Content uploaded by G. Guaglio
Author content
All content in this area was uploaded by G. Guaglio on Apr 28, 2014
Content may be subject to copyright.
EUROPEAN ORGANIZATION FOR NUCLEAR RESEARCH
CERN ⎯ AB DEPARTMENT
Geneva, Switzerland
February, 2006
CERN-AB-2006-009 BI
Beam Loss Monitoring System for the LHC
E.B. Holzer, B. Dehning, E. Effinger, J. Emery, G. Ferioli, J.L. Gonzalez,
E. Gschwendtner, G. Guaglio, M. Hodgson, D. Kramer, R. Leitner, L. Ponce, V.
Prieto, M. Stockner, C. Zamantzas.
CERN – Geneva - Switzerland
Abstract
One of the most critical elements for the protection of CERN's Large Hadron Collider (LHC) is its
beam loss monitoring (BLM) system. It must prevent the superconducting magnets from quenching
and protect the machine components from damages, as a result of critical beam losses. By
measuring the loss pattern, the BLM system helps to identify the loss mechanism. Special monitors
will be used for the setup and control of the collimators. The specification for the BLM system
includes a very high reliability (tolerable failure rate of 10
-7
per hour) and a high dynamic range of
10
8
(10
13
at certain locations) of the particle fluencies to be measured. In addition, a wide range of
integration times (40 μs to 84 s) and a fast (one turn) trigger generation for the dump signal are
required. This paper describes the complete design of the BLM system, including the monitor types
(ionization chambers and secondary emission monitors), the design of the analogue and digital
readout electronics as well as the data links and the trigger decision logic.
Presented at IEEE NSS – 23-29 Oct. 2005 – San Juan/Puerto Rico
Beam Loss Monitoring System for the LHC
Eva Barbara Holzer, Bernd Dehning, Ewald Effinger, Jonathan Emery, Gianfranco Ferioli, Jose Luis Gonzalez,
Edda Gschwendtner, Gianluca Guaglio, Michael Hodgson, Daniel Kramer, Roman Leitner, Laurette Ponce,
Virginia Prieto, Markus Stockner, Christos Zamantzas
Abstract— One of the most critical elements for the protection of
CERN’s Large Hadron Collider (LHC) is its beam loss monitoring
(BLM) system. It must prevent the superconducting magnets from
quenching and protect the machine components from damages, as
a result of critical beam losses. By measuring the loss pattern, the
BLM system helps to identify the loss mechanism. Special monitors
will be used for the setup and control of the collimators. The
specification for the BLM system includes a very high reliability
(tolerable failure rate of
per hour) and a high dynamic range
of
( at certain locations) of the particle fluencies to be
measured. In addition, a wide range of integration times (40
s to
84 s) and a fast (one turn) trigger generation for the dump signal
are required. This paper describes the complete design of the
BLM system, including the monitor types (ionization chambers
and secondary emission monitors), the design of the analogue and
digital readout electronics as well as the data links and the trigger
decision logic.
Index Terms— beam loss, beam instrumentation, machine pro-
tection, quench protection, damage protection.
I. INTRODUCTION
T
HE start-up of the LHC is scheduled for 2007. An
unprecedented amount of energy will be stored in its
circulating beams. The loss of even a very small fraction of
this beam may induce a quench in the superconducting magnets
or cause physical damage to machine components. A fast (one
turn) loss of
and a constant loss of times the
nominal beam intensity can quench a dipole magnet. A fast loss
of times nominal beam intensity can damage a magnet.
The stored energy in the LHC beam is a factor of 200 (or more)
higher than in existing hadron machines with superconducting
magnets (HERA, TEVATRON, RHIC), while the quench levels
of the LHC magnets are a factor of about 5 to 20 lower than
the quench levels of these machines. The detection of the lost
beam particles allows protecting the equipment by generating a
beam dump trigger when the losses exceed certain thresholds.
These thresholds depend on the momentum of the stored beam,
the duration of the beam loss and on the location of the beam
loss monitor. In addition to the quench prevention and damage
protection, the loss detection allows the observation of local
aperture restrictions, orbit distortions, beam oscillations and
particle diffusion, etc. Since a repair of a superconducting
magnet would cause a down time of several weeks or months,
the protection against damage has highest priority.
