From: "Frank Zimmermann"
To: "Frank Zimmermann" ; ; "Francesco Ruggiero" ; "Daniel Schulte" ; "Elias Metral" ; "Frank Schmidt" ; "Gilbert Guignard" ; "Jacques Gareyte" ; "Werner Herr" ; "Luc Vos" ; "Alex Koschik" ; "Bruno Muratori" ; "Tommaso D'Amico" ; "Rita Paparella" ; "Walter Wittmer" ; "Gregory Penn" ; "Lifshitz Ronen" ; "Maxim Korostelev" ;
Cc: "Jean-Pierre Riunaud" ; "Karlheinz Schindl" ; "Louis Rinolfi" ; "Michel Martini" ; "Oliver Bruning" ; "Roberto Cappi" ; "Charles Hill" ; "Gianluigi Arduini" ; "Helmut Burkhardt" ; "Daniel Brandt" ; ; "Ralph Wolfgang Assmann"
Subject: Minutes of LHC Collective Effects meeting 11/04/2003
Date: Monday, April 14, 2003 4:58 PM
------------------------------------------------------------------------
minutes ABP-LCE meeting 11.04.03
present: EB, JJ, AK, EM, DS, FR, LV, FZ
special guests: G. Bellodi, R. Gluckstern, G. Rumolo
excused: WH
------------------------------------------------------------------------
(1) comments on the minutes from the last meeting
------------------------------------------------------------------------
B. Zotter and A. Koschik clarified some remarks concerning Gerry Dugan's
paper. AK could indeed reproduce the simulation results by G. Dugan for
CESR, where a single bunch can oscillate at large amplitudes without
coupling to other bunches in the presence of a tune shift with
amplitude. This decoupling, however, is a simple consequence of the fact
that
the bunch oscillation frequencies are different. It is not thought to be a
manifestation of a soliton, as suggested in the paper. For the LHC,
with only 9 bunches, neither multi-bunch modes not (pseudo-)solitons
were seen in the simulation.
ACTION AK -> continue simulation studies for LHC
(2) luminosity scaling with collimator aperture (FR)
-------------------------------------------------------------------------
Two effects contribute to the aperture in the final triplet, the beam
size and the orbit offset from the center, which are related to the beta
function at the IP and to the crossing angle, respectively. If the
normalized separation at the parasitic collision points is kept
constant, they scale in the same way. The aperture at the triplet, in
units of sigma, must be 1.4 times the aperture at the primary
collimators. Combining the various factor Francesco Ruggiero found that
an increase in the collimator aperture from 6 to 10 sigma results in a
luminosity loss by a factor 1.8. In other words, a realistic beta* would
be more like 1 m instead of 0.5 m, for the present triplet. The argument
above contained some consideration of beta-beating and residual spurious
dispersion, whose changes largely cancel each other.
(3) impedance challenge by Fritz Caspers (FR)
-------------------------------------------------------------------------
Fritz suggested that a proper assessment of the collimator impedance
should proceed in 3D and not in 2D. He expects that the low-frequency
impedance of a collimator jaw will be small if the jaw is electrically
isolated from the rest of the vacuum chamber. An example of a similar
system with small impedance is a lambda/4 strip line. In addition, to
suppress the impedance at higher frequency Fritz conceived the idea of
using rotated or skewed jaws (e.g., X-shaped with about 5 degree angle).
FR asked the LCE team, why, if Fritz is right, this scheme is not used
in any storage ring to reduce the resistive impedance, or, for example,
in light sources, to reduce the impedance of insertion devices. His next
question is whether such a scheme could create additional problems
(longitudinal impedance, resonances?), and what conclusions can be drawn
from existing measurements.
ACTION LV -> estimate the impedance around different revolution
harmonics, e.g., at 20 MHz and higher, and confirm suspicion that the
reduction of the overall impedance by the electrical isolation is small,
even if the impedance is reduced at 0 frequency.
ACTION EM -> learn HFSS and run it to check Fritz' idea.
(4) analytical approaches (R. Gluckstern)
-------------------------------------------------------------------------
After some historical remarks on N. Christofilos, a Greek elevator
engineer who invented strong focusing two years before Courant,
Livingston, and Snyder, flooded BNL with (too) many interesting ideas,
and was later "forced" to learn some maths to check his own ideas (see
http://www.bnl.gov/bnlweb/history/Blewett_talk.htm), Bob presented some
analytical calculations for the collimator impedance problem.
He considers a somewhat simplified model, where a gap with a resistive
medium is embedded in an otherwise perfectly conducting pipe.
He described the various steps of the impedance calculation, for example
the distinction between pipe kernel and kernel of the resistive medium,
the matching of boundary conditions for the magnetic fields of even and
odd modes, and the resulting integral equation. He first computed the
longitudinal impedance, and outlined the analogous steps for the
transverse case.
FR pointed out that terms propagating at much less than the speed of
light may be important, and he also asked whether the use of impedance
boundary conditions for E/H could simplify the problem.
