Jefferson Lab Experience with
Beam Halo, Beam Loss and
Related beam diagnostics
Pavel Evtushenko, Steve Benson, Dave Douglas,
Kevin Jordan, Carlos Hernandez-Garcia, Dan Sexton,
Jay Benesch, Arne Freyberger
UBW2012, Berlin, December 2012
Outline
1. IR/UV Upgrade (JLab FEL) overview
2. Different sources of unwanted beam
3. Beam dynamics example
4. Drive Laser related remarks
5. Setting up for high current operation
6. Dynamic range of diagnostics
7. Recent LDR measurments
8. CEBAF overview
9. RF trip rate
10. Direct measurements
11. Vacuum “events”
12. Summary
UBW2012, Berlin, December 2012
JLab IR/UV Upgrade
Ebeam 135 MeV
Bunch charge: 60 pC – UV FEL
135 pC – IR FEL
Rep. rate up to 74.85 MHz
25 μJ/pulse in 250–700 nm UV-VIS
120 μJ/pulse in 1-10 μm IR
UBW2012, Berlin, December 2012
Flavors of Unwanted Beam
Four sorts of the unwanted beams
1. Fraction of the phase space distribution that is far away from the
core (due to the beam dynamics)
2. Low charge due to not well attenuated Cathode Laser (ERLs) –
but real bunches that have proper timing for acceleration
3. Due to the Cathode and Laser but not properly timed (scattered
and reflected light on the cathode and in the DL transport)
4. Field emission: Gun (can be DC or RF), LINAC itself
(is accelerated in both directions)
5. Actually, there is one more – ions that accumulate in are true
CW electron beam, travel in both directions with thermal
velocities in side the electron beam, reduce Q.E. of the cathode
one really does not want this beam.
UBW2012, Berlin, December 2012
FEL Injector as an example of #1
Measured in JLab FEL injector,
local intensity difference of the
core and “halo” is about 300.
(500 would measure as well)
10-bit frame grabber & a CCD
with 57 dB dynamic range
PARMELA simulations of the same setup with 3E5 particles:
X and Y phase spaces, beam profile and its projection show
the halo around the core of about 3E-3.
Even in idealized system (simulation) non-linear beam
dynamics can lead to formation of halo.
UBW2012, Berlin, December 2012
Drive Laser “ghost” pulses
 Using a Log-amp is an easy way to diagnose presence of the “ghost” pulses
 Log-amps with dynamic range 100 dB are available
631 uA (100%)
135 pC x 4.678 MHz
5.7 uA (~0.9 %)
4.678 MHz “ghost”
pulses
UBW2012, Berlin, December 2012
DL light scattered on photo cathode
 a view of GaAs photo cathode when
running beam (probably 6 % duty
cycle or 1.5 %)
Wafer 25 mm diameter
Active area 16 mm diameter
 measured with simple vis. CCD
camera
Drive laser 8 mm diameter
 locations of the wafer and active
area are knows from the same view,
HV off and white light on
 we are looking in to a gap between
two non-flat mirrors
 with a brand new wafer (no heat
cleaning) one would not see any light
from the DL spot
 At least two processes contribute to the
generation of scattering centers
 Heat cleaning of the cathode (made periodically,
every 4-5 re-Cs)
 Visible (green) DL preferable over UV
 Preserving cathode surface will be very helpful
 Get rid of heat cleaning for GaAs (H – cleaning)
 HV breakdowns can result in rather large pits –
scattering and field emission
UBW2012, Berlin, December 2012
Courtesy of C. Hernandez-Garcia
Cathode Laser pulse via streak camera
 appears to be close to Gaussian on linear scale; tails not so much Gaussian
 the difference from Gaussian distribution is obvious on log scale
 realistic (measured) distribution must be used for realistic modeling
 especially is the calculations are intended for large dynamic range effects
UBW2012, Berlin, December 2012
High current operation
 JLab FEL driver is setup for high current operation in three steps/phases
 Most of the measurements are made with low duty cycle beam beam
(this is Step 1 that establishes best RMS setup for FEL performance)
- setting up injector (RF phases and solenoids)
- transverse match
- longitudinal match
 Step 2 is to increase the duty cycle, usually to 6 %, and look at
the beam loss, small adjustments in transverse and long. match often are
required; the adjustments must preserve the high performance of the FEL
this is the reason the adjustments have to be small
 When beam loss is small enough high average (9 mA) current can be
operated and the long term trends in pressure (vacuum) are used for
Step 3 of machine adjustment, also very small.
UBW2012, Berlin, December 2012
General FEL remarks
1. JLab FEL is a 9 mA average current machine, despite the fact that all four sorts of
beam halo are present (9 mA and 135 MeV  1.2 MW)
2. Setting up for high current operation requires time, but has been done routinely
3. To properly (and quickly) deal with first kind of beam halo Large Dynamic Range
diagnostics are needed; until then takes time and trail and error, and origins of
problems are not well understood (not confirmed by measurements)
4. For Drive Laser transport Brewster angle windows (input and output); essentially light
tight beam line; laser transport with spatial filter to mitigate scattering
5. Scattered DL light on the cathode is a reality one has to leave with, i.e., run beam
when it is small enough and replace cathode when it is not.
6. Gradient in the LINAC is limited via requirements to keep dose rate below certain level
(especially at the wiggler), but also due to other effects the same as at CEBAF (trip
rate)
7. Instruments are:
 Beam Loss Monitors (BLM) of the MPS
 Rad.Con. calibrated ionization chambers
 Radiation survey just after beam operation ended (for chronic losses)
 LDR diagnostics have been developed
UBW2012, Berlin, December 2012
Why large dynamic range diagnostics?

