PALM-3000 High-Order Wavefront Sensor Alignment Guide

sunglowcitrineUrban and Civil

Nov 15, 2013 (3 years and 11 months ago)

175 views

-

COO
-


CALTECH OPTICAL OBSERVATORIES

CALIFORNIA INSTITUTE OF TECHNOLOGY






Rev 1.0 10/28
/0
9












Caltech Optical Observatories

California Institute of Technology

Pasadena, CA 91125


Caltech Instrumentation Note #636




PALM
-
3000 High
-
Order Wavefront Sensor
Alignment Guide



C.
Baranec






2

Abstract


This document describes the

step
-
by
-
step alignment procedure for the high
-
order wavefront
sensor for PALM
-
3000 as well as

other supporting information
.


Revision Sheet


Release
No.

Date

Revision Description

Rev. 0.1

07/18/09

Initial draft
of outline
by C. Baranec
.

Rev. 0.2

07/19/09

Finished up to 4.3.1 with 4.3.2 partially finished.

Rev. 0.3

10/15/09

Modifications and
extensions during October alignment.

Rev. 1
.
0

10/28
/09

Major alignment revisions and additions.

Procedure leading to the
October 18, 2009 alignment of the HOWFS is fully documented.

















3

Table of Contents

1

Introduction

................................
................................
................................
.............................
4

1.1

Acronyms and Definitions
................................
................................
................................
.........

4

1.2

Purpose

................................
................................
................................
................................
.......

4

1.3

Scope

................................
................................
................................
................................
...........

4

1.4

Related Documents

................................
................................
................................
....................

4

1.5

Optical Design Summary

................................
................................
................................
..........

4

1.5.1

Basic System Parameters

................................
................................
................................
.......................

4

1.5.2

Design and component functions

................................
................................
................................
...........

5

2

Telescope Simulator

................................
................................
................................
................
5

2.1

Overview

................................
................................
................................
................................
.....

5

2.2

Alignment

................................
................................
................................
................................
...

6

3

LabVIEW Alignment Tool

................................
................................
................................
......
8

3.1

Overview

................................
................................
................................
................................
.....

8

3.2

Procedure fo
r starting

................................
................................
................................
...............

8

3.3

Explanation and use of the tool

................................
................................
................................

9

3.3.1

Camera and image controls

................................
................................
................................
....................

9

3.3.2

Camera display

................................
................................
................................
................................
....

10

3.3.3

An
alysis display

................................
................................
................................
................................
...

10

4

HOWFS

................................
................................
................................
................................
.
10

4.1

Optical Design

................................
................................
................................
..........................

10

4.2

Mechan
ical setup

................................
................................
................................
.....................

12

4.3

Optical alignment

................................
................................
................................
....................

12

4.3.1

Definition of focal point and optical axis

................................
................................
.............................

12

4.3.2

Alignment of collimating and fold mirrors

................................
................................
..........................

13

4.3.3

Installation of relay optics

................................
................................
................................
....................

15

4.3.4

Installation of microlens arrays and exchanger

................................
................................
....................

17

4.4

Optical testing and final alignment

................................
................................
........................

19

5

Appendix

................................
................................
................................
................................
21

5.1

Discussion of
additional errors

................................
................................
...............................

21

5.1.1

Spherical (and other high order radially symmetric modes)

................................
................................

21

5.1.2

Coma

................................
................................
................................
................................
....................

22

5.1.3

Astigmatism

................................
................................
................................
................................
.........

22

5.2

Appendix A


October 18
th
, 2009 alignment

................................
................................
.........

22

5.3

Appendix B


HOWFS alignment table worksheet

................................
..............................

23






4

1

INTRODUC
TION

1.1

Acronyms and Definitions

FWHM



Full
-
Width at Half
-
Maximum

PSF



Point Spread Function

RMS



Root Mean
-
Squared

HOWFS



High
-
order wavefront sensor

sXX

Spatial sampling mode of the HOWFS. X
X denotes the number of
samples, either 64, 32, 16 or 8.

subaperture size


the size of a lenslet
-
array lenslet, projected onto the detector

1.2

Purpose

The purpose of this document is
to present the step
-
by
-
step alignment procedure for the high
-
order wavefront sensor for PALM
-
3000 as well as other supporting informa
tion
.

1.3

Scope

This document
attempts to provide all of the information required, including alignment tools and
supporting software
,

for optical alignment of the high
-
order wavefront sensor
.

1.4

Related Documents




C. Baranec
,

High
-
order wavefron
t sensing system for PALM
-
3000,”

Proc. SPIE
Adaptive
Optics Systems, eds. N. Hubin, C. Max & P. Wizinowich, 7015, 2008.



A. Bouchez, R. Dekany, J. Angione,
C. Baranec
,

K. Bui, R. Burruss,
J. Crepp, E. Croner,
J. Cromer, S. Guiwits, D. Hale. J. Henning, D. Palmer, J. Roberts, M. Troy, T. Truong &
J. Zolkower,
“Status of the PALM
-
3000 high
-
order adaptive opt
ics system,”

Proc. SPIE
Astronomical Adaptive Optics Systems and Applications IV, eds. R. Tyson & M. Hart,
7439B, 2009.

1.5

Optical Design Summary

1.5.1

Basic
System
Parameters


PALM
-
3000

Plate scale:

0.390 mm/arc second

F/#
:


15.77

Number of illumin
ated subapertures in s64 mode:



63.95 wide by
62.88 tall. (
Centered on a lenslet array
)

Telescope pupil reimaged to lenslet array.

