Tuxedo

1
Laboratory Demonstration of In-chamber Target
Engagement
Mark Tillack, Lane Carlson, Jon Spalding
Dan Goodin, Graham Flint, Ron Petzoldt, Neil
Alexander
HAPL Project Meeting Rochester, NY 8-9 November
2005
2
We are attempting to demonstrate the
pessimistic version in-situ target engagement
system proposed by Flint 3/05 (Gen II)
Goals ? Full integration of all key elements of
target engagement ? Benchtop demo first
identify and solve problems before investment in
full-scale, high-performance demonstration
Key Requirements 20 mm accuracy in
(x,y,z) 1 ms response time
3
The system consists of Poisson spot detection,
Doppler fringe counting, a simulated driver with
steering, and a glint-based alignment
4
Initial Poisson spot results were reported at the
previous HAPL meeting
L. Carlson, M. Tillack, D. Goodin, G. Flint,
RD Plan for Demonstrating Elements of a Target
Engagement System
We demonstrated Poisson spot detection with 5 µm
accuracy in lt1 ms using a translation stage and
an ex-situ centroiding algorithm
5
To perform real-time target engagement on the
benchtop, we needed a target transport method
We are using various translation stages and rail
systems
Were still working on a more prototypical
surrogate transport method
6
Our in-line benchtop centroiding system now runs
continuously at lt20 ms per measurement
1 cm/s target speed over 1 m travel 100 fps
Basler camera Labview running on Windows XP
this allows us to begin real-time feedback to
beam steering higher speed will require
real-time OS and a faster camera
7
Integration of Poisson spot detection with a
fast steering mirror was implemented
We passed a pseudo driver beam through a 10x beam
expander to magnify the range of motion of FSM
(1.5 mm)
Determining the location of the driver on the
target is difficult the accuracy of engagement
is confirmed with an offset PSD as a surrogate
target
8
Open loop Poisson spot tracking The Movie
1. At t0, PSD initialized at (0,0) 2. Start
train moving 3. Measure Poisson spot
(x,y) 4. Move FSM to follow sphere 5. Measure
accuracy using PSD
yellow dot PSD
white dot Poisson spot
3 mm CMOS 1.5 mm PSD
9
Engagement is performed in 23.5 ms, but dynamic
errors are too large
x-axis comparison of PS and PSD readings
Breakdown of times
Sources of errors rocking of PSD
target speed limitations in PC hardware/
software overly simplistic gain curves
FSM quality
10
We characterized the Thorlabs piezo cage mirror
mount using a signal generator
Beam deflection is nonlinear with drive voltage
and exhibits severe resonant behavior
1 ms
595 Hz
617 Hz
higher performance will require a better FSM
11
Work has begun on Doppler fringe counting

• Restrictions on laser power limit the use of a
  metal sphere, so were using an n2 sphere and
  flat mirror
• Single-wavelength (632.8 nm)
• Errors due to translation stage, vibration, air
  flow

Repeatability demo using micrometer travel of 5
mm with 10 mm increments
An N2 ball lens is a retroreflector
12
Curvature in the target trajectory allows us to
avoid a shutter mirror for a range of velocities
We performed a fast tracking demo at 1000 Hz
using a high-speed pellet and post-shot centroid
analysis
Speed of gun is too fast, speed of tracking too
slow ? work on the benchtop
1000 fps, 10 ms per frame video sequence of
surrogate target coming into, then out of the
cameras FOV, at 150 m/s (Photron camera)
13
Next Steps more integration and more
prototypical
Poisson system Acquire a faster camera and
real-time OS Doppler system Demonstrate
counting on metal spheres with longer
paths Implement dual-wavelength
countingIntegration of Doppler and
Poisson On-axis demonstration
(pseudo-integration) Off-axis demonstration
(true integration)Integration of Poisson and
FSM Improve control of the environment, acquire
a high-end FSMGlint system Install glint
laser and coincidence sensor, align 2 beamlets