Operation of JLab FEL with high average current requires a compromise (in terms of
match) between high peak beam brightness (required by FEL) and very low beam loss

The match is iterative process and often does not converge easily (if at all…)

For the transverse beam profile measurements and transverse match JLab FEL relies
heavily on beam imaging (2D distribution) large number of beam viewers

LINAC beams have neither the time nor the mechanism to come to equilibrium
(unlike storage rings, which also run high current)

When setting up a high current accelerators with tune-up beam, halo is something
invisible (due to the dynamic range of the measurements) during the setup, yet causing a
lot of difficulties when trying to run high current

Increase the DR significantly to make the halo measurable visible with tune-up
beam already; measure the phase space distribution with the LDR and use such
information for the match. When DR is large enough no need to separate what is
core and what is halo.
UBW2012, Berlin, December 2012
Imaging Sensor(s) Dynamic Range

The first issue to overcome is the DR of a single imaging sensor

The main principle is to use imaging with 2 or 3 sensors with different effective
gain simultaneously and to combine data in one LDR image digitally
(single sensor dynamic range 500..1000 if cost is kept reasonable)

From experience (calculations tested by experiments) we know the safe level of beam
current/power for a low duty cycle (tune-up) beam

With typical beam size of few hundred μm OTR signal is attenuated by ~ 10 to keep CCD
from saturation. For phosphor or YAG:Ce viewers attenuation of at least 100 is used.

Using OTR there is enough intensity
to measure 4 upper decades;
lower two decades need gain of
about 100 to be measured.