Bandwidth:

450
-
95
0 nm








5

1.5.2

Design and component functions


The entrance to the wavefront sensor will be fitted with a field
stop that will also function as a
spatial filter. The

spatial filter consists of two blades forming a square aperture which can be
adjusted remotely in size via a piezo
-
electric

flexure mechanism.

The spatial filter used in
PALM
-
3000 will have a maximum a
pert
ure size of 4 arc sec (1.56 mm)
for guiding on solar
system objects, and a minimum size of 0.48 arc sec (190 μm) to filte
r out spatial frequencies
above
λ/d for the 16 across pupil sampling mode. Support for spatial filtering for the 8 across
pupil sam
pling mode will not be
supported because aliasing error will not be the dominant error
term in wavefront reconstruction.


The second element in the system is a reflective collimator
which images the pupil onto
the
microlens array. The

collimator is reflect
ive to control errors in the pupil size as a function of
color over the large bandwidth of the sensor. It

also introduces enough astigmatism error to
match the geometries of the projected pupil
, at a 10.5
°

angle
-
of
-
incidence,
on the deformable
mirror to th
e

projected pupil on the microlens array. A flat

is included to fold the
beam to ease
packaging.


The space between the fold flat and microlens array is reserved for a set of optional
filters.


Immediately before the microlens array is a cylindrical lens w
hich removes the
astigmatism induced by the collimating mirror, flattening the local tilts for
each of the

suba
perture

while still allowing the projected pupil geometries to be matched. This ensur
es that
the spots in the Shack
-
Hartmann pattern are located
at the center of pixel boundaries and do not
use up dynamic

range on the wavefront sensor.


There are four different microlens arrays which can be selected to be used for wavefront
measurement.


Downstream of the microlens arrays are the relay

optics and
detector. The pair

of optics
act as a 0.3
2
00

demagnification

relay to reimage the spots created by the microlens array to the
detector.
There is room for a beam
chopper to support range gating of the Sodium laser guide
star
between the relay lenses
.


Note
that an atmospheric dispersion corrector is being considered upstream of the spatial
filter to control the guide

source elongation due to the chromati
c effects of the atmosphere. If
atmospheric dispersion is not controlled at low

zenith angles, the spatial

filter will selectively
clip the low and high ends of the wavelength range.

2

TELESCOPE SIMULATOR


2.1

Overview

The telescope simulator is the main optical tool that will be used in al
ignment of the HOWFS. It
comprises an
on
-
axis laser and an F/15.6175

beam with a pupil i
maged at infinity. Figure 1

shows the optical layout of the
telescope simulator

and figure 2

shows the assembled telescope
simulator.



Figure
1

Optical design of the telescope simulator. These dimensions are for
red
(λ = 632 nm) light
.





6




Figure
2

Assembled telescope simulator in Cahill lab 15.


The telescope simulator is mounted on a 36” by 6” by ½” breadboard. It is supported on the
corners by 3” tall fixed posts. The optical beam height of the optics is approximately 4” above
the breadboard for an overall beam height of ~7.5”. The fiber source
, fed by a 632 nm diode
laser, is mounted on an X
-
Y
-
Z stage which has all of its actuators set to mid
-
range.

2.2

Alignment

In preparation for alignment of the telescope simulator, make sure that all of the components
seen in figure 3 are available. This inclu
des both a red and green diode laser.


1.

Lay out all of the components on the breadboard in approximately the correct positions.
This is to make sure that everything will fit. Try to get the two fold mirrors as close to
the edge as possible.

2.

Define the axis

of the simulator.

a.

Install and bolt down the green laser and fold mirror at the far end of the
breadboard.

(A red HeNe laser can be used instead of the green laser. It will have
a more Gaussian beam profile but it will be more difficult to obse
rve the retr
o
-
reflections off of

refractive surfaces.)

b.

Bolt down the two fold mirrors at the near end of the simulator.





7

c.

Turn the laser on. It may need to be warmed up by gently placing your hand next
to it for several seconds before it starts.

(If using the HeNe, wai
t 30 minutes or
until the laser is in thermal equilibrium; it should be almost hot to the touch.)

d.

Using the tip/tilt adjustments of all of the mounts, establish a beam running from
one end of the breadboard to the other which is parallel to and 4” above th
e
breadboard.

Make sure that the beam coming off of the last mirror is roughly
normal to the end of the breadboard.

3.

Install the fiber source.

a.

With the fiber removed, place the FC fiber chuck in the beam at its approximate
z
-
position. Make sure that the bea
m is normal to the fiber chuck.

b.

Use the x and y actuators to center the hole in the fiber chuck

with respect to the
green laser
.

c.

Insert the fiber into the fiber chuck.

d.

Place the red laser and co
llimating coupler off to the side out of the main optical
path.

Using the tip/tilt adjustments, couple the red laser into the fiber and observe
the output from the other end of the fiber. (Alternatively, use a fiber coupled laser
source.)

e.

Tip and tilt the fiber chuck to make sure the fiber output is approximately

centered with the on
-
axis laser path. Remove the fiber, and re
-
center the fiber
chuck hole on the on
-
axis laser as it will have shifted slight
l
y.

4.

Install the first lens.

a.

Note the orientation of the lens, with the flat side towards the fiber source.

b.

Place

the first lens in the beam at approximately 150 mm from the fiber source
taking care to roughly center the lens on the green laser.

c.

Install the fiber source and check the collimation of the red light with a shear
plate and adjust the z
-
position until the
beam is collimated.

d.

Remove the fiber source and check the alignment of the lens with respect to the
green laser. Adjust the tip/tilt and x
-
y position until the three retro

reflections off
of the lens surfaces overlap at the

fiber chuck hole.

e.