Manuscript received November 11, 2005.
E.B. Holzer is the corresponding author (e-mail: barbara.holzer@cern.ch).
All authors are with CERN, CH-1211 Geneva 23, Switzerland (e-mails:
firstname.lastname@cern.ch).
II. SPECIFICATIONS AND REQUIREMENTS
The functional specifications of the BLM system are defined
in [1].
A. Families of BLMs
There are four different families of beam loss monitors. They
are listed in Table I. The highest number of monitors, BLMA,
will be installed around the quadrupole magnets all around
the ring (six per quadrupole). They constitute local aperture
minima, and therefore likely loss locations. The second family,
BLMS, will be installed at global aperture limits other than
the collimators (i.e. final focus magnets of the experimental
insertions at 7 TeV) and other critical loss locations (e.g.
losses due to beam injection or extraction errors). One set of
detectors, BLMC, will be installed after each collimator. They
will be used to set the position of the collimator jaws and to
continuously monitor their performance. All these monitors use
the same particle detectors (see Section III-A) and the same
readout electronics (described in Section III-B). The extended
dynamic range of
is realized by installing two detectors
with different sensitivity next to each other. The minimum
acquisition time of this electronics is 40 s, covering both
required time resolutions. In addition there will be a set of
movable BLMs to cover unexpected loss locations. For beam
studies there is the possibility to ignore a beam abort signal
from the maskable monitors, if the stored energy in the beam
does not reach damage potential. The BLMS and BLMC are
not maskable. All non-maskable monitors have to be available
to allow beam injection into the LHC.
The fourth family, BLMB, will not be described in this paper.
It will only be installed after the commissioning of the LHC,
to be used for dedicated beam studies on a bunch scale.
B. Quench Levels and Observation Range
The dynamic range of the system is given by the calculated
damage and quench levels (see Figure 1) and the expected
usage from pilot beam intensity to ultimate beam intensity. The
dynamic range for the arc monitors (BLMA) is . At a certain
number of the BLMS locations higher loss rates could occur
due to machine component failures. The collimation sections
(BLMC) will have the highest continuous loss rates. Therefore,
the dynamic range at these monitor locations is extended to
. The observation time range is defined by the fastest
possible use of the trigger signal by the beam dump (on the
TABLE I
FAMILIES A N D LOCAT I O N S O F BEAM L O S S M O N I T O R S
Type Locations Purpose Mask-
able
Dynamic
Range
Time
Resolution
Number of
Monitors
BLMA All along the rings
(6 per quadrupole)
Protection of superconducting
magnets
yes 2.5 ms
BLMS Critical aperture limits
or critical positions
Machine protection and diag-
nostics of losses
no or 1 turn
(89 s)
BLMC Collimation sections Set-up the collimators and
monitor their performance
no 1 turn
(89 s)
BLMB Primary collimators Beam studies yes 1 bunch
side of short integration intervals) and the response time of the
helium temperature measurement system (on the long side). The
required time resolution for the arc monitors is 2.5 ms and for
all other monitors 89 s (1 turn). With this 1 turn resolution it
will be possible to allow beam extraction with a maximal delay
of three turns. The longest integration time is 84 s.
C. Failure Rate and Availability
The measurement system failure rate and the availability
requirements have been evaluated using the Safety Integrity
Level (SIL) approach [2]. A downtime cost evaluation is used
as input for the SIL approach. The beam loss monitor system
is critical for short and intense particle losses, while at medium
and longer loss durations it is assisted by the quench protection
system and the cryogenic system. The required probability of
not detecting a dangerous beam loss, and therefore losing a
magnet, is per year, which corresponds to SIL3. The
unavailability of the BLM system has been calculated (using
the program Reliability Workbench V10.0, ISOGRAPH) to be
per channel. Assuming 100 dangerous losses per year
this satisfies the SIL3 requirement. The required probability of
generating a false dump is calculated to be to per
hour (SIL 2) per channel, corresponding to 20 false dumps per
year. The simulation of the BLM system yields 10 to 17 false
beam aborts per year, again satisfying the SIL2 requirement.