The key elements:




image intensifiers
alignment and linearity
combining algorithm(s)
understanding CCD saturation
Intensity range that can
be measured without
additional gain
Intensity range where
additional gain of ~ 100
is needed.
(not high for an image
intensifier)
UBW2012, Berlin, December 2012
To be measured with
imaging sensor #1
and attenuation ~ 10
To be measured with
imaging sensor #2
and gain ~ 200
Raw images and combining algorithm
Data combining algorithm
 Two images (on the left) measured simultaneously with integration
times 20 us and 400 us
 Background measurements and subtraction is crucial!
Made separately for two sensors and subtracted on-line.
 Combining algorithm is efficient enough to provide 5 Hz rep. rate for
1024x768 images
 At the time of measurements was limited by the flexibility of DLPC
 Demonstrated dynamic range of ~ 5E+4 (factor of 100 increase)
 Integration time is used for normalization and overlap (sufficient)
 Averaging also improves SNR and therefore DR (beam stability)
UBW2012, Berlin, December 2012
linear & log; the “trouble” with the RMS
The two images show exactly the same data (beam profile - (x,y))
but in linear and log scale
Next step is to use such measurements for beam characterization, emittance and Twiss
parameters measurements (add x’ and y’)
Ultimately tomographic measurements are planned; but first just quad scan
=
òx
For non-Gaussian beam RMSXbeam width is a 2tricky thing!
It depends on how much of tails of the distribution function f(x)
RMS
is taken in to account.
w
f (x)dx
UBW2012, Berlin, December 2012
X
wRMS
=
2
x
ò f (x)dx
Quadrupole scan raw data
Level of interest
(LOI)
more tails included
less tails included
UBW2012, Berlin, December 2012
Emittance and Twiss parameters
less tails included
RMS emittance
more tails included
beta function(s)
alpha function(s)
UBW2012, Berlin, December 2012
Diffraction limit and PSF

Imaging  measured distribution is a convolution of source
distribution and so-called Point Spread Function (PSF)

PFS determined by optical system angular acceptance but also
by the source angular distribution. Different beam viewers have
different PSF.

Diffraction determines rather hard limits to the DR

Ways to mitigate: increase angular acceptance, use spatial
filter, coronagraph-like optics
UBW2012, Berlin, December 2012
Objective Lens Pupil Apodization
 First a Lyot’s coronagraph was
considered to improve the PSF, but this
would not allow for simultaneous
measurements of the beam core and
halo, but it is a good exercise
r
Uniform pupil function
 Domain of Fourier optic, always Fresnel
approximation – numerical calculations
required for most of the interesting cases
– becomes demanding on CPU and
memory quickly due to large apertures
and optical wavelength (~ 0.5 um)
z
 Implemented and used quasi-discrete
Hankel transform for optics modeling
(allows to do 1D calculations vs. 2D)
Gaussian
pupil with s=r0/3
optical field propagation by means
of qDHT
 Fourier optics  mage plane = Fourier
transform of pupil function for a point
source (this is the PSF)
(false colors – intensity in log scale)
 Then it is easy to see that the uniform
pupil function, i.e., the harp lens edge
is the problem (besides the uncertainty
principal, which also adds to the
problem)
 Apodization – modification of the pupil
function; First considered Gaussian
amplitude apodization
UBW2012, Berlin, December 2012
Objective Lens Pupil Apodization
 First a Lyot’s coronagraph was
considered to improve the PSF, but this
would not allow for simultaneous
measurements of the beam core and
halo, but it is a good exercise
Point Spread Functions
 Domain of Fourier optic, always Fresnel
approximation – numerical calculations
required for most of the interesting cases
– becomes demanding on CPU and
memory quickly due to large apertures
and optical wavelength (~ 0.5 um)
 Implemented and used quasi-discrete
Hankel transform for optics modeling
(allows to do 1D calculations vs. 2D)
 Fourier optics  mage plane = Fourier
transform of pupil function for a point
source (this is the PSF)
 Then it is easy to see that the uniform
pupil function, i.e., the harp lens edge
is the problem (besides the uncertainty
principal, which also adds to the
problem)
 Apodization – modification of the pupil
function; First considered Gaussian
amplitude apodization
UBW2012, Berlin, December 2012
Convolutions: PFS and 2D Gaussian
Wire scanner measurements
A. Freyberger, in DIPAC05 proceedings,
Measurements made at CEBAF
 wire with diameter much smaller than
the beam size interacts with beam as it
is scanned across it
 there is a number of interaction
mechanisms:
 beam capture
 secondary emission
 scattering
 different ways to detect the signal
 induced current
 secondary particles (counting)
 Only 1D projections (no 2D distribution)
 Takes time (“patience limited”)
 LINAC beams are non-equilibrium
(non Gaussian)
UBW2012, Berlin, December 2012
at JLab FEL
Wire-scanner viewer combination
 Designed to allow measurements of OTR photons,
not Bremsstrahlung X-rays, gammas
 low Z wires for OTR; still can use high Z wires if
Bremsstrahlung is better
 Viewer can be OTR or YAG, or other scintillator
(normal to the beam)
 Mirrors must be very good in terms of scratch and
dig spec. – minimize scattering
 Must have impedance shields (at JLab FEL)
 Two diagnostics at one location
UBW2012, Berlin, December 2012
Wire-scanner electronics etc.