Iteratively re
peat steps b. and c. until the red light is collimated and the lens is
aligned to the green laser.

5.

Install the stop.

a.

Use an inside micrometer to place the stop at the proper z
-
position. For a stop
with a thickness of 0.0315”, the distance should be 5.8320”

(~5.8290” with
PTFE

covering the end of the micrometer
touching the lens surface)
from the first lens
surface to the first plane of the stop.

b.

Roughly c
en
ter the stop on the green laser.

c.

Ensure that the stop is normal to the on
-
axis laser by placing a flat

mirror against
the stop and check the location of the back reflection.

Tip and tilt the stop as
necessary. Shims in the stop holder may be helpful to get proper tilt.

6.

Install the second lens.

a.

Again, note the orientation of the lens with the flat side now
away from the fiber
source.

b.

Using an inside micrometer, place the lens at a distance of 5.8320”
(~5.8290”
with PTFE covering the end of the micrometer touching the lens surface)
away
from the back surface of the stop.





8

c.

Adjust the tip/tilt and x
-
y position o
f the lens until the three retro reflections off
of the lens surfaces overlap and are directed back towards the green laser. It may
be necessary to put a card with a small hole near the stop to see the retro
reflections.

7.

Check focus.

a.

Install the fiber sour
ce into the fiber chuck.

b.

Use a card to estimate the position of the focal point after the last fold mirror.

8.

Align the stop.

a.

Shine the on
-
axis laser normal against a flat surface (lab wall) a large distance
away, and mark its location.

b.

Install the fiber and

look at its footprint. Shift the stop in x and y until the
footprint pattern is centered on the on
-
axis laser location mark.

3

LABVIEW

ALIGNMENT TOOL

3.1

Overview

The
LabVIEW

alignment tool was developed to use feedback from the CCD50 in order to aid in
the ali
gnment of the HOWFS.

This tool is a modified version of the server/viewer software
originally developed by Thomas Stalcup (
tstalcup@keck.hawaii.edu
) for the MMT LGS AO
system. I have since augmented it to provide real time image and wavefront sensor
analysis
including center
-
of
-
mass determination, projection into Zernike and rotation modes and other
merit function capability.

3.2

Procedure for starting


1.

Once the host computer is powered on and started, log on to windows.

2.

Load PDVshow
.

a.

It will prompt for
a progr
am; select the ‘SciMeasure CCD50: Lil Joe 128 x 128 16
Bit CL using Generic DLL’ mode.

b.

Click on the menu bar: Camera


Programming.

c.

Enter ‘@PRG5’ into the box and hit Enter.

d.

Close PDVshow.

3.

Load Slim Joe.

a.

Click on the ‘1246 Hz 128^2’ tab.

b.

Slide the g
ain control to ‘HI’.

c.

Close Slim Joe.

4.

Load the Camera Server.

a.

Click on the CCD_Server shortcut on the desktop. (This launches the
camera_server.vi within
LabVIEW
.)

b.

Once loaded, push the right arrow near the menu bar to run. If the program needs
to be stoppe
d at a later time, push the ‘STOP’ button in the middle of the window.

5.

Load the Camera Viewer.

a.

Click on the P3K_WFS shortcut on the desktop. (This launches the
p3k_viewer_with_recon.vi within
LabVIEW
.)

b.

Once loaded, push the right arrow near the menu bar to

run. If the program needs
to be stopped at a later time, push the ‘EXIT’ button in the upper right of the
window.





9

3.3

Explanation
and use
of the tool

Figure 3

shows the front panel of
the LabVIEW

alignment tool.
The tool is divided into three
general areas:
Camera and image controls in the upper left, camera display on the right and
analysis of the images on the lower left.



Figure
3

LabVIEW

alignment tool front panel/GUI.


3.3.1

Camera and image controls

There are four tabs used for controlling the camera. Gener
ally the Image controls tab is the most
useful and will be all that is covered here.


The ‘Dark Sub’ tab toggles between doing nothing and subtracting the stored dark frame from
the camera data before being sent to the camera display.


The ‘Take Dark’ tab

will create a temporary dark frame to be used by the ‘Dark Sub’ tab. Enter
the number of frames to average over to the right of the tab, turn off the source, and hit the tab.
In the Command Response window, the number of frames will run up to the requested

number
and return ‘dark complete, EOF’ when finished.


The ‘Avg Frame’ tab will, when activated, display an average frame of the indicated number
instead of the live image. Hit the ‘Frame Updates’ tab to restore a live image.






10

The ‘Save Frames’ tabs will
save the indicated number of frames to a file on disk. There will be
a single file [unresolved format


to be updated] with the current dark frame saved as the first
image.


The conversion method controls the image scaling on the right.

This can be set to
a number of
different values. For the ‘Given Range’ option, select the upper and lower limits with the nearby
sliders.


3.3.2

Camera d
isplay

The camera display shows the output of the CCD50 camera. Various controls can change what is
seen in the display.


The ‘
Frame Updates’ tab changes between capturing and displaying new images as they are
captured and halting the display of new frames.


The ‘Boxes’ tab turns the green boxes which show the location of subapertures on and off. The
type and number of boxes will
change depending on the option selected in the analysis display
‘Sampling Selector’ box. For the ‘Frame COM’ mode, various box sizes centered on the display
will appear.


3.3.3

Analysis display

The analysis display will perform a real time analysis of the current frame
shown

in the
camera
display.


The ‘Sampling Selector’ lets the user switch between analysis of the four different lenslet array
modes (s64, s32, s16 and s8) and a ‘Frame COM’ mode.