A detailed record of the reliability calculations for the BLM
system and for the whole LHC can be found in [3], [4] and [5]
respectively.
III. ARCHITECTURE OF THE BLM SYSTEM
A. Detectors
Signal speed and robustness against aging were the main
design criteria for the detectors. Because of the high dynamic
range two types of detectors will be used. The standard moni-
tors are ionization chambers with parallel aluminum electrode
plates separated by 0.5 cm, as shown in Figure 2. The detectors
are about 50 cm long with a diameter of 9 cm and a sensitive
volume of 1.5 liter. The collection time of the electrons and
ions is of the order of 300 ns and 80 s respectively. The
chambers are filled with N at 100 mbar overpressure. The
composition of the chamber gas is the only component in the
BLM system which is not remotely monitored (see Section III-
B). The properties of the chamber gas are sufficiently close to
the ones of air at ambient pressure (i.e. inside a detector which
has developed a leak) not to compromise the precision of the
BLM system, but sufficiently different to detect a leak during
the scheduled annual test of all the chambers with a radioactive
source.
Fig. 2. Photograph of the inside of an ionization chamber. The stack of
aluminum electrodes with the insulator ceramics at both ends can be seen.
At locations with very high (potential) loss rates (about 300)
the ionization chambers will be complemented by secondary
emission monitors. They are based on the same design, but hold
only three electrodes. The signal (middle) electrode is made out
of titanium, as its secondary emission coefficient shows better
stability as the integrated dose increases [6]. The chamber is
10 cm long, the pressure inside has to stay below bar.
The sensitivity is about a factor of smaller than in the
ionization chamber.
Both chambers are operated at 1.5 kV and are equipped with
a low pass filter at the high voltage input. The dynamic range
of the detectors is higher than . It is limited by leakage
currents through the insulator ceramics at the lower end and by
saturation due to space charge at the upper end.
The estimated radiation dose on the detectors during 20 years
of LHC operation is Gray in the collimation sections and
Gray at the other locations. To avoid radiation aging
(electronegative gases, organic compounds) a strict cleaning
procedure for the chambers is followed (including glow dis-
charge cleaning for the collimation section detectors). Impurity
levels due to thermal and radiation induced desorption are
estimated to stay in the ppm range. No organic material is
1.E+04
1.E+05
1.E+06
1.E+07
1.E+08
1.E+09
1.E+10
1.E+11
1.E+12
1.E+13
1.E+14
1.E+15
1.E+16
1.E+17
1.E+18
1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06
Duration of loss [ms]
Quench levels [proton/s ]
BLMS* & BLMC
Quench level and observation range
450 GeV
7 TeV
Damage levels
Special &
Collimator
1 turn
Arc
2.5 ms
He heat flow
He heat reserve
heat flow between
cable and He
heat reserve
of cable
Fig. 1. Quench levels of the LHC bending magnets as function of loss duration at 450 GeV and at 7 TeV (dark green and dark blue). The required observation
range for both energies is indicated in light green and light blue color. The damage levels for short and long duration losses are indicated as well.
present, neither in the production process (pumping, baking
and filling) of the detectors, nor in the detectors themselves.
The positioning of the detectors was determined by simu-
lation studies (see Section IV-A). In the arcs, three monitors
per beam will be installed around each quadrupole located in
the horizontal plane defined by the beam vacuum tubes (see
Figure 3). At this position the secondary particle fluence is
highest and the best separation of the losses from the two
beams is reached. Their longitudinal positions are about 1 m
downstream of the most likely loss locations.
x
y
z
MQML
right detector
left detector
D
F
beam1
iron yoke
copper
coils
vacuum vessel
beam2
shrinking
cylinder
Fig. 3. Cross section of a quadrupole cryostat with the two beams. The
location of the BLMs is indicated on the outside of the cryostats.
