W. Hofle
CERN AB/RF
Transverse Feedback System
(LHC Damper)
presented by
Wolfgang Hofle
CERN AB/RF/FB
Acknowledgements:
P. Baudrenghien, A. Blas, T. Bohl, A. Findlay, G. Kotzian,
R. Louwerse, T. Linnecar, E. Montesinos, V. Rossi, D. Valuch
& JINR, Dubna, Russia, V. Zhabitskiy
20 October 2006
LHC-LUMI-06 Workshop - Valencia
1
W. Hofle
CERN AB/RF
Transverse Feedback System
(LHC Damper)
Outline

Overview: Transverse feedback systems (dampers) at CERN’s
synchrotrons (PS Booster, AD, LEIR, PS, SPS and LHC); example:
SPS damper; RF group is responsible for all these systems

Functional Specification and overview of LHC Damper System

Expected noise performance for LHC transverse feedback

Possible upgrades of LHC damper system

Upgrades and consolidation of injector transverse feedbacks,
including new PS2

Conclusions
20 October 2006
LHC-LUMI-06 Workshop - Valencia
2
W. Hofle
CERN AB/RF
LHC Transverse Feedback (“LHC Damper”)
Transverse multi bunch feedback
Kicker
Need real-time digital
signal processing
t signal
Signal
processing
t beam
gain g
Match delays:
t signal = t beam + MT 0
T0 : beam revolution time
D
Pick-up 1
Pick-up 2
M=1: very common ->
“One -Turn-Delay” feedback
But M>1 also possible
phase and delay adjustments
• damping: of transverse injection oscillations
• feedback: curing transverse coupled bunch instabilities
• excitation: of transverse oscillations for beam measurements & other applications
20 October 2006
LHC-LUMI-06 Workshop - Valencia
3
W. Hofle
CERN AB/RF
LHC Transverse Damper
Transverse Feedback Systems in Synchrotrons at CERN
Currently installed transverse feedback systems in CERN Synchrotrons
Accelerator
Digital / analogue processing
Power / kicker / bandwidth
Usage in operation
PS Booster (protons)
50 MeV – 1.4 GeV kin. E.
multi turn injection from Linac 2
analogue beam offset signal
suppression, analogue delay
(cables & switches)
100 W, 50 W stripline
Limited to 13 MHz in operation
But built for 100 MHz
bandwidth, baseband
H-plane: used and required
V-plane: beam stable w/o FB
AD (anti-proton decelerator)
3.57 GeV – 0.1 GeV
Copy of booster system
100 W, 50 W stripline
100 MHz bandwidth baseband
used only for excitation purposes
SPS (protons, ions)
(14 – 450) GeV/c protons FT
(26 – 450) GeV/c LHC beam
digital notch filter and 1T-delay
(Altera FPGA, 80 MHz clock)
commissioned in 2000/2001
tetrode amplifiers with two
30 kW tetrodes in push-pull
directly coupled to a kicker (base
band); feedback bandwidth
~10 kHz to 20 MHz
H-plane: used in operation
V-plane: used in operation
used
and
required
for
operation
above 5x1012 protons
(max ~5.5x1013 ppp accelerated)
20 October 2006
LHC-LUMI-06 Workshop - Valencia
4
W. Hofle
CERN AB/RF
LHC Transverse Damper
Transverse Feedback Systems in Synchrotrons at CERN
Current transverse feedback system projects at CERN
Accelerator
Digital / analogue processing
Power / kicker / bandwidth
Planned commissioning
and usage
LEIR ( ions: Pb54+ )
4.2 MeV/u – 72 MeV/u
copy of PS Booster System
New: remote control of pick-up
vector sum
100 W, 50 W stripline
2005
damping during e-cooling may
be necessary
PS (protons, ions)
1.4 GeV – 25 GeV (kinetic E)
synergy with LHC Damper
3 kW pulsed, 200 W CW,
upgrade towards more CW
power possible (loads); 112 W
stripline (0.9 m length), planned
with ~30 MHz bandwidth in
baseband, lower cut-off ~50 kHz
upgrade possible with magnetic
kicker (0.9 m length, 12.5 W, but
bandwidth
considered
too
limited);
H-plane magnetic
kicker built already, but BW
limited
2006
injection damping and feedback
will be beneficial in particular
for high intensity CNGS beams
and LHC beams. Currently
horizontal instabilities are cured
by introducing coupling to the
vertical plane which constrains
the tunes
LHC (protons, ions)
protons: 450 GeV/c – 7 TeV/c
digital notch filter and 1T-delay,
built-in diagnostics
14 bit ADC/DAC
Altera FPGA, 40 MHz clock
new development in progress
tetrode amplifiers with two
30 kW tetrodes in push-pull
directly coupled to kicker (base
band) similar to SPS system
3 kHz -> 20 MHz
2007
injection damping
feedback loop closed during
ramp
switch off during physics?
20 October 2006
LHC-LUMI-06 Workshop - Valencia
5
W. Hofle
CERN AB/RF
LHC Transverse Feedback (“LHC Damper”)
Example SPS (CNGS type beam, 1 batch, 1013 protons):
3 ms growth rate
0.5 ms damping time
injection with
damper off
damper being switched on 7 ms
(300 turns) after injection
20 October 2006
LHC-LUMI-06 Workshop - Valencia
6
W. Hofle
CERN AB/RF
LHC Transverse Feedback (“LHC Damper”)
Example SPS (LHC beam, 48 / 72 bunches spaced 25 ns)
injection with damper off
(~4x1012 p total in 48 bunches)
20 October 2006
injection with damper on
(~6x1012 p total in 72 bunches)
LHC-LUMI-06 Workshop - Valencia
7
LHC Transverse Feedback (“LHC Damper”)
W. Hofle
CERN AB/RF
LHC Damper is being installed in point 4 of LHC along with
the accelerating 400 MHz RF system
20 October 2006
LHC-LUMI-06 Workshop - Valencia
8
Transverse Feedback (“LHC Damper”)
W. Hofle
CERN AB/RF
Nominal LHC filling pattern: 2808 bunches
 Basic bunch spacing is 25 ns (every 10th bucket of 400 MHz RF)
 Running-in with 75 ns bunch spacing possible
 Trains of 72 bunches created in PS accelerator
 Up to 4x72 bunches accelerated in SPS and injected into LHC
 1 ms gap between batches injected into LHC
20 October 2006
LHC-LUMI-06 Workshop - Valencia
9
LHC Transverse Feedback (“LHC Damper”)
W. Hofle
CERN AB/RF
Performance specification (1)
(LHC Design Report)
Beam parameters and requirements for nominal intensity:
Injection beam momentum
Static injection errors
ripple (up to 1 MHz)
resistive wall growth time
assumed de-coherence time
tolerable emittance growth
Overall damping time
450 GeV/c
2 mm (at bmax=183 m)
2 mm (at bmax=183 m)
18.5 ms
68 ms
2.5 %
4.1 ms (46 turns)
bunch spacing
minimum gap between batches
lowest betatron frequency
highest frequency to damp
25 ns
995 ns
> 2 kHz
20 MHz
20 October 2006
LHC-LUMI-06 Workshop - Valencia
10
LHC Transverse Feedback (“LHC Damper”)
W. Hofle
CERN AB/RF
Performance specification (2)
Equipment performance specification:
choice:
aperture
“electrostatic kickers” (“base-band”)
52 mm
kickers per beam and plane
length per kicker
nominal voltage up to 1 MHz at b=100m
kick per turn at 450 GeV/c
4
1.5 m
+/- 7.5 kV
2 mrad
rise-time 10-90%, DV= +/- 7.5 kV
rise-time 1-99%, DV= +/- 7.5 kV
350 ns
720 ns
must provide sufficient gain from
1 kHz to 20 MHz
noise must be less than quantization noise due to 10 bit / 2s
20 October 2006
LHC-LUMI-06 Workshop - Valencia
rise time fast
enough for gap of
38 missing bunches
11
W. Hofle
CERN AB/RF
Transverse Feedback (“LHC Damper”)
The LHC Transverse Damping System (high power part)
Damper system
IP4
H
H
H
H
V
V
V
V
V
V
V
V
H
H
H
H
Beam 1
Beam 2
“Electrostatic”
kicker
Wideband
amplifier
Unit