A slider called ‘Leak’ with values from
0 to 1 will enable leaky integration on all displayed analysis values.



When in any of the lenslet array modes, the histogram window will show the
reconstructed Zernike mode amplitudes in waves at 632 nm
from order
s 1 to 8

with many of
the modes labeled by name. In addition, the RMS slope values in pixels, RMS wavefront
err
or in nm over Zernike orders 1
-
8, RMS wavefront error calculated from the pixel slope
values

and the rotation angle in degrees will be displayed.

The large rotation radial dial is
useful for checking the Shack
-
Hartmann rotation from a distance and displays 100ths of a
degree. The X
-
COM and Y
-
COM values are not calculated in the lenslet array modes.



When in the ‘Frame COM’ mode, only the X
-
COM and

Y
-
COM values will be
displayed. They will indicate the center
-
of
-
mass of the image in units of pixels

from the
center of the detector
.

4

HOWFS

OPTICS

4.1

Optical Design

The optical design and purpose of the high
-
order wavefront sensor (HOWFS) is described in
de
tail in Baranec (SPIE 2008.) In summary, there is first a field stop/spatial filter, then




11

collimating and fold mirror, another fold, a cylindrical lens, lenslet arrays and a relay system
before light is incident on the CCD50 detector. A Zemax layout of the

design with important
dim
ensions is presented in figure 4
.
A picture of the HOWFS is shown in figure 5.



Figure
4

Optical design of the high
-
order wavefront sensor with important dimensions noted.


The
most current

Z
emax design
of the system
is ‘HOWFS_
FINAL_Oct3_2009
.zmx’ and is
embedded in this document below.




Figure 5
Assembled HOWFS in Cahill lab 15. The two simulator fold mirrors are in the lower right.





12


4.2

Mechanical setup

In preparation for alignment, it is necessary to first assemble all of the mechanical parts of the
HOWFS. Most importantly, the large base plate has a smaller interface plate underneath which is
held on by 8x ¼
-
20 bolts. The interface plate should be separ
ated and installed on the optical
bench where the HOWFS will be assembled or realigned. With the central counter
-
bored holes
facing up, attach the plate to 4 fixed posts that each have a height of 2.75”. Clamp the posts onto
the optical bench.
Make sure th
at small optics platform that holds the collimating mirror, fold flat
and cylindrical lens is assembled on the large base plate before proceeding.
Now, place the large
base plate onto the smaller interface plate and bolt them together. When integrating wit
h the
adaptive optics system, the smaller interface plate will be attached instead to the Aerotech ATS
-
150 focus stage.


Attach the Newport
focus stage (UTS50) to the large base plate using
M6

bolts. Attach the focus
stage base plate to the UTS50 using
M5

bolts. Run the stage to its most forward position
(towards the field stop.) Install the lenslet array assembly leaving a small (slightly less than a
business card) gap between the focus stage base plate and the vertical Newport MPA
-
CC linear
stage.

4.3


Optic
al alignment

4.3.1

Definition of focal point and optical axis

The first part of the optical alignment involves defining the focal point and optical axis of the
HOWFS. This will involve adjustment of the telescope simulator location and its output fold
mirrors.
F
igure 6 shows the target point in space for the telescope simulator and HOWFS
common focus.



Figure 6
Solidworks model showing the physical relation between the large base plate and the HOWFS optics.

Measured values are based off of the assembled HOWFS

as of October 2009.


1.

Draw the optical axis and focal point on the HOWFS structure.

(These have been scribes
as of October 2009.)





13

a.

Draw a line parallel to the front of the base plate that is 0.
948
” back from the
front. See the red dY dimension in figure 6.


b.

Mark the optical axis on the front part of the bas
eplate. Draw a line that is 3.12
1”
over from the edge and pointing back. See the blue dZ dimension in figure 6.

c.

On the CCD50 front plates, there should be a scribe mark showing the
horizontal
center of the

CCD50 as well
as an offset line which is 13.983 mm below the
center of the camera head
. If not, then draw these in or scribe them with a height
gauge.

2.

Install the temporary field stop.

a.

Setup an adjustable sized iris (not a zero aperture iris) on an adjust
able height
post.

b.

Close down the iris as far as possible and use a height gauge to set

the height of
the hole to 3.871


c.

Place the iris
on the base plate such that the
opening
is
as close as possible to the
intersection of
the lines drawn on the base plate
in steps 1.a. and b. when viewed
from above.

It may be helpful to use squares lined up with the scribe marks when
viewed at angles.

3.

Positioning of the simulator on
-
axis laser.

a.

Turn on the green on
-
axis laser.

b.


Move the simulator around, and adjust the two output fold mirrors to get the laser
going through the field stop.

c.

Adjust the two mirrors to get the beam to both go through the field stop and hit
the intersection of the horizontal and offset lines on the CC
D50 front plate. (See
1.c.)

4.

Adjust the simulator z
-
position to get proper focus.

a.

Install the red fiber source.

b.

Check that the red light also goes through the field stop opening.

c.

Close down the stop and place a piece of card behind the opening.

d.

Without tilt
ing or rotating the simulator, push forward and backwards to get the
fiber source in focus on the
card
.

5.

Final adjustment.

a.

Reiterate through steps 3. and 4. to get the axis in the correct position and
direction, and the focus to be in the correct location.


4.3.2

Alignment of collimating and fold mirrors

The next part of the optical alignment involves installing the two mirrors in the HOWFS. The
focal length on the collimating mirror will determine the overall size of the reimaged pupil on the
lenslet
arrays and
the angle of incidence on the mirror will determine the amount of induced
astigmatism which will compress the pupil in the vertical d
irection. For an angle of incidence of
10.5° on the high
-
order deforma
ble mirror, the ratio of major to minor axes in the p
upil will need
to be 1.017
0.

