Transcript IPM_RWilliamson_2016_finalx

IPM Simulation and
Correction with Strong
Space Charge
R Williamson, B Pine, C Warsop, C Wilcox
With many contributions from ISIS Diagnostics and
Accelerator Physics Groups
IPM Simulation Kickoff
Workshop 2016, CERN
Contents
• Introduction to ISIS
• ISIS Residual Gas IPM
• IPM Simulations
• Measurement errors
• Comparison measurements to
simulation
• Profile correction
• Plans for the future
Introduction
Target Station 1 (40 pps)
H- ion source, 35 keV
H- RFQ, 665 keV
H- Linac, 70 MeV
Target Station 2 (10 pps)
800 MeV Synchrotron
Introduction
Circumference:
Energy:
Repetition Rate:
Intensity:
Mean Power:
Tune Shift:
Injection:
RF system:
Extraction:
Extraction
163 m
Injection
70–800 MeV
50 Hz
Up to 3x1013 ppp
Up to 190 kW
~0.5
130 turn
charge exchange
h=2, 1.3-3.1 MHz, 160 kV/turn
h=4, 2.6-6.2 MHz, 80 KV/turn
single turn, vertical
Residual Gas Ionisation Profile Monitor
- ISIS Design
• Compensating electrodes double as
calibration device – single channel
• Up to 60 kV drift field
• 300 π mm mrad, ~100 mm wide average
• 10-7 mbar in synchrotron
340 mm
Residual Gas Ionisation Profile Monitor
- Detector Development
• Single channel electron multipliers channeltrons®
• Upgrade from one moveable channeltron®
(averaged profile) to 40 fixed channeltrons®
(single pulse profile)
• 6 mm centre to centre separation
• Single, moveable channeltron® used to
calibrate multichannel
S Whitehead, S
Payne, A Pertica
et al.
BIW’ 10
3D Electrostatic Model
• Simplified geometry
• Compensating field
• Beam modelled as
concentric cylinders
– Centred
– “Normal” distributions
Measurement Error
– 2D Drift Field
• Non-uniform electric field
• Pushes ions outward from central axis
• Measured profile widened
• Feature of monitor geometry
y
x
Ideal ion path
Actual ion path
Measurement Error
– Space Charge
• Pushes ions out from
centre of beam
• Measured profile
widened
• Dependent on beam
charge, size, energy,
distribution, intensity
and drift field voltage
B. Pine
EPAC ‘06
1x1013 ppp
Space Charge
Field
+ High Voltage Electrode
Ion Tracking
– Full 3D
• Simple in house code
– Load electric field
– Create ions distributed evenly
throughout beam volume
– Track ions with zero initial velocity until
enter the detector or exit the monitor
– Weight tracks with beam distribution
Profile Simulator Workflow – Setup & Ion Tracking
Inputs
(Data from ISIS
machine run)
3D E-Field
Calculation
(CST EMS)
Beam profile information from 2D SEM grid
Horizontal and vertical radii, 2D beam position.
Beam
Intensity
IPM Drift Field
Strength
CST EM Studio IPM models (x2) – Calculate E-fields within monitor volume
2 different E-fields are used for a time dependence approximation.
Model 1 calculates E-field with beam inside monitor, Model 2 calculates the E-field with no beam.
C++ tracking code setup
Ion Tracking
Create randomised ion distribution within beam volume.
Load the E-Field matrix from CST model 1.
(C++)
t=t+Δt
Loop for each time step (t) until all ions have left tracking region
i=i+1
Loop for each ion in the distribution (i)
Calculate nearest E-Field matrix point
No interpolation of E-Field used as point spacing <1% of detector width.
See appendix for
equations
Calculate ion motion during time step
Displacement and new velocity are calculated using initial position, initial
velocity and Lorentz force applied by E-Field.
If t=200 ns, replace E-field matrix
Load the E-field from CST model 2 and use for the remainder of the tracking time (ttotal~1500 ns)
Profile Simulator Workflow – Post Processing
Post
Processing
Code (IDL)
Load ion motion tracks calculated in C++ ion tracker code
Create simulated profile measurement from each ion that has a
final position on the detector face
Measured profile consists of a histogram with 40 bins, each with a width of
6mm. Each bin represents one channeltron detector.