ACTION RG -> continue calculations for transverse impedance
(possibly during the vacation period!)
(5) frequency maps for electron-cloud simulations (FZ)
------------------------------------------------------------------------
Frank Zimmermann presented two frequency maps computed by Yannis
Papaphilippou (ESRF) for 1000 particle trajectories obtained in
electron-cloud HEADTAIL simulations by E. Benedetto. A large tune spread
was found extending over several low-order resonance lines, but not
revealing any influence by these resonances. Yannis had pointed out that
the frequency maps for this example are not very accurate, because of
the presence of synchrotron motion and since the trajectories are the
result of a multi-particle simulation, not governed by a weak-strong
Hamiltonian for which KAM tori are known to exist. Nevertheless it surprises
that any resonance structure is conspicuously absent. DS interpreted this
result as a confirmation that the emittance blow-up is caused by a coherent
phenomenon, and not by incoherent single-particle motion. JJ suggested
to look at the wider frequency spectrum following his earlier work with
R. Burgess.
(6) emittance growth due to quadrupole wakes (FZ)
------------------------------------------------------------------------
Responding to a question by AK, Frank first reviewed the definition of
quadrupole wake fields. A. Chao and R. Cooper defined quadrupole wakes
as wakes driven by a (normal or skew) quadrupole moment of the beam. In
addition, the wake force on a test particle resembles the force by a
quadrupole magnet with a linear dependence on the offset of the test
particle. By extension, later any wake force which is linear in the
offset of the test particle has been called a quadrupole wake. If the
surrounding structure is not cylindrically symmetric, the beam does not
need to contain a quadrupole moment to drive such a wake. In particular
such generalized quadrupole wakes exist for the geometry of a flat
collimator. The term quadrupole wake in this second sense was used by J.
Irwin (NLC ZDR, 1996) and S. Heifets, A. Wagner, and B. Zotter (SLAC-AP
note, 1998).
Using the resistive-wall wake field for a flat parallel-plate
collimator, computed by A. Piwinski (DESY HERA 1992) or K. Yokoya (Part.
Acc. 1993), one can compute the deflection of a particle at longitudinal
position z
and transverse position y in the bunch. Computing the projected
emittance before and after the kick by calculating the second moments y^2,
yy',
y'^2 before and after the collimator, Frank estimated the growth in
projected emittance for a single traversal through the collimator.
It scales with the fourth power of the beam size and the inverse sixth
power of the aperture. Insering numbers for a single 1-m long Al
collimator jaw with 1 mm half gap, Frank estimated an emittance doubling
time of 1 billion turns for the LHC at 7 TeV. This estimate assumes that
the emittance growth is additive from turn to turn, which may not fully
be the case, due to synchrotron motion. A simulation study by E.
Benedetto is underway to explore the accumulated growth over many turns.
(7) report from SNS ASAC and KEKB (FZ)
------------------------------------------------------------------------
The SNS Accelerator Systems Advisory Commmittee meeting was held March
10-12. Highlights included the numerous counter-measures and remedies
taken against the electron cloud (TiN coating of all components,
solenoids at the collimators, electric fields and electron collector at
stripping foils, electron-catcher viewing system), actions to stabilize
the beam against any potentially occurring electron-cloud instability
(such as enhanced momentum spread, chromaticity, and wide-band
feedback), and the development of the ORBIT code, which is rapidily
expanding, and will be the work horse of the SNS commissioning.
The program ORBIT was written by a collaboration involving ORNL, FNAL,
BNL, IUCF, TRIUMF, etc. The code combines all or most features of MAD
with an accurate representation of transverse and longitudinal impedance
and space charge in 3D. Comparison of simulated transverse and
longitudinal beam profiles with measurements at the Los Alamos PSR
showed an impressive agreement. The modelling of scattering in a
stripping foil has been greatly improved, and the resulting tail
distribution has changed compared with earlier models. In the future, it
is planned to include an electron cloud module in ORBIT. FZ suggested to
install this code at CERN.
Recent electron-cloud observations at KEK include a substantial change in
the measured blow-up threshold from one day to the next due to unknown
origin, a non-saturation of the threshold increase for maximm available
solenoid strengths (about 50 G), which is different from earlier
findings when only part of the ring was covered by solenoids, and
saturation (of the lower threshold) was observed below 30 G. BPM and PMT
data showing that above the blow-up threshold individual bunches of the long
train
suddenly become unstable in an apparent random fashion. The beam size
of individual bunches spontaneously increases and damp down to the
original level within about 1500 turns (compared with a transverse
damping time of 9000 turns). Simultaneously, above the threshold the
tune line, seen on both BPMs and PMT, splits into two, which separate
for increasing beam current.
Presentations by WH (Crossing angle at IP2) and EM (Highlights from
Oxford workshop) will take after Easter, at the next LCE team meeting of
25 April.
Slides from presentations by Francesco and Frank Z. are attached.