The pupil will ultimately need to be 9.
5925

mm in width, and
9.4320

in
height to
achieve
the correct ratio

and size
. In addition, the on
-
axis laser beam reflecting off of the fold
mirror will need to be centered on the CCD50 an
d be parallel to the axis of the UTS50 stage.


1.

Installation of the collimating mirror.





14

a.

Attach

the ½”

tip/tilt mount to

the custom tilted fixed post
on the small optics
platform in the middle of the HOWFS. Make sure the

mirror will be pointing
upwards at
~7
.5
°

and such that the returning folded beam will clear the top of the
mount.
Install the mirror in the mount.

b.

Remove the fiber source and turn on the green on
-
axis laser.

c.

Place the mirror in the beam in approximately the correct location.

d.

The beam should
be slightly high of center on the mirror.

e.

Add shims to the underside of the mount until the beam is centered on the mirror.
Suggested shims are
0.0125”,
0.015” or 0.020”.

2.

Installation of the fold mirror.

a.

Make sure there is a lens cap over the CCD50. While
the green laser will not
harm the CCD50, it is best to avoid blasting it anyway.

b.

Place the fold mirror in its mount and attach it to the small optics platform.

c.

Slide the mount left and right until it is centered on the green laser and lock down
the mount.

d.

Use an inside micrometer to check the distance between the two mirrors is
54.512

mm or
2.146
”. If not, use all three fold mirror actuators to piston the mirror to the
correct position.

e.

Adjust the tip/tilt of the fold mirror to place the green laser in the
center of the
lens cap.

3.

Rough alignment of mirrors.

a.

Switch to the fiber source.

b.

Check the collimation of the beam going to the CCD50 in the left
-
right direction.
This can be done with a shear plate

when viewed from the side
-

not the top
-

of
the wavefront

sensor.

c.

If the collimation needs to be changed, unclamp the collimating mirror mount and
move it forwards and backwards until the beam is close (but not necessarily
exactly) collimated.

d.

Check the footprint of the beam on the collimating mirror with a piec
e of
transparency and make sure it is centered. If not, recenter and repeat the
collimation check and adjustment.

4.

Co
-
alignment with UTS50 axis

a.

Setup the
LabVIEW

GUI so that it is reading frames off of the CCD50 and set it
to the
‘Frame COM’

mode. This will

measure the cen
ter
-
of
-
mass of the entire
frame, so take
new backgrounds of

~5000 frames when appropriate.

b.

Load up the
ESP

utility. Setup up the focus stage axis so that you can move the
stage from one end of its travel to the other with a single click.

c.

Pl
ace an ND (4
-
6) either right in front of the CCD50 aperture or near the laser
aperture. [Still need to decide what’s best because no matter what you do the
beam walks or tilts with the addition on the NDs]

d.

Move the focus stage all the way forward.

e.

Tip and
tilt the fold mirror until the beam is centered on the CCD50.
These
actuators require 2 mm Allen keys for adjustment.
Use the CoM
-
X and CoM
-
Y
measurements. Get the absolute value of these numbers to be less than 0.1000
pixels.

See figure 7.






15


Figure 7
Ima
ges of the on
-
axis laser when aligning the laser beam to the UTS50 axis and CCD50 center.


f.

Next, move the UTS50 stage all the way (or ~50 mm) back. Notice the direction
and magnitude that the spot moves.

g.

Return the stage to the forward position.

h.

Using the
T/T adjustments on the collimating mirror, move the spot in the same
direction that was observed in step f
.
, with a magnitude of ~5
-
10

times greater.


i.

Use the T/T adjustment of the fold mirror to
recenter the spot on the camera to
within 0.1 pixels again.

j.

Repeat steps f. through i. until the spot position in less than 0.1 pixels from zero
(the center of the CCD50) in both extreme positions of the UTS50.
[
This is a
relatively fast process compared to
other tip/tilt convergence methods.]

5.

Install the
cylindrical lens

a.

Install the cy
li
ndrical lens in the 1” lens mount making sure the proper orientation
of the flat (towards the spatial filter) and concave surfaces (towards the CCD50.)

b.

Check the mark on the cylindrical lens and rotate it such that there is

no power in
the sideways direction and negative power in the vertical direction.

c.

Install the lens into the optical path.

d.

Using retro
-
reflections from the on
-
axis laser, center the cylindrical lens.

e.

Decenter the lens vertically by using the top actuator on

the len
s mount. Turn the
actuator by 3.65

turns

clockwise

whic
h corresponds to a distance of 0
.9
27

mm.

f.

Go back and repeat all of step 4, realigning the on
-
axis laser

with the CCD50, this
time getting the spot position to less than 0.05 pixels from the CCD
50 center.

g.

Insert the fiber source and check the collimation with a shear plate in both the
horizontal and vertical directions.

4.3.3

Installation of relay optics

This next step involves the installation of the relay optics refer
red to as R1 (closest to the lens
let
array) and R2 (closest to the CCD50.)






16

1.

Installation of R2.

a.

With the on
-
axis laser on and no NDs installed, let the laser fall directly on the
CCD50.

b.

Drop in R2 to the approximate corr
ect position as seen in figure 5
. Note the
correct orientation of th
e lens, flat negative element towards the CCD50.

c.

Using the retro

reflections off of the lens surfaces, normalize and center the lens
on the on
-
axis laser.

d.

Switch to the fiber source.

e.

Image the beam footprint on the CCD50.

f.

