-----------------------------------------------------------------------
Minutes of the ABP-RLC section meeting of 14.01.05
present: WH, JJ (partly), EM, TP, FR, DS, EV, FZ
excused: EB
-----------------------------------------------------------------------
(0) New organization (FR)
-------------------------
FR announced the new name of the section: R&D and LHC Collective Effects (RLC).
The R&D activities are new. They include CLIC and LHC upgrade, but not
the SPL. Roberto Corsini, Hans Braun, and Helmut Burkhardt (50%) have joined
the section, and two new posts will be filled during the year.
In addition, the team will comprise several new students and fellows.
Meetings will be held separately for LHC & LHC upgrade on the one hand
(continuation of LCE meetings) and for CLIC studies on the other hand
(Wednesday beam-dynamics meetings organized by D. Schulte).
(1) Minutes from previous meetings and pending actions
------------------------------------------------------
The last minutes did not mention that an AB report by EV
on 'simulation of transient effects in beam transverse-feedback
interactions with application to the CNGS beam from the SPS'
has been approved by ABP and RF group leaders.
Comments and pending actions also included beam-beam studies with
different crossing schemes, flange impedance, and electron cloud.
They are addressed in the presentations and discussions below.
(2) Beam-beam studies
---------------------
There was a discussion whether the team should write a common
or separate paper(s) describing the results for different crossing
schemes. Studies are still ongoing. There may be only small
differences between the different calculations.
Yannis Papaphilippou had offered to perform frequency-map
analysis for an example case, to help in the benchmarking.
He would need a SIXTRACK or MAD file as input.
ACTION => WH will provide him with a first test file.
Dobrin Kaltchev is working full time on this problem,
performing tune scans with optimization for nominal and
PACMAN bunches, for all options. This topic is to be followed up,
but there is no urgent deadline at this stage.
(3) Numerical calculations of collimator impedance (AG)
-------------------------------------------------------
Gdfidl time-domain simulation shows two longitudinal modes,
at about 0.6 GHz and 1.24 GHz. HFSS calculation for these
two modes then determined detailed parameters. The mode
frequency changes with the size of the gap. A similar effect
was measured by Fritz Caspers in the SPS MD. The measured
change with gap size is somewhat larger than simulated,
but in the same direction.
ACTION => invite Fritz or his colleague to present the
SPS measurement results
The first (lower-frequency) mode appears quite harmless.
The power loss to this mode is 10 W. The second mode at
1.24 GHz is concentrated near the rf fingers, also extending
into the collimator tank. It has a higher Q, and strongly
couples to the beam. The power loss into this mode is
about 280 W.
ACTION => inform collimation project
AG also computed the longitudinal geometric wake for a
very short bunch.
ACTION => compare the size of the longitudinal geometric
wake with RW longitudinal wake from A. Koschik (FR)
AG suggested that the longitudinal wake could be
fed into the Haissinski equation to estimate the
associated bunch lengthening.
ACTION => compute the transverse modes (AG)
(4) Calibration of new ECLOUD simulations with SPS measurements (DS)
--------------------------------------------------------------------
An error in the angular dependence of the secondary emission
yield was discovered by G. Bellodi and has been fixed. The effect
on the LHC heat load predictions is small. However, the agreement
and consistency with SPS measurements is improved. DS compared
simulations for the strip detector with different numbers
of batches and batch spacings with the experimental data to
infer the maximum secondary emission yield and the low-energy
electron reflectivity. Using the ratio in the current measured for
1 and 2 batches, and also comparing the currents observed for
nominal and 2-microsecond spacing, he determined
delta_max=1.35 and R=0.5 as intersection between the two fits.
Remarkably, the total current measured at the detector also
agrees with the value simulated, taking into account all
electrons with 'vertical kinetic energy' above 1 eV. Setting the
'vertical energy' cut at 10 eV instead, the fitted delta_max and
R do not change very much, but the currents no longer agree.