B. Acquisition System
The electrical signals of the detectors are digitized with a
current to frequency converter and these pulses are counted
over a period of 40 s (see Figure 4). The counter value
is transmitted every 40 s to the surface analysis electronics
using a high speed optical link (with a cyclic redundancy
check). The signal treatment and transmission chain is doubled
after the current to frequency conversion to meet the required
failure rate probability of to per hour. The surface
electronics calculates the integrated loss values and compares
them to a table of loss duration and beam energy depended
threshold values. Warning information is transmitted by a
software protocol. The beam abort signals are transmitted to
the beam dump kicker magnets using the LHC beam interlock
system (LBIS). The beam energy information is received over a
dedicated fiber link. Details to the readout system can be found
in [7] and [4].
The analog electronics is located below the quadrupole
magnets in the arc. For all detectors of the dispersion suppressor
and the long straight sections the electronics is located in side
tunnels to the LHC. All components of the tunnel electronics
are radiation certified to 500 Gray. The dose expected at the
electronics locations is about 20 Gray per year. The analog
signal transmission cables have a length of a few meters in
the LHC arcs and up to 500 m in the long straight sections.
This part of the transmission is subject to the injection of
electromagnetic crosstalk and noise.
The availability of all electronics channels is constantly
monitored and radiation dose induced drifts in the electronic
channels are corrected for (up to a maximum level, which
corresponds to 10% of the lowest beam abort threshold value).
The availability of all detectors, the acquisitions chain and
the generation and communication of the beam abort signal is
verified for each channel before each injection into the LHC.
The BLM system will drive an online event display and write
extensive online logging (at a rate of 1 Hz) and postmortem
data (up to 1000 turns plus averages of up to 10 minutes) to a
database for offline analysis.
R
C
R
C
R
C
R
C
R
C
High Voltage
R
C
BLM 1
ARC Quadrupole
8 m
5 m 5 m
5 m 5 m
3 m
3 m
3 m
3 m
6 m
10 m
11 m
7 m
4 m
Optical link
Beam loss installation Q12 --- Q34
Sout
HVin
1 2 3
4
5 6 7
8
IN
BLM frontend
electronics
BLM 2 BLM 3
BLM 6 BLM 5
BLM 4
Current to
frequency
converter
Multi-
plexer
LASER
Diode
Demulti-
plexer
Threshold
comparator
LBIS (Dump)
Beam Permit
Beam Energy
VME Bus
Analog Electronics
Digital Electronics
Below Quad. in ARC
elsewhere in RR, UA, UJ
Ionisation
chamber
or SEM
At surface (SR, SX)
Fig. 4. Schematic view of the signal transmission chain and the BLM
installation around one arc quadrupole.
IV. THRESHOLD CALIBRATIONS
The BLM interlock limits can be set for each of the about
4000 chambers individually. In the arcs they will be set to 30%
of the magnet quench levels. They vary with integration time
(12 integration time intervals between 89
s and 84 s) and the
energy of the beam (32 energy ranges). The determination of
the thresholds is based on simulations. A number of simulations
have to be combined for these calibrations. Whenever possible,
crosschecks of the simulations by measurements are performed
before the start-up of the LHC. A factor of 5 and a factor
of 2 are the specified initial and final absolute precisions on
the prediction of the quench levels respectively. The relative
precision for quench prevention is requested to stay below 25%.
A. Simulations
The aim of the simulation is to relate the BLM signal to the
number of locally lost beam particles, to the deposited energy
in the machine components and ultimately to the quench and
damage levels. The distribution of the loss locations along the
LHC is simulated by particle tracking with a detailed aperture
model [8], see e.g. Figure 5.
The lost beam particles initiate hadronic showers. Proton
induced showers through cold magnets in the LHC arc and
dispersion suppressor [9] and through the collimators [10] have
been simulated. These simulations yield the heat load on the
magnets (or the collimators) and the particle fluence at the
location of the beam loss monitors, see e.g. Figure 6.