Adjust the z
-
position of R2 until
the beam footprint is 105.05 pixels wide when
using the telescope simulator. (If using the testbed or bench relay, this should be
104.04

pixels). It may not necessarily be centered, this is ok.

In ‘Frame COM’
mode, turn the boxes ‘on’ and switch to the 104

‘box size’.

g.

Remove the fiber and go back to step 3. Repeat until the lens is both aligned with
the laser axis and the beam footprint is the correct size.

2.

Establish fiber axis.

a.

With the fiber installed, now check the centration of the beam footprint on the

CCD50.

b.

Similar to
step 4 of § 4.3.2.,
center the footprint as seen on the CCD50 by
adjusting the tip and tilt on the collimating and fold mirrors for both the near and
far positions of the UTS50. Try to get the centration to less than 0.03 pixels (or
best

effort.)

3.

PixeLINK

alignment.

a.

Remove the CCD50 and put the
PixeLINK

camera in its place. There is a small
jig to hold the
PixeLINK

with a translation stage.

b.

Install the fiber source.

c.

Load the
PixeLINK

software and image the beam
footprint. Measure the ver
tical
height of the beam footprint by first saving an image to a .tiff and opening the file
with Microsoft Paint. MSPaint displays pointer coordinates, so measure the
maximum y extent of the beam.

d.

Adjust the z position of t
he
PixeLINK

to get
measurement of

the beam at
720
.3

(vertical) and 708
.2

(horizontal)

pixels
. I
f us
ing the simulator source, 713
.4

(vertical) and 701
.4

(horizontal)
pixels if using the testbed/bench relay optics.

e.

Once the beam size is correct, note the positions of the vertical and horizo
ntal
extrema of the footprint and calculate its center position.

4.

Installation of R1





The idea in this step is to use the laser to get the approximate tilt correct of R1

(since it
now does not exactly define the axis)
, while using the footprint size and location on the
PixeLINK

camera to establish the x
-
,

y
-

and z
-
positions of R1.




a.

Similar to the installation of R2, put the lens in the approximate corr
ect position
as seen in figure 5
, taking note to get the

proper or
ientation of the lens, flat
negative element towards the lenslet arrays.

b.

Using the on
-
axis laser, align the lens to the laser.

c.

Look at the vertical size of the beam footprint on the
PixeLINK
.





17

d.

Adjust the z position of R2 until the vertical footprint size is

885.6 (vertical) and
871.0 (horizontal) pixels for the simulator, 877.2 (vertical) and 862.8 (horizontal)

pixels for the testbed/bench relay.

e.

Once the proper footprint size is established, decenter the lens until the footprint
is centered in the same loca
tion on the
PixeLINK

as was established in step 2.e.

f.

Zero the tip and tilt of the lens again with respect to the on
-
axis laser and repeat
step e. The centration of the lens with respect to the laser may not be established.

g.

Repeat steps b. to f
. until the f
ootprint is the proper size and the lens is centered
on the laser.


4.3.4

Installation of
microlens

arrays and exchanger


The
microlens

arrays are already all installed in the exchanger and in the proper positions. I
f
necessary, instructions on mounting arrays
will be documented

in a future release.

The linear
stages of the HOWFS can all be driven in Windows using the ESP
-
utility. Figure 8 shows the
features of this program that will be used most frequently.



Figure 8
The ESP utility for driving the linear sta
ges in Windows. The Jog window should be set to the ‘Indexed’
mode. Select the axis on the left, set distance in the top and click a button to move either positive or negative in that
axis. The Home window can be used to move either axis one to the ‘Positi
ve limit and index’ position or axes 2 and
3 to the previously established ‘Zero’ position.





18

1.

Measurement of

microlens positions.

a.

If not already done, move the exchanger such that the simulator light goes
unobstructed through the central hole. By eye, center

the beam on the hole.

b.

In the ESP Position window, hit the Zero button on channels 2 and 3. It should
never be necessary to zero these channels again, so cover up that part of this
window.

c.

In the ESP Home window, set Axis 1
to ‘Positive limit and index’
. T
his will move
the UTS50 stage all the way forward.

d.

Start by finding the position of the first microlens array, go to
-
10 mm in both
channel 2 and 3.

e.

There will be light coming through part of one of the microlens arrays (s8).

f.

At this point, move the
PixeLI
NK

camera in z so that the camera slightly
overfocuses on the microlens array surface and the microlens boundaries appear
bright. See figure 9.


g.

Adjust channels 2 and 3 until the central microlens of the array are centered on
the illumination pattern. Note

that there are more microlenses than the actual
sampling: s8 has 10 across, s16 has 20 across, s32 has 40 across, and s64 has 80
across.

Use the patterns made by the grid of lines to center the arrays. Align the
arrays to the nearest 10 µm, which correspo
nds to approximately 1 pixel
resolution on the
PixeLINK

camera.

h.

Once the array has been centered, write down the values for channels 2 and 3.

i.

Return to the center by using the Home window’s Axis 2 and 3 ‘Zero position”
command.

j.

Find the center positions fo
r

the other three lenslet arrays and write them down.



Figure 9
Slightly

over
-
focused

images of the s8 (left) and s64 (right) microlens arrays on the
PixeLINK

camera.


2.

(Optional) Re
-
establish the proper focus positions of the microlens arrays.

a.

Go back to

the zero position on channel 2 and 3.





19

b.

Move

the
PixeLINK

in z such that the footprint corresponds to that seen in
§

4.3.3
step 4.d.

c.