20 electrostatic kickers
40 wideband amplifiers,
i.e. 40 tetrodes
(RS2048 CJC, 30 kW)
20 amplifier cases
20 October 2006
+
Spare units
H
H
V
V
Module
LHC-LUMI-06 Workshop - Valencia
12
W. Hofle
CERN AB/RF
LHC Transverse Damper
LHC optics at injection in IR4
(beams do not cross!)
Beam 1
high horizontal beta left of IP4
high vertical beta right of IP4
Beam 2
high horizontal beta right of IP4
high vertical beta left of IP4
Optics 6.4, sorry, 6.5 not very different for ADT)
20 October 2006
LHC-LUMI-06 Workshop - Valencia
13
W. Hofle
CERN AB/RF
LHC Transverse Damper
Achievable performance (“baseline”) without upgrade
LHCADT performance in LHC optics version 6.500 compared to original assumptions (at 450 GeV/c),
assuming 7.5 kV maximum kick voltage
b=100 m performance
Optics 6.500 performance
Kick per turn in s
Kick per turn in s @ b in m
ADTH beam 1
0.2 s
0.271 s at b=183 m
ADTH beam 2
0.2 s
0.274 s at b=189 m
ADTV beam 1
0.2 s
0.301 s at b=227 m
ADTV beam 2
0.2 s
0.331 s at b=275 m
Estimate of maximum capabilities (usage as beam exciter, abort gap cleaning etc.), assumes optics 6.5
as in table above, 450 GeV/c and 7 TeV and running with up to ~15 kV DC for tetrode anode voltage (at 1 MHz 1.4x nominal)
1 MHz
10 MHz
20 MHz
100 kHz
ADTH
0.42 s / 0.11 s
0.38 s / 0.10 s
0.12 s / 0.03 s
0.044 s / 0.011 s
ADTV
0.46 s / 0.12 s
0.42 s / 0.12 s
0.14 s / 0.04 s
0.049 s / 0.012 s
20 October 2006
LHC-LUMI-06 Workshop - Valencia
14
W. Hofle
CERN AB/RF
LHC Transverse Damper
Physical layout in point 4 underground LHC
ADT racks
(driver amplifiers,
PLC controls, fast
interlocks)
ADT (4 modules)
left of IP4
+ space for 2 more
modules (upgrade)
ADT (4 modules)
right of IP4
+ space for 2 more
modules (upgrade)
Beam 2
Beam 1
20 October 2006
LHC-LUMI-06 Workshop - Valencia
15
W. Hofle
CERN AB/RF
LHC Transverse Damper
Overview of one Damper system
Most of electronics on surface -> easy access !
20 October 2006
LHC-LUMI-06 Workshop - Valencia
16
Transverse Feedback (“LHC Damper”)
W. Hofle
CERN AB/RF
LHC ADT Low level electronics

