination, then solar
(field angles > 30° for safety), then lunar. Look for gaps
in baffles and glints.
- Use script to take integrating CCD images of baffles from edge
of focal surface ("Detection of scattered light in telescopes",
Frank Grundahl and Anton Norup Soerensen, Astronomy and
Astrophysics Supp., 116, 367-371). Begin with the moon
at 1.8°, 2.7°, 4.5°, 7.4°, 12.1°,
20°, 25°, 30°, 35°, 40° and 50°
off-axis and left, right, up and down with respect to the
boresight. Look for asymmetries, glints and deviations from
Breault Research Organization report dated Aug 5, 1996.
- Take images along four radii centered on the moon (left,
right, up and down) at 0.25° intervals from 1.8° to
20° off-axis (264 images at 10 seconds per image is 44
minutes). These images provide a permanent archive that can be
consulted if a scattered light problem is suspected.
- Using photometric camera, take images with Jupiter or Venus
just outside the field of view. Rotate camera 180° between
images and compare images to get the large field angle PSF.
- Obtain PSFs over the photometric field of view, at least one
per photometric CCD.
Record
- Images from items 3 and 4.
- PSFs from items 5 and 6.
Comment
- It would be helpful to have software tools to reduce the
baffle images to determine quantitatively the amount of scattering
from each baffle surface.
Test collimation
Purpose of test
Align optics with the instrument rotator axis. This test insures
that image quality is not compromised by optics misalignment. Also,
it insures that the point spread function depends only on field
radius.
Criteria
On the surface of best focus, image size and aberrations should be
only a function of field radius. The image shape should be consistent
with the optical design convolved with the seeing.
Conditions
- W < 5 m/s; Z don't care
- a = 30°, >80°; A don't care
- seeing < 1 arc second
- enclosure open
- spectrographic corrector installed
Special tools
- Collimation computer
- Integrating CCD camera and mounting
Technique
- Coma: At field center, defocus until central obscuration is
visible in the image. Insure that central obsuration is centered
in image (“Proposed Process for 2.5m Collimation” E.
Mannery, 18 July, 1996 and "Collimation of Fast Wide-Field
Telescopes", Brian A. McLeod, PASP 108:217-219, 1996
February).
- Astigmatism: At field edge, defocus until astimatism can be
measured. Insure that astigmatism is independent of rotator
angle.
- With the telescope in focus, verify that the PSF core is
consistent with the optical design convolved with the seeing over
the field of view.
Record
- Magnitude of residual coma
- Uniformity of astigmatism
- Uniformity of PSF
Test mirror ventilation
Purpose of test
- Mirror temperature nonuniformities cause figure errors.
- Thermal inertia contributes to temperature differences between
the mirror surface and the ambient air and thereby contributes to
image degradation.
- Axial temperature gradients change the power of optics,
require refocussing and change the scale of the focal surface
thereby degrading astrometric precision.
Criteria
- The temperature uniformity of the primary mirror must be 0.1
°C peak - valley after the axial gradient and mean face plate
and back plate temperature distribution is removed.
- The maximum mirror surface - air temperature difference is
0.25°C.
- These criteria are for an ambient temperature gradient of
0.25°C/hour.
- In terms of time constants, the thermal time constants of the
face and back plates must be uniform to 0.4 hours peak - valley
after the axial gradient and mean face plate and back plate
thermal time constant distribution is removed.
- The mean thermal time constant of the face plate may not be
larger than 1 hour.
- For the secondary, the mean thermal time constant of the face
plate may not be larger than 1 hour.
- The thermal time constant distribution need be uniform to 0.8
hours peak - valley.
Conditions
- W don't care; Z don't care
- a = 30°, >80°; A don't care
- enclosure don't care
Special tools
- temperature measurement system
- temperature analysis software
Technique
Record
- Thermal time constant distributions and temperature v. time
for the primary and secondary.
Comments
- These tests should indicate if active ventilation of the
secondary is needed to achieve the project astrometric goal.
- Several days of data acquisition will be necessary, more if
problems are found. Data acquisition can occur in parallel with
other tests.
Date created: 10/9/96
Last modified: 10/9/96
Walter A. Siegmund
siegmund@astro.washington.edu