For each of the different microlens configurations, change channel 1 until the
spots on the
PixeLINK

are at ‘best focus.’ S
tartin
g values are: s8 (
-
40.0), s16 (
-
18.0), s32 (
-
11.9) and s64 (
-
7.9)
.

4.4

Optical testing and final alignment


This is the final stage of the alignment process, putting the CCD50 back in the system, putting
the final tweaks on the R1 position and the rotati
on of the lenslet arrays.


1.

Install the CCD50 camera.

a.

Remove the
PixeLINK

camera.

b.

Make sure that the small Kapton spacers are near the three camera mounting bolt
-
holes. If not, tape 2 small squares of Kapton tape over each hole and use a hobby
knife to cut
out clearance for the ¼
-
20 bolts.

c.

Put the CCD50 back in its original position

and make sure the bolts are in
reasonably tight
.
(It might be worth putting a torque specification here

for
repeatability
.)
Note that the camera is held down on a three point mo
unt with
three bolts.


2.

Make initial measurements of the alignment errors.

a.

For each of the different
microlens array configurations, input

the premeasured
values for channels 1, 2 and 3 on the ESP controller
.

b.

For each configuration, select the appropriate r
econstructor from the LabVIEW
GUI.

c.

Set the fiber source such that the peak pixel value is around 4000. This should
avoid exceeding the putative full
-
well depth and avoid any intensity non
-
linearities
.

d.

Block or turn the fiber off and take a background, befo
re turning the source back
on.

e.

One of the first things to notice is how the tip and tilt measurements are large.
This should be due to a slight misalignment of the mircolens array, since it was
only possible to get it registered in section 4.3.4 to ~10 µm.

f.

Adjust channels 2 and 3 of the ESP controller by steps of 1 µm to get the tip and
tilt modes to b
e

roughly

zero.

Moving positive in channel 2 will cause tip to go in
the positive direction; moving positive in channel 3 will cause tilt to go in the
negativ
e direction.

g.

Once the system has settled, hit the ‘Avg Frame’ button and take an average of a
few thousand images. At the end of the average, the GUI will update, displaying
the alignment metrics of the averaged frame.
Hit the ‘Frame Updates Off’ to go
bac
k to a live image.

h.

At this point it is useful to make a chart of the summarized alignment metrics and
delta positions in channels 2 and 3. See table 1.

Fill in the table for each
configuration.

i.

Ideally, the magnitude of the Δ2 and Δ3 columns should be less

than 10 µm,
although up to 20 µm may be acceptable. If they are not, this may be indicative of
a misalignment somewhere in the previous steps.





20

Table
1

Spreadsheet for recording alignment positions and metrics.


Ch.1

Ch. 2

Ch. 3

Δ2

Δ3

Foc.

A45

A90

Sph.

RMS

WFE
-
P

WFE
-
Z

Rot.

s

mm

mm

mm

µm

µm

λ

λ

λ

λ

pix.

nm

nm

deg

8














16














32














64















3.

Microlens array rotation adjustment

a.

Go back to each microlens configuration
from

step 2
, including the Δ

positions
.

b.

Confirm the rotation measurement in degrees.

c.

See figure 10. The microlens arrays are mounted to cylinders that can rotate.

The
cylinders have a set of two small holes drilled ~1 mm behind the array mounting
surface in which a 0.05” hex wrench (
or other appropriate instrument) can be
inserted to rotate the array.
There are two set screws that hold in the
cylinders to a
larger rectangular block which mounts to the two linear stages.



Figure
1
0
Image of the rotation adjus
tment for the microlens arrays in their exchanger mechanism.


d.

If the rotation is far off, then unlock the set screws and get the rotation reasonable
by eye and lock the set screws. They should be reasonably snug, but not over
tightened.

e.

Place the 0.050” h
ex wrench in the adjustment holes and give a quick firm tap of
the wrench. This should rotate the array by a measurable amount. It may take
some time to master how hard to tap the wrench to get the array to rotate.

f.

Adjust the rotation of the array, using f
eedback from the LabVIEW GUI, until the
magnitude of rotation is less than 1E
-
3 degrees, it is difficult to get the rotation




21

much less that 5E
-
4. (The large visible dial that shows rotation in the GUI is used
for final touches on the rotation and displays
units of 1E
-
2 degrees.)

g.

Once all of the arrays are rotated, null out the tilt modes like in step 1.f.
, take an
average frame and record the new alignment positions and metrics.

4.

Focus adjustments

Typically there will be a discrepancy in the measured focus m
ode between the
different configurations.
This is due primarily to a magnification error in the relay
lens pair.
Proceed with this adjustment if this is the case.

a.

Switch to the s8 configuration, making sure to zero out any additional measured
tip and tilt
by moving the array position (shouldn’t be more than 2 or 3 µm.)

b.

Make note of the focus position with another average frame measurement.

c.

Go back to a live image, possibly turning up the leak coefficient to 0.5 to keep the
measurement from moving too much.

d.

Using the three 100
-
TPI actuators on the back of the R1 lens mount, piston

the
lens forwards or backwards until the measured focus mode goes to 0 waves.

As a
reference, moving each actuator clockwise by approximately half a turn will cause
the focus measur
ement in s8 to go from +0.2 waves to 0 waves.

e.

Repeat a measurement of the alignment positions and metrics for all of the modes.

f.

The focus measurement for each of the modes should all be closer together and all
closer to zero.

g.

Repeat the piston adjustment
of R1 until all of the modes are close to 0 waves of
focus.

(See appendix 6.1 for the best focus in each configuration.)

h.

