Injector Commissioning Problems

1
Injector Commissioning Problems

• Analysis of Work Package 1

2
Outline

• Image processing
• GTF transverse data
• GTF longitudinal data
• Tomography
• LCLS simulation data
• Summary

3
Image Processing

• Background removal
• Region of interest
• Calculate moments and profiles

4
Background Sources

• From accelerator
• Dark current
• X-rays
• Non-uniform ambient light
• From camera
• Damaged pixels
• Gaussian noise distribution
• From transmission line
• Line noise

5
Background Removal

• Dark current and ambient light Subtract
  background image.
• X-Rays damages pixels Median filter.
• Gaussian noise Set zero in histogram.
• Line noise Digital signal transmission.
• Background image subtraction
• Removes static features in background.
• Increases random noise and varying features like
  line noise.
• Average many background images to get static
  background.

6
Region of interest

• Find coarsely beam centroid and size from entire
  image.
• Determine ROI as multiple of beam size.
• Best multiple depends strongly on beam shape.

7
GTF Image Processing

• Processing w/ and w/o background subtraction
• OTR
• Horizontal lines do not vanish with median
  filter.
• Lines different in background image.
• YAG
• Gaussian noise additional 1/f like damaged
  pixels
• Line noise
• Line noise enhances by background subtraction
• 1/f like noise removed by subtraction

8
Sample Image Processing
w/o BG Subtr.
w/ BG Subtr.
Histogram
BG
Beam
1.5s Cut
Raw image
Cleaned image
3.6s ROI
9
GTF Quad Scan Analysis

• Calculate mean and std of measured set of beam
  sizes.
• Generate 1st order beam transport matrices.
• Least square fit of s-matrix to squared beam
  sizes.

OTR
YAG
10
GTF Transverse Results
11
GTF Slice Emittance
Get time axes from energy-time correlation in
longitudinal phase space
Slice
BG No BG
Projected
Mismatch parameter ? (ß0?-2a0a?0ß)/2
12
GTF Horizontal Phase Space
Simultaneous Algebraic Reconstruction Technique
Maximum ENTropy
SART, enx 0.51 µm (5 cut)
MENT, enx 0.55 µm
Profiles from YAG screen measurement Linear
momentum correlation removed
13
Tomography Projections (SART)

• Reconstruction error
• E v(?(p0 p)²) / ?p0
• SART 1.2
• MENT 1.8

14
GTF Longitudinal Data

• Fit second order polynomial to beam energy.
• Determine crest phase and energy gain in
  accelerator (assume 5 MeV gun energy).
• Calculate linear longitudinal transport
  matrix.E E ?E kRF sin(f0) z
• Fit longitudinal beam ellipse to energy spread
  data.

15
GTF Longitudinal Emittance
16
GTF Longitudinal Phase Space

• Include 2nd order beam transport due to rf
  curvature.
• E E ?E kRF sin(f0) z ½ ?E kRF² cos(f0) z²

SART, enz 3.3 µm (5 cut)
MENT, enz 3.1 µm
5 cut
no cut
17
Tomography Projections (MENT)

• Reconstruction error
• SART 2.3
• MENT 2.5

18
Time Profile and Energy Correlation
Projected energy spread 8.4 keV Average slice
energy spread 5.8 keV Minimum slice energy
spread 3.0 keV
SART
MENT
19
LCLS Simulation Data

• Images with gaussian background noise only
• Bounding box 3.6s of beam size
• Histogram cut at 1.5s of noise width
• No error on beam size from single image
• Linear fit of sigma matrix to squared beam size

20
Results of Simulated Data Analysis
21
Slice Emittance
Slice
Projected
ee1_K ee2_K
Mismatch parameter ? (ß0?-2a0a?0ß)/2
22
Summary

• Different methods of image processing available.
• Best method depends on statistical and
  systematic noise sources in images.
• Beam sizes and emittances depend on method used.
• Design image acquisition with minimal noise
  sources for method independent beam moments.