Magnet quench levels as a function of beam energy and
loss duration have been calculated [11] and will also be
Fig. 5. Simulated proton loss locations along the LHC at injection energy for
beam 1. The lost protons are given in bins of 10 cm. In the upper picture losses
in warm magnets are marked in red, losses in the collimators in black and losses
in cold magnets in blue. The lower picture shows the loss pattern around one arc
quadrupole of beam halo particles escaping the betatron cleaning insertion in
IP7. In blue color the quadrupole magnet and in green color the neighboring
dipole magnets are shown. The beam beta functions are given as well. The
dispersion function indicated is not to scale. The loss pattern (horizontal losses
in this region) is explained by the beam optics functions. The strong peak at the
entrance of the quadrupole magnet is due to the enlarged vacuum chamber at
the bellows between the magnets. (Data courtesy of R.W. Assmann, G. Robert-
Demolaize and S. Redaelli.)
simulated [12]. The signal response of the ionization chamber to
the mixed radiation field in the tail of the hadronic shower has
been simulated [13] and measured in various beams at CERN.
The corresponding simulations for lead ion beams are being
performed as well [14].
B. Measurements
The uncertainties in the threshold level determination are
dominated by our knowledge of the longitudinal loss distri-
bution and of the magnet quench levels. Hence, the future
investigations will concentrate on these points. Quench level
measurements on LHC magnets for different time constants
(without beam) are planned [12]. A beam loss measurement
program at HERA/DESY has started in 2004. It aims to
crosscheck the simulation of hadronic showers through super-
conducting magnets and of the chamber response to the mixed
radiation field in the tail of the hadronic shower.
In the LHC itself a sector test with beam could take place in
2006. It would allow to crosscheck the threshold simulations
120.PROX
0
20
40
60
80
100
120
9000 10000 11000 12000 13000 14000
z [cm]
10
-4
ch./p/cm
2
210.PROX
0
25
50
75
100
125
150
175
200
9000 10000 11000 12000 13000 14000
z [cm]
10
-4
ch./p/cm
2
MBB MBB
MQML
MCBCB
MBB
MCS MCSMCS
MBAMCDO
MBB MBB
MQML
MCBCB
MBB
MCS MCSMCS
MBAMCDO
Q10
Q10
beam1
beam2
loss@ 11448cm
loss@ 11448cm
left detector signal
right detector signal
shower maximum @ 11560cm
shower maximum @ 11360cm
Fig. 6. Simulated longitudinal particle shower distribution at the outside of
the cryostat in the horizontal plane of the beam pipes. Proton impact positions
are in the center of the quadrupole. The maximum signal occurs one meter
after the proton impact position.
for instant losses at 450 GeV. The tuning of the BLM interlock
levels will begin with the first beam, using the beam loss and
magnet quench data of the logging and post mortem database
respectively. Dedicated beam tests will be required if the
“parasitic” tuning speed of the BLM system cannot keep pace
with the increasing demand on precision, as the beam intensity
and energy increase during LHC commissioning. Apart from
damage protection, the threshold levels also have to be precise
enough not to compromise the operation efficiency by false
dumps or magnet quenches.
V. CONCLUSION
The LHC tolerates less fractional beam loss than any existing
hadron machine because it features higher stored beam energy
in combination with superconducting magnets which withstand
less energy deposition. This lead to challenging requirements
on the beam loss monitoring system and to several novel
features in respect to existing systems. They include a large
dynamic range and a high reliability. To satisfy the reliability
requirements radiation tolerant electronics and a failsafe sys-
tem design is employed (reliable components, redundancy and
voting for less reliable components and a constant monitoring
of the availability of the channels). The beam abort thresholds
in the LHC system change dynamically not only with the beam
energy but also with the duration of the loss. A high accuracy in
the quench level determination (requiring extensive simulation
studies) became necessary for machine protection on the one
hand and for operational efficiency (to avoid false dumps) on
the other hand.
ACKNOWLEDGMENT
The authors would like to acknowledge the contributions
of various members of the LHC Machine Protection Working
Group [15] and the LHC Collimation Working Group [16].
REFERENCES
[1] B. Jeanneret et al., Functional Specification on the Measurement of the
Beam Losses in the LHC Rings, LHC-BLM-ES-0001.00 v.2.0, 2004.
[2] Functional Safety Of Electrical/Electronic/Programmable Electronic
Safety Related Systems, IEC 61508 First edition 1998-11.