It is possible that the beam isn’t perfectly collimated, in which case it is more
important to get all of the different configurations
measuring the same focus
coefficient. Once they all read basically the same value, move the fiber source
forward in z to null out the focus. Make sure to adjust the fiber x and y position to
not add any tip or tilt as the z
-
stage that the fiber is on may n
ot be parallel with
the optical axis of the telescope simulator.

5

APPENDIX

5.1

Discussion of additional errors

5.1.1

Spherical (and other high order radially symmetric modes)


If there is any residual spherical aberration measured, this could be due to distortion in

the relay lens pair. From the relay design, this is primarily caused by an error in distance between
R2 and the CCD50. This can be adjusted by pistoning R2 in a similar manner to R1. Warning:
since the beam is only ~1.6 mm wide on R2, it is very sensitive

to any mis
-
alignments. Adjusting
piston may result in slight tilting or decentering of R2 from which it is difficult to recover and
will likely set you back to section. The Zemax file for the relay is attached below
:







22

5.1.2

Coma


The largest

amount of uncorrectable internal optical error in the HOWFS
should be coma.
In Zernike coefficients, the coma should be 18.5 nm RMS or 52.4 nm P
-
V. The LabVIEW GUI
should therefore be displaying ~+0.03 waves at λ

= 635 nm, but is currently tending to hove
r
around the +0.08±0.01 waves, which is more consistent with the P
-
V coefficient. The Zemax
design that measures the wavefront error going up to the microlens array is included below for
reference in order to solve this mystery in the future:



5.1.3

Astigmatism

From the

Zemax design presented in 5.1.2, it can be seen that there should be 0 nm of 45
°

astig
matism and only a very small, 5

nm
RMS
, amount of 90
°

astigmat
ism. This should
correspond to less than 1/100 of a wave. Therefore the A45 and

A90 measurements should be
much smaller than measured in Appendix A.

5.2

Appendix A


October 18
th
, 2009 alignment


Ch.1

Ch. 2

Ch. 3

Δ2

Δ3

Foc.

A45

A90

Sph.

RMS

WFE
-
P

WFE
-
Z

Rot.

s

mm

mm

mm

µm

µm

λ

λ

λ

λ

pix.

nm

nm

deg

8

-
40.0

-
11.521

-
11.254

-
1

+6

-
0.01

+0.02

-
0.01

-
0.005

0.1464

187

156

-
4E
-
4

16

-
18.0

11.321

-
11.206

+1

+4

-
0.05

-
0.08

-
0.03

+0.03

0.0851

218

141

1E
-
4

32

-
11.9

11.156

11.613

-
14

+3

+0.11

+0.06

-
0.09

0.000

0.0433

128

116

8E
-
4

64

-
7.9

11.415

11.592

-
5

+2

+0.12

+0.06

-
0.08

0.000

0.0265

109

116

-
3E
-
4







23

5.3

A
ppendi
x B



HOWFS a
lignment table worksheet


Alignment Description/Changes:


Ch.1

Ch. 2

Ch. 3

Δ2
=
Δ3
=
c潣.
=
䄴^
=
䄹^
=
p灨p
=
o䵓
=
tcb
J
m
=
tcb
J
w
=
o潴o
=
s
=

=

=

=

=

=
λ
=
λ
=
λ
=
λ
=
灩p.
=

=

=
摥g
=
U
=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
䅬楧湭敮琠䑥sc物灴楯港n桡湧n猺
=
=
C栮h
=
C栮′
=
C栮″
=
Δ2
=
Δ3
=
c潣.
=
䄴^
=
䄹^
=
p灨p
=
o䵓
=
tcb
J
m
=
tcb
J
w
=
o潴o
=
s
=

=

=

=

=

=
λ
=
λ
=
λ
=
λ
=
灩p.
=

=

=
摥g
=
U
=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
䅬楧湭敮琠䑥sc物灴楯港n桡湧n猺
=
=
C栮h
=
C栮′
=
C栮″
=
Δ2
=
Δ3
=
c潣.
=
䄴^
=
䄹^
=
p灨p
=
o䵓
=
tcb
J
m
=
tcb
J
w
=
o潴o
=
s
=

=

=

=

=

=
λ
=
λ
=
λ
=
λ
=
灩p.
=

=

=
摥g
=
U
=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
䅬楧湭敮琠䑥sc物灴楯港n桡湧n猺
=
=
C栮h
=
C栮′
=
C栮″
=
Δ2
=
Δ3
=
c潣.
=
䄴^
=
䄹^
=
p灨p
=
o䵓
=
tcb
J
m
=
tcb
J
w
=
o潴o
=
s
=

=

=

=

=

=
λ
=
λ
=
λ
=
λ
=
灩p.
=

=

=
摥g
=
U
=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
䅬楧湭敮琠䑥sc物灴楯港n桡湧n猺
=
=
C栮h
=
C栮′
=
C栮″
=
Δ2
=
Δ3
=
c潣.
=
䄴^
=
䄹^
=
p灨p
=
o䵓
=
tcb
J
m
=
tcb
J
w
=
o潴o
=
s
=

=

=

=

=

=
λ
=
λ
=
λ
=
λ
=
灩p.
=

=

=
摥g
=
U
=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
䅬楧湭敮琠䑥sc物灴楯港n桡湧n猺
=
=
C栮h
=
C栮′
=
C栮″
=
Δ2
=
Δ3
=
c潣.
=
䄴^
=
䄹^
=
p灨p
=
o䵓
=
tcb
J
m
=
tcb
J
w
=
o潴o
=
s
=

=

=

=

=

=
λ
=
λ
=
λ
=
λ
=
灩p.
=

=

=
摥g
=
U
=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=

=
=
=
=
=
=
=
=
=
=
=
=
=
=
=