Transcript Document

Recent Design Improvements to the Canadian Light Source
(CLS) Optical Metrology Facility Long Trace Profilometer
*
*
*
Brian Yates , Alan Duffy , Shawn Carriere (* CLS)
Motivation
A design upgrade was recently initiated for the long trace profilometer
(LTP) at the Canadian Light Source that allows radius of curvature
measurements, on large mirror benders, as a function of bending force
prior to beamline installation. The LTP vertical clearance can now handle
mirror benders up to approximately 500 mm in height. A decoupled
external enclosure reduces air convection current effects, and vertical
adjustment of the main beam can be safely performed by one person, with
no realignment of the LTP reference light beam. Design criteria and recent
stability scan results will be discussed.
Motivation
The need arose from two CLS hard x-ray beamlines that use relatively
large mirror bender mechanisms in their designs. There was a requirement
prior to installation to measure these mirror bender systems to determine
the radius of curvature as a function of bending force, and to verify that the
overall slope error (mirror plus bender) met specifications. A design that
provides adequate vertical clearance between the LTP optical head and the
mirror bender system was required. Fig. 1 shows the LTP prior to the
system upgrade.
Figure 1: Front view of the CLS Long Trace Profilometer prior to the system
upgrade shown with a small mirror under test.
Design Criteria
• Mirror bender mechanisms up to 500 mm in height should be
measurable.
• More rigid support of the LTP main beam to achieve sub-microradian
stability.
•Eliminate realignment issues of the reference light beam with a change in
LTP main beam height.
• An external enclosure that is decoupled from the LTP and vibration
isolation table to reduce air convection currents.
• All operations of the LTP should be directly viewable and accessible from
the front access doors.
• Vertical adjustment of the LTP main beam can be safely performed by one
person.
Our Operating
Funding Partners
Discussion
The design used extruded aluminum beams
from MK Profile [1] to build a rigid support
system for the main LTP beam and
enclosure. The surrounding enclosure is
isolated from the vibration isolation table and
LTP main beam by a small clearance gap.
The enclosure has three floor-mounted
support legs and a fourth support at the back
wall. Transparent 0.25” polycarbonate sheets
enclose the top and side panels of the
external isolated enclosure, using rubber
grommets for additional vibration isolation.
Fig. 2 shows the LTP and enclosure with a
mirror bender under test, and Fig. 3 shows a
back view of the LTP.
The two aluminum towers supporting the LTP
main beam were precision machined from
solid aluminum 4”x4” bar (Type 6061-T6),
with thirty five precisely drilled holes in each,
allowing LTP beam height adjustments over a
total range of 510 mm, to within 15 mm. An
expanding pin is used to align both sides of
the main beam to equal height, to better than
1 mrad levelness. Once level the two bolts
beside each expanding pin are tightened to
lock the LTP main beam down, and the
expanding pins removed.
The reference
mirror is mounted fixed relative to the main
beam,
eliminating
reference
mirror
readjustments. Fig. 4 shows an expanding
pin locked in one of the hole.
A removable worm gear hand winch
(McMaster-Carr Part. No. 3205T16, load
capacity of 2000 lbs.) was used to lift the
main LTP beam using 3/16” stainless steel
braided cables. The hand winch employs a
41:1 gear ratio and is of a “no back drive”
design so a single person can raise or lower
the LTP main beam.
Two cables are
employed on the worm gear – one to raise
the right portion of the LTP beam, the other
to simultaneously raise the left portion
through a longer cable and adjustment
turnbuckle.
The turnbuckle allows the
tension of the two cables to be matched.
Typically the entire worm gear mechanism
and cables are removed once the desired
height has been achieved.
[1] MK is the copyright for Maschinebau Kitz.
Stability Tests
A series of stability tests were carried out as a function of scan duration time
to estimate our “noise floor”, keeping the LTP position fixed. For this test
essentially “identical” super-polished silicon mirrors from General Optics
(Catalogue Number GO-S100-1) were used for both the reference beam and
the sample. Fig. 5 shows the LTP stability test with scan duration times
ranging from 30 to 600 seconds.
Figure 2: Front view of the upgraded CLS Long Trace Profilometer (front
doors open), showing adjustable height adjustment and external decoupled
enclosure. Under measurement is a 1.1 m length vertically collimating mirror
bender system mounted on an adjustable tip/tilt stage.
Figure 5: LTP stability versus scan duration time. The sample and reference beam data have been shown separately for
comparison purposes, and have had a best fit constant removed. The ±1s error bars shown were determined from seven
measurements.
Figure 3: Back view of the upgraded LTP showing the rigid support structure
for the main beam (before external decoupled enclosure assembly). The
worm gear hand winch (center portion) and wire/pulley system (middle-top
portion) raises or lowers the main LTP beam to the desired height.
Figure 4: Front view of the left tower support structure showing the
wire/pulley system and anchor shackle/eyebolt that can be used to raise or
lower the main LTP beam. Also visible are the precision drilled holes in the
tower (15 mm spacing over a range of 510 mm) to which the height can be
adjusted, and the lever/expanding pin that is used to align and level both
sides to better than 1 mrad, before lock-down.
It is interesting to note that although the average sample beam slope error
contribution is 83 nrad RMS, the “bottleneck” appears to lie more with the
reference beam which has a much higher average slope error contribution of
270 nrad RMS. Typically when measuring a mirror the two signals are either
added or subtracted from one another (phasing depends on the setup) to
correct for the fluctuations in the straightness of the LTP main beam. As a
result an actual mirror measurement is currently limited to this noise floor of
~300 nrad. A series of test fans was set up to mix the air through which the
reference beam passes, to break up convection cells and avoid air
stratification, but the stability tests were either unchanged or slightly worse
than that shown in Fig. 5. In order to further reduce the LTP noise floor we
are looking into improving our room temperature stability (from ±0.1 to ±0.01
degrees centigrade), and replacing the main LTP aluminum beam with a
ceramic beam from Coorstek, Inc., with substantially better flatness of ~5
μrad RMS, higher stiffness, and lower coefficient of thermal expansion.
Acknowledgements
• The CLS is supported by NSERC, NRC, CIHR, and the University of
Saskatchewan.
www.lightsource.ca