From: Francesco Ruggiero
Sent: Friday, March 12, 2004 7:15 PM
To: Bruno Zotter; Daniel Schulte; Elias Metral; Frank Schmidt; Frank
Zimmermann; Gianluigi Arduini; Helmut Burkhardt; Hiroshi Tsutsui; John
Jowett; rlg@physics.umd.edu; Stephane Fartoukh; Werner Herr; Luc Vos;
Maxim Korostelev; Alexander Koschik; Eric D'Amico; Walter Wittmer;
Gregory Penn; lronen@techunix.technion.ac.il; Elena Benedetto; Elmar
Vogel; Javier Resta Lopez; Peder Eliasson
Cc: Fritz Caspers; Jean-Pierre Riunaud; Karlheinz Schindl; Louis
Rinolfi; Michel Martini; Oliver Bruning; Charles Hill
Subject: Minutes of LHC Collective Effects team meeting 12/03/2004
Minutes of the ABP-LCE team meeting on 12.03.04
-------------------------------------------------------------------------
present: EB, WH, AK, EM, FR, DS, EV
FS, FC (part-time)
excused: FZ
-------------------------------------------------------------------------
(1) Follow up on action items
-----------------------------
OLD ACTIONS (mainly DS): concerning LHC collimator impedance
=> clarify geometry of trapped modes -> partly done
=> compute longitudinal impedance -> done
The real longitudinal impedance due to em trapped modes below cut-off
frequency is about 50 mOhm for collimator jaws in "closed" position.
Evaluation of Q-factor for modes near cut-off is non trivial with GdfidL
and DS contacted W. Bruns (author of GdfidL) to include losses at the edge
of the mesh. To better distinguish between wave-guide and trapped modes a
time-domain analysis was used, clearly showing that some of the modes do
propagate. These modes could however be reflected by nearby collimators.
=> practical advice for collimator design -> partly done
DS has suggested a modified design of the collimator tank to reduce
its coupling to the beam. The longitudinal impedance of a trapped mode,
localized near RF contacts and pipe walls, is 1 mOhm for m=4.
This mode may be damped by an antenna-like structure (not favoured by
collimator design engineers).
=> redo estimate for larger gap -> done
=> understand stability diagram for m=4
Stability diagrams allowing tune spread in the two transverse planes
or in the longitudinal plane are discussed in LHC-Project-Report-121.
To get a first conservative estimate, it is suggested to use a stability
diagram independent of the longitudinal azimuthal mode number m.
=> what can be measured in the SPS? -> pending ACTION (AK, EM, FR, DS,
FZ)
Typical tune shifts of about 5x10^-5 are expected in the SPS from
geometric collimator impedance for m=4 or m=5. These are very difficult
(impossible?) to measure. Further analysis is needed, also for resistive
wall impedance, in view of a dedicated APC meeting on April 2nd (see
http://ab-div.web.cern.ch/ab-div/Meetings/APC/2004/apc040402/agenda.html)
=> perform full simulation for the SPS (AK) -> pending ACTION
=> compare results and mathematica notebooks (FZ, EM, AK) -> partly done?
=> suggest that F. Caspers may attempt a measurement in a parameter region
that is closer to the LHC collimators (FR) -> done
(see minutes of discussion prepared by EM)
=> check epsilon and mu of collimator material for impedance calculation
(mu of the graphite was checked by EM and it is close to 1)
-> pending ACTION for epsilon (EM)
=> perform comprehensive HEADTAIL simulation study for LHC (EB)
-> partly done (EB will present results next week)
=> FR will investigate whether there exists a special NAP priority queue
or whether priorities are automatic -> done
FR clarified this with FS. Job submission is normal, but high priority
requires users to be in a special list called u_slap.
EB will be included in the list u_slap.
(2) Computer needs and CPU estimates (FS)
-----------------------------------------
FS reported that Lyn Evans agreed to buy 64 new dual processor machines leading to a tenfold increase of our current computing power. The procurement of these machines will take time and they will not be available before August. Meanwhile IT will allocate about 40 machines, but we need to assess specific computing needs/plans over the next 6 months to allow a better coordination of resources. Our system is currently under-used.