Beam
distribution
weighting
Weighting to
simulate
detector
behaviour
Apply an elliptical weighting to the measured profile, using the
initial (ionisation) positions of each ion
The ion tracker created a uniform distribution of ions within the beam volume,
but in reality the ion creation will more concentrated in the beam centre.
Apply a weighting to simulate the behaviour of the channeltron detection behaviour
The channeltron detection efficiency decreases as the angle of incidence increases. It also varies with
the energy of the incoming ion.
Calculate & output results
Ion Tracking Results
• Combined errors can double the
measured profile width in normal
running conditions
• Long tails present on profiles
Ideal profile
Simulated profile
Measurement Error
– Longitudinal Drift Field
• Non-uniform electric field
• Longitudinal electric potential saddle point
• More complicated ion trajectories
• Not a pure 1D projection
Ideal ion path
Actual ion path
y
z
Initial Position of Detected Ions
• Distorted 2D
slice
• More ions from
monitor centre
• Some oscillatory
trajectories
Projection of
detector aperture
Edge of
electrode
Benchmark – Transfer Line
• Check understanding of profile monitor
in EPB with other diagnostics
Profile
– Position monitor (planned) Grid
– Ionisation profile monitor
– 3 x SEM grid profile monitors Position
SEM Grid
Ionisation Profile Monitor
Grid
Grid
SEM Grid Vs Ionisation Profile
• Measurement at 2.5x1013 ppp
• Nominal 15 kV drift field
• 95% width:
SEM Grid – 100 mm
IPM
– 158 mm
Accuracy +/- 3 mm
Simulation Vs Measurement
• Model beam in ionisation monitor as measured
using SEM grid profile monitor
• Simulate resulting profile
• Compare with measured ionisation profile
• Many profile comparisons in
the synchrotron confirm
model with space charge and
drift field
• EPB comparisons promising
Simulated profile (black)
Measured profile (blue)
Profile Correction
Drift field correction
Space charge correction
Beam intensity
ksc pg I
wm, p = kd wt, p +
V
Relativistic gamma
Drift field voltage
R. Williamson
EPAC ‘08
Ionisation monitor measured width and true beam width at percentage p
Simulated
data
• Derived from
combination of
simulation and
measurements
• Correct synchrotron
profiles to within 6 mm
• Profile tails
Ideal profile (purple)
Simulated profile (red)
Corrected profile (blue)
Error Summary
In decreasing order of importance
• 2D non-uniform drift field
–
–
–
–
Verification at low intensities
~30% error on width
Simple correction
Gradient electrodes
• Space charge
– Verification with altering drift field
– Error dependent on beam parameters, drift
field, ~30% error on width
– Simple correction for centrally peaked,
centred distributions
– Increase drift field voltage
• 3D non-uniform drift field
– Needs to be experimentally verified
– Small effect
– Remove compensating field?
Summary
• Good model of the ISIS residual gas
ionisation profile monitor
• Measurement errors
– 2D non-uniform drift field
– Space charge
– 3D non-uniform drift field
• Simulations Vs Measurement
• Correction works well for “normal”
centred beams
• Key measurements for understanding
machines with high space charge
Plans for the Future
• Effectiveness of profile correction for
different beam types
– Intensity
– Off-centre
– Unusual beam distributions
• Experimental check of beam width by
different methods in the synchrotron
• Characterisation of the channeltrons® in
new diagnostic vacuum vessel
• Development of the correction model
and the monitor itself
Additional slides
HEDS IPM
•
•
•
•
•
Newer IPM installed in HEDS
between the Linac and ring
in 2010.
Beam profile here is narrow
(~19mm width) so
channeltrons not suitable.
32 channel micro channel
plate (MCP) used to detect
ions instead.
MCP can be rotated 90o to
allow for calibration of each
channel.
Shaping field electrodes
added to the design to
remove widening effect of
the drift field.
Stepper motor
and resolver
32 channel
MCP detector
Beam
Shaping field
electrodes
P. Barnes
IPAC ‘11
Profile Simulator Workflow - Setup
Inputs
(Data from ISIS
machine run)
3D E-Field
Calculation
Beam profile information from 2D SEM grid
Horizontal & vertical radii, positions.