System will be based on experience gained with SPS Damper

Two coupler type pick-ups with betatron phase advance of ~90 degrees will be
used (BPMC/BPMCA, provided by the BDI group and integrated in quad)

Spectrum repeats every 40 MHz (25 ns bunch spacing) -> analogue down
conversion with a multiple of 40 MHz to baseband , LO of 400 MHz chosen

One of the horizontal pick-ups is installed at ~zero dispersion and it will be
possible to use this pick-up and a Hilbert transformer (FIR filter) for the
phase adjustment if wanted; Hilbert filter tested in SPS but not used in
operation (reduced phase stability margin at very high gain)

FPGA technology for implementing most of the functions digitally (notch
filter, fine delay adjustment, betatron phase adjustment by two pick-ups or
Hilbert filter and single pick-up use)
20 October 2006
LHC-LUMI-06 Workshop - Valencia
17
Transverse Feedback (“LHC Damper”)
W. Hofle
CERN AB/RF
Coupler type pick-ups

Length of electrodes 150 mm

Frequency domain: maximum of transfer impedance 6.46 W @ 500 MHz

Peak voltage for ultimate beam @ collision: ~140 V -> very large

Sensitivity: 0.314 mW/mm => 9.6 V/mm peak in time domain after ideal hybrid

install hybrid on surface (12 dB cable attenuation will avoid saturation) -> peak
voltage reduced to ~25 V, OK for wideband hybrid H9

Hybrid: 2 dB (+3 dB -1 dB), estimate: cable and bandpass filter generating) 8
periods of 400 MHz

after filtering: ultimate beam: ~0.38 V/mm @ 400 MHz; nominal 0.26 V/mm
(duty cycle 0.8)
20 October 2006
LHC-LUMI-06 Workshop - Valencia
18
W. Hofle
CERN AB/RF
LHC Transverse Damper
Signal Levels from pick-ups
LHC Beam Parameters
Injection
Collision
Beam Energy
450 GeV
7000 GeV
g
479.6
7461
RMS bunch length in cm
11.24
7.55
FREV
11.245
11.245
400.789
400.790
FRF
in kHz
in MHz (h=35640)
^ (w) = 6.46 W)
Range of intensities for LHC beam and expected pick-up signal levels (ZT = 3.23 W; Z
T
Pilot
Nominal beam
Ultimate beam
Particles per bunch
5x109
1.15x1011
1.7x1011
Number of bunches
1
2808
2808
Circulating current (DC) in A
9 x 10-6
0.582
0.860
Bunch peak current @ injection in A
0.9
19.6
29.0
Bunch peak current @ collision in A
1.3
29.2
43.1
Peak Voltage from PU @ inj. in V
2.8
63.4
93.6
Peak Voltage from PU @ coll. in V
4.1
94.3
139.4
20 October 2006
LHC-LUMI-06 Workshop - Valencia
19
Transverse Feedback (“LHC Damper”)
W. Hofle
CERN AB/RF
Performance with respect to noise:
aim at a better than 1 mm resolution

Assume bunches oscillate with 1 mm rms (bunch-to-bunch)