[3] G. Guaglio, B. Dehning and C. Santoni, Reliability of LHC Beam
Loss Monitors System for the Large Hadron Collider, ICFA HB2004,
Bensheim, Germany; AIP Conf.Proc.773:191-196, 2005.
G. Guaglio, B. Dehning and C. Santoni, Reliability Considerations of
Beam Loss Monitors System for the Large Hadron Collider, CERN-AB-
2004-029-BDI, BIW 04, Knoxville, USA; AIP Conf.Proc.732:141-149,
2004.
[4] G. Guaglio, Reliability of the Beam Loss Monitor System for the Large
Hadron Collider at CERN, unpublished, PhD Thesis, Universit´e Clermont
Ferrand II - Blaise Pascal, to be submitted 2005.
[5] R. Filippini et al., Reliability Assessment of the LHC Machine Protection
System, CERN-LHC-Project-Report-812, PAC 05, Knoxville, Tennessee,
USA, 2005.
[6] G. Ferioli and R. Jung, Evolution of the secondary emission efficiencies
of various materials measured in the CERN SPS secondary beam lines,
CERN-SL-97-71-BI, DIPAC 97, Frascati, Italy, 1997.
[7] W. Friesenbichler, Development of the Readout Electronics for the Beam
Loss Monitors of the LHC, CERN-THESIS-2002-028, Diplomarbeit,
Fachhochschule Wr. Neustadt, 2002.
C. Zamantzas et al., The LHC beam loss monitoring system’s real-time
data analysis card, CERN-AB-2005-082, DIPAC 05, Lyon, France, 2005.
C. Zamantzas, The Design, Construction, Test and Integration of the LHC
Beam Loss Monitor Threshold Comparator (BLMTC), unpublished, PhD
Thesis, Brunel University, London, UK, to be submitted 2006.
[8] R.W. Assmann et al., Expected Performance and Beam-based Optimiza-
tion of the LHC Collimation System, EPAC 04, Lucerne, Switzerland,
2004.
E.B. Holzer and B. Dehning, Longitudinal Loss Distribution along the
LHC, EPAC 04, Lucerne, Switzerland, 2004.
S. Redaelli, LHC Aperture and Commissioning of the Collimation System,
Chamonix XIV workshop, CERN, 2005.
G. Robert-Demolaize et al., A new version of SixTrack with collimation
and aperture interface, CERN-AB-2005-033, PAC 05, Knoxville, Ten-
nessee, USA, 2005.
[9] A. Arauzo and C. Bovet, Beam loss detection system in the arcs of the
LHC, CERN-SL-2000-052-BI, 2000.
A. Arauzo and B. Dehning, Configuration of the beam loss monitors for
the LHC arcs, LHC Project Note 238, 2000.
E. Gschwendtner et al., Ionization chambers for the LHC beam loss
detection, DIPAC 03, Mainz, Germany, 2003.
E. Gschwendtner et al., The beam loss detection system for the LHC ring,
EPAC 02, Paris, France, CERN SL-2002-021 BI, 2002.
[10] I.A. Kurotchkin et al., unpublished simulation studies.
[11] J.B. Jeanneret et al., Quench levels and transient beam losses in LHC
magnets, LHC Project Report 44, CERN, 1996.
A. Arauzo-Garcia et al., LHC Beam Loss Monitors, DIPAC 01, Grenoble,
France, CERN-SL-2001-027-BI, 2001.
[12] A. Siemko, Beam loss induced quench levels, Chamonix XIV workshop,
CERN, 2005.
[13] M. Hodgson, Beam Loss Monitor Design Investigations for Particle
Accelerators, Masters Thesis, University of Surrey, Guildford, Surrey,
UK, 2005.
[14] H.-H. Braun et al., Collimation of Heavy Ion Beams in LHC, EPAC 04,
Lucerne, Switzerland, 2004.
J.M. Jowett et al., Limits to the Performance of the LHC with Ion Beams,
EPAC 04, Lucerne, Switzerland, 2004.
J.M. Jowett, LHC Operation with Heavy Ions, Chamonix XIV workshop,
CERN, 2005.
[15] http://cern.ch/lhc-mpwg
[16] http://cern.ch/lhc-collimation