Beam
intensity
IPM Drift Field
Strength
CST EM Studio IPM Models – Calculate E-Fields within monitor
2 CST IPM models are used for time dependence approximation:
1 model with beam inside monitor, 1 model without beam
(CST EMS)
Export E-Fields as 3D grids with data point spacing of 2mm in
each plane
Data point spacing <1% of IPM detector width (240mm).
Ion Tracking
Setup
C++ Tracking Code Setup
Import E-Field values from the CST model with beam inside the monitor.
(C++)
User Inputs
‐ Beam radii & position
‐ Residual gas ion mass
& charge
‐ Tracking region
boundaries (detector
& monitor wall
positions).
Create uniform ion distribution within beam
volume then randomise initial positions
Initial ion velocities are set to 0.
Begin tracking loop
Profile Simulator Workflow – Ion Tracker
Ion Tracking
Loop
t=t+Δt
Loop for each time step (t) until all ions have left tracking region
Time step size, Δt = 1 ns
(C++)
i=i+1
Loop for each ion in the distribution (i)
Calculate nearest E-Field matrix point and retrieve field values
No interpolation of E-Field used as point spacing <1% of detector width.
Calculate ion motion during time step
New position and velocity calculated using initial position, initial
velocity and Lorentz force applied by E-Field.
*The user
defines the
“tracking
region” during
setup. The
tracking region
represents the
volume inside
the monitor,
with
boundaries on
the ion
detectors and
the monitor
walls.
Has ion
left tracking
region*?
No
No
Calculate final time step size, Δtf
Yes
If the ion reached the detector, calculate
exact time of arrival (<1ns). Use this new
time step to recalculate exact position &
velocity at contact with detector.
Have all ions been
tracked in current
time step?
Exclude ion from further
tracking loops
Yes
If t=200 ns, load the beam independent E-Field matrix
Replace the initial E-Field matrix, calculated with the beam inside the monitor,
with the E-Field from the CST model with no beam present.
Profile Simulator Workflow – Post Processing
Simulation
Post
Processing
Code
(IDL)
Load ion motion tracks calculated in C++ ion tracker code
Create simulated profile measurement from each ion that has a
final position on the detector face
Measured profile consists of a histogram with 40 bins, each with a width of
6mm. Each bin represents one channeltron detector.
Beam
distribution
weighting
Weighting to
simulate
detector
behaviour
Apply an elliptical weighting to the measured profile, using the
initial (ionisation) positions of each ion
The ion tracker created a uniform distribution of ions within the beam volume,
but in reality the ion creation will more concentrated in the beam centre.
Apply a weighting to simulate the behaviour of the channeltron detection behaviour
The channeltron detection efficiency decreases as the angle of incidence increases. It also varies with
the energy of the incoming ion.
Calculate & output results
Injection Painting Measurements
• Closed orbit and Relative Betatron
amplitude measured over injection
• Results used in ORBIT simulation of
painting
• Beam distributions generated by ORBIT
compared to MCPM measurements at
high and low intensity
• Space charge correction
applied to HI data
B. Jones
EPAC ‘08
Low intensity
High intensity
Drift corr.
Drift and SC corr.
2.5x1012 ppp
2.5x1013 ppp
-0.3ms
-0.3ms
-0.2ms
-0.2ms
-0.1ms
-0.1ms
Half Integer Studies
Half integer resonance with space charge
• Key loss mechanism
• Experimental studies 2D coasting beam
𝜀𝑥= 𝜀𝑦 , 𝜀𝑟𝑚𝑠 ≈20 π mm mr, 2Qy=7 driving term, Qy=3.6
Ramp intensity (1E13 ppp), push onto resonance
• Study evolution of corrected profiles, not just loss
Observations agree with ORBIT models
Clear formation of core and lobes
ORBIT
Measured
• Using to understand loss
mechanism
1D and 2D models of core-tohalo dynamics
• Good model of profile
monitor key
C. Warsop
Space Charge ‘15
Driving
Phase 1
Driving
Phase 2