Therefore, the detector appears to measure all electrons.
From the further current decay DS conjectures that the final
value of delta_max was 1.25 after scrubbing, though an independent
determination of delta_max and R by the same method as before
is not possible in this case, since the current measured
for the 1st batch alone decreased below the measurement
resolution.
(5) Flange impedance (FZ)
-------------------------
FZ showed a revised estimate of the LHC flange impedance,
based on a formula of K.Y. Ng. It is different from Alex Chao's
formula by a factor d/g (where d is the depth of the cavity
and g the size of the gap). Using Ng's formula and revising
some other numbers as well, the total longitudinal impedance
of the LHC flanges would exceed half the total budget and
the transverse broadband impedance would amount about 20%
of the total. These numbers are significant. In addition,
there is a possible danger from trapped modes.
The flange impedances were a problem in the PS booster at various
times over a 30-year period, especially when the harmonic number
was lowered to h=1 for the LHC beam. The PS booster flanges are e
quipped with bypasses around the insulating layer of the flange.
These bypasses had to be upgraded to lower the impedance at low
frequencies. Helpful discussions with K-J. Schindl, C. Carli,
M. Chanel over the last week have provided us with many ideas
and with a set of relevant papers. Extrapolation to the LHC is
not straightforward, as the LHC flanges do not contain any
insulating layer.
ACTION => perform numerical simulations of flange to
determine broadband impedance and trapped modes (AG)
ACTION => obtain more information on the flanges from
M. Jimenez, e.g. exact total number, material,
chamber dimensions, details of the drawing,
location of flanges and possible presence
of a beam screen (FZ)
(6) Ecloud model for SPS lifetime (FZ)
--------------------------------------
Stimulated by HHH-2004 a collaboration was started with
Giuliano Franchetti (GSI) to model the combined effect of
space charge, electron cloud and synchrotron oscillation, by
a resonance trapping mechanism which would lead to loss
of off-momentum particles (as observed in the SPS) and to
halo generation. Eventual effects of the linear coupling
resonance and the possibility that the pinched electron
cloud may excite the Montague resonance are also being
explored.
(7) Trip Report PEP-II MAC and LBNL HIF VNL (FZ)
------------------------------------------------
PEP-II integrated luminosity per month was increased
by 40% thanks to 'trickle injection' and 'reduced abort
rate', before all accelerators were stopped in response
to accident. DOE and Stanford investigations and
remedial actions are still ongoing. For PEP-II upgrade,
parasitic collisions and the effect of a crossing angle
are a big concern. One can be traded against the other.
These two effects have been investigated in simulations
and machine experiments. A few 100 microrad crossing angle
leads to a significant luminosity loss, and so does any
reduction in the beam-beam separation at the parasitic
collision points. Powerful PC farm with close to 400 nodes
is used for SLAC beam-beam simulations The farm exceeds
the capacity of the NERSC supercomputer. Up to 250% beta
beating in both rings has not been corrected since several
years (a failed attempt lowered the luminosity). Previous
mysterious results of beam-based alignment could be
understood by including linear coupling in the analysis.
Recommendations include IP tuning knobs and lowering
the vertical emittance, e.g., by dispersion-free
steering.
Afternoon visit to VNL. Two prototype accelerators
exist for high-current ion-beam generation and transport
(HCX) and charge neutralized focusing (NTX). HCX shows
dramatic e-cloud effects. A series of PIC simulation codes
were written. Recent flagship is the WARP code, in which
M. Furman's POSINST algorithms are currently implemented.
In a video conference with LLNL, R. Cohen described
3-D simulation results, including scheme which speeds up
the modeling of electron motion in magnetic fields by
a factor of 25, without losing physics contents. This
scheme bridges the gap between electron and ion time
scales.
FR suggested to explore the possibility of establishing a
collaboration on e-cloud simulations with the VNL team.
Attached: Slides by AG, DS, FZ