Power available from pick-up @400 MHz (+/- 20 MHz): 433 pW (nominal
beam)

Thermal noise at 290 K: kBT = 4x10-21 W/Hz; in 40 MHz BW: 0.16 pW

Digitization with 14 bit: 16384 discrete levels, assume 1 mm -> 4 steps then 14
bit are sufficient to cover +/- 2 mm

Margin with respect to thermal noise: To use this margin we need to limit
orbit variations at the pick-ups to less than +/- 2 mm.
20 October 2006
LHC-LUMI-06 Workshop - Valencia
20
Transverse Feedback (“LHC Damper”)
W. Hofle
CERN AB/RF
LHC ADT upgrade possibilities (1)
In case of lack of kick strength

Push system to its limits +40% feasible (pulsed mode, reliability?)

Increase of beta functions at kickers (to be checked), scales with b1/2

Additional kickers (expensive) but space is reserved (4 MCHF for +50%)
In case of large instability growth rates @ limit of bandwidth (20 MHz)

New power amplifiers with increased bandwidth, possibly dedicated damping
of modes which are unstable

New kickers (matched strip-line) and power amplifiers -> expensive; signal
processing electronics can be re-used
20 October 2006
LHC-LUMI-06 Workshop - Valencia
21
Transverse Feedback (“LHC Damper”)
W. Hofle
CERN AB/RF
LHC ADT Upgrade possibilities (2)
In case of insufficient Signal-to-Noise ratio (S/N)

Digitization of pick-up signals underground -> save 600 m of cable -> gain 12 dB
(factor 4), but need to develop hybrids, filters and electronics to process
large amplitude signals; alternatively use 1 5/8” cable: gain ~6 dB (factor 2)

If limited by interference & noise on link from surface to tunnel (3 kHz20 MHz) -> built digital link and have Digital-to-Analogue conversion as
close as possible to amplifiers (possibly find new space for driver
amplifiers that are now in UX45)

Increase beta functions at pick-ups (not very likely but worth a thought
since otherwise no additional cost); in particular vertical beta function at
BPMC.7L4.B2 for beam 2 somewhat low in optics version 6.5 (=87 m)

Average in the signal processing over several turns, i.e. 9 turns, gain of a factor 3
in S/N if noise turn-by-turn uncorrelated - more turns not feasible due to limited
signal processing power and sensitivity to tune.

Use additional pick-ups; strip-line pick-ups at Q8 and Q10 may be available;
gain in S/N 50%, additional cables and electronics must be installed; further
improvement by installing additional pick-ups, spread around the ring?
20 October 2006
LHC-LUMI-06 Workshop - Valencia
22
Transverse Feedback (“LHC Damper”)
W. Hofle
CERN AB/RF
Strategy for transverse feedback projects and
upgrades in injectors

Goal: Move towards common hardware for signal processing for all dampers

Currently strong push to commission PS transverse feedback system -> in work

Concentrate resources on LHC transverse feedback commissioning in 2007-2008

Consolidate Booster/AD/LEIR Systems from 2008 onwards; use of LHC signal
processing hardware to be checked (bandwidth required, large frequency swing!)

Retrofit LHC hardware to SPS in 2009; possibly install third horizontal system to
cure ongoing e-cloud problem (horizontal systems are at limit); SPS damper needs
upgrade to go to ultimate

Use as much as possible the experience with the new PS transverse feedback
system (to be gained during next years) when defining system needed for the
PS2

Keep experience in operation and electronics development close together (same
team) to be able to have short development cycles
20 October 2006
LHC-LUMI-06 Workshop - Valencia
23
Transverse Feedback (“LHC Damper”)
W. Hofle
CERN AB/RF
Potential benefits of injector transverse feedbacks
upgrades for LHC

Upgrade a must for SPS [unless one changes vacuum chamber]: resistive wall
and electron cloud instability; upgrade required to go to the ultimate beam
intensity; damper control upgrade to LHC standards

It is likely that we need in the PS the transverse feedback working for ultimate
beam

Booster requires in the horizontal plane a transverse feedback system for
stabilization, future upgrade required due to obsolete hardware -> assure future
reliability
20 October 2006
LHC-LUMI-06 Workshop - Valencia
24
Transverse Feedback (“LHC Damper”)
W. Hofle
CERN AB/RF
Conclusions

Before considering to implement upgrades in injectors, finish LHC
transverse feedback first

Better than 1 mm resolution seems achievable for LHC
transverse feedback

Upgrade possibilities for LHC transverse feedback exist (kick
strength, bandwidth, noise performance

Upgrade and consolidation needs in injectors for ultimate beam need
attention

Build PS2 with transverse feedback from start
20 October 2006
LHC-LUMI-06 Workshop - Valencia
25