One key problem is the required measurements time
should keep as short as possible to make the diagnostics practical

Maximum useful counting rate is the important parameter

The quest - counting system with the max rate of 100 MHz

PMTs themselves are fast enough for this

PMT base (HV divider) is tricky though, high rate  high current 
loading of HV power supply  changes HV  changes gain (passive
divider are not good for high rate)

There are a few different options for the PMT base, we plat to use socalled transistorized bases (high beta, high voltage transistors)

Preamplifiers, discriminators, logic etc. – commercially available

Need time resolution !!! Options are: gated counting, Time Correlated
Single Photon Counting (TCSPC)

Counter, Multi-channel counters and TCSPC: FPGA implementation
UBW2012, Berlin, December 2012
LDR measurements state

we have demonstrated beam imaging with DR increased by ~ 100

applied the LDR imaging to beam characterization and have shown
that for LINAC non-Gaussian beam the DR has strong impact on the
measurements results

have modeled optics required to improve the DR range to reach 106

new diagnostic station for LDR imaging and cross-check with wire
scanner was designed and built

next1 - practical implementation of the apodization optics
(manufacturing, error sensitivity study, optimization)

next2 - beam measurements with new diagnostics (tomographic
phase space measurements based on LDR imaging)
UBW2012, Berlin, December 2012
CEBAF: overview
Ebeam was 6 GeV is being upgrade to 12 GeV
Bunch charge: 0.2 pC
Repetition rate: 499 MHz (x3)
Three independent beams (3 Halls)
1. Beam halo hitting beam pipe would create
background in the NP detectors
2. FE in LINAC cavities affects the trip rate,
which reduces up time and must be limited
UBW2012, Berlin, December 2012
CEBAF: trip rate, statistics
JLAB-TN-05-57 J. Benesch,
Field Emission in CEBAF's Superconducting RF Cavities and Implications
for Future Accelerators
JLAB-TN-10-008 J. Benesch,
Comparison of arc models from March 2003/Nov 2004 and December 2009
JLAB-TN-12-049 J. Benesch, A. Freyberger,
CEBAF Energy Reach and Gradient Maintenance Needs
 Uses “accounting” and statistical analysis of the trip rate and its dependence on the
cavities gradient
 For 12 GeV CEBAF; 400 cavities + each cavity trips 1/(2 days) would result in on
average 8 RF trips per hour
 Original C25 design / unfortunate feature / RF window has a direct line of sight to the
beam – charges up / eventually break down
 With time performance of cavities degrades i.e. at the same gradient trip rate goes up
exact mechanism is not known (speculated that # of FEs goes up)
 Conclusion – gradient maintenance is needed (reprocessing cavities and refurbishing
the cryo modules)
UBW2012, Berlin, December 2012
CEBAF: trip rate
Initial distribution of gradients of C50 cavities
Distribution of gradients in C25
cavities that resulted in 1/(2 days)
trip par cavity
Distribution of gradients of the same C50 cavities at
the end of 6 GeV operation (~ 4 years later)
Courtesy of J. Benesch
UBW2012, Berlin, December 2012
CEBAF: no Halo ?
 One ways to make large dynamic range measurement is to arrange it to be
frequency measurement
 Then make it work for 1 Hz and for 100 MHz and this is 108 dynamic range.
 For instance use PMT and keep them working in counting mode
Courtesy of A. Freyberger
UBW2012, Berlin, December 2012
CEBAF: vacuum
 Despite the idea/claim that CEBAF beam is quite Gaussian and has
no or very little large amplitude non Gaussian tails,
there are vacuum “events”
 Two types of events:
1. Burn through that require a new piece of beam pipe to be fabricated as it has
a hole drilled into it.
2. Low current, very low intensity lose (chronic lose) that heats up a flange.
This requires Rad. Con. to identify the hot spot, and then
the flange is tighten up and the region recovers quickly.
 Frequency of such events is 1-2 per year (35 weeks of operation)
 Type2 is due to some kind of beam that is not seen (not looked for)
 Type1 (some of them) related to rapid energy change due to RF
changes
 Fortunately it did not happen close to the SRF LINAC
UBW2012, Berlin, December 2012
Conclusion / Summary
 JLab FEL (IR/UV Upgrade)
 RF gradients in LINAC always require attention, set radiation background level (F.E.)
 Drive Laser transport if made very carefully, seems to be not a problem
 Drive Laser rep. rate control (EO cells) always need attention (extinction ration drifts)
 Cathode suffers when conditioning and from breakdowns, still makes beam as needed,
but scatters DL light – generates some halo
 Non-linear beam dynamics is responsible for some fraction of the halo. When setting
up for high current operation, a lot of effort and time goes in to “fitting” the halo
through the recirculator, such that peak beam brightness does not suffer.
 Radiation monitors, BLMs and vacuum are used as tuning diagnostics
 CEBAF
 NP detectors (background) require essentially no beam halo
 Large statistics of cavity performance and its evolution (F.E.)
 Direct effects of F.E. – RF trip rate, reduction of max. possible energy
 Vacuum events related to beam loss (both high and very low current)
UBW2012, Berlin, December 2012
Backup slides
UBW2012, Berlin, December 2012
FEL Injector as an example of #1 (1/6)
downstream of
the gun
UBW2012, Berlin, December 2012
FEL Injector as an example of #1 (2/6)
upstream of the
buncher cavity
UBW2012, Berlin, December 2012
FEL Injector as an example of #1 (3/6)
downstream of the
buncher cavity
UBW2012, Berlin, December 2012
FEL Injector as an example of #1 (4/6)
upstream of the
SRF cavity 1
UBW2012, Berlin, December 2012
FEL Injector as an example of #1 (5/6)
downstream of the
SRF cavity 1
UBW2012, Berlin, December 2012
FEL Injector as an example of #1 (6/6)
downstream of the
SRF cavity 2
UBW2012, Berlin, December 2012
Beam Loss Monitors

The primary BLM at the JLab is a 931B Hamamatsu photo-multiplier tube,
operated with a fixed integrator and individually variable HV power supply

The BLM electronics are 12 channel VME boards.





PMTs are used in current (analog) mode
There is a single FSD fiber output to the MPS for each VME board
Ch3:
Trigger
All 12 channels have analog monitors
thatFault
are connected
H = No Fault
to the Analog Monitoring System (AMS)
L = FSD Fault
These are used as tune-up diagnostics in the control room
Calibration procedure




machine is locked into 1 uA CW operation
beam is driven into chamber and detector gain is varied by changing HV
CH1: Drive Laser Intensity
the HV is adjusted until the system trips
this new “gain” setting is saved in EPICS and accounts for aging of tube.
Courtesy of K. Jordan, D. Sexton
UBW2012, Berlin, December 2012