ACTION => WH, DS, and FZ will assess specific CPU needs/plans for
LHC beam-beam tracking and e-cloud simulation campaigns.
DS asked whether a parallel code such as GdfidL could be run on the NAP. The answer is probably no. In addition there would be licence problems.
ACTION => FR will talk to J.-P. Delahaye to see if the CLIC parallel cluster can be used (with low priority) also for GdfidL simulations for the LHC, possibly by contributing 1-2 dual processors paid by ABP. In addition he will propose to provide DS with a fast desktop machine for e-cloud and possibly HFSS calculations.
FS mentioned the project currently carried out with Eric McIntosh to run LHC simulations as a "screen saver" on all CERN PCs. Scalability is under study.
(3) Beam-Beam simulations, follow-up of LTC open action, and MAD work
(WH)
--------------------------------------------------------------------------
WH reported that F. Jones will visit CERN in May. Beam-Beam simulations with HFMM can be run on a parallel cluster at TRIUMF.
In an LCC meeting at the end of 2002 it was agreed to perform
- tracking studies with Beam-Beam and linear imperfections,
- tune and phase advance scans with alternating and non-alternating crossings. Some limited work on the first point was done (and may still be done?) by B. Muratori. The second point requires tracking without imperfections to have a reference for non-alternating crossing planes and then extending the studies to include imperfections and IP phase advance scans. A related goal is to establish an updated list of tolerances associated with beam-beam effects, including tolerances on beam separation, crossing angles, beta*, bunch populations, and emittances.
ACTION => WH will organize a meeting to discuss beam-beam tracking studies
WH reported that he is re-writing the MAD orbit correction package to allow a user friendly two-beam orbit correction. This does not require any modification of the data structure, but relies on name extensions for beam 1 and beam 2.
ACTION => WH will discuss MAD name extensions with TR
(4) HEADTAIL simulations of Q' effect on high intensity
single bunch stability at injection in the SPS (EB+EM)
----------------------------------------------------------
SPS observations in 2003 with a special high-density bunch of reduced longitudinal emittance indicated a fast TMCI-like instability. This is expected from a simplified analytic criterion (see enclosed slides by
EM) and confirmed by HEADTAIL simulations performed by EB. Also E. Shaposhnikova had presented MOSES results at the last Chamonix workshop showing this effect. The instability threshold increases linearly with chromaticity and the corresponding slope computed by EM's formula is in excellent agreement with HEADTAIL simulation results for a single broad-band impedance model. However HEADTAIL shows little or no sensitivity to the resonant frequency f_r of the impedance, contrary to EM's analytic expression which is proportional to f_r. Further experimental data will become available this year.
FR suggested to include a more realistic impedance model in HEADTAIL, to be able to make SPS predictions for different impedance sources (e.g. more MKE kicker modules with various degrees of shielding).
(5) New results by BZ on Resistive Wall impedance for collimators (AK)
----------------------------------------------------------------------
AK presented plots (see enclosed slides) for various models of the resistive wall impedance down to very low frequency. For LHC graphite collimators, modelled by a uniform multi-layer resistive wall, the real transverse impedance obtained by BZ by exact mode matching for a ring charge with cos-like azimuthal distribution at ~ 10 kHz is two orders of magnitude larger than predicted by LV or Burov-Lebedev. The latter is also obtained by mode matching, but only the vector potential is used. The imaginary part of the impedance obtained by BZ changes sign around 10 kHz, consistently with the fact that the transverse wake goes to zero at zero distance from the drive (ring) charge. For the test bench measurements by F. Caspers et al, all the real impedance results agree above a few hundred Hz. The change of sign for the imaginary impedance predicted by BZ may be consistent with experimental data so far considered as being in the noise and attributed to a measurement noise effect.
(6) Green light for installation of LHC collimator prototype in the SPS
(DS)
----------------------------------------------------------------------------
DS presented two slides (enclosed) for the LHC Collimation Project meeting scheduled in the afternoon. There is no problem from the point of view of the geometric (and resistive wall) impedance. RF contact resistance will be reviewed separately.
(7) US-LARP collaboration meeting (FR)
--------------------------------------
FR reported about SLAC now contributing LHC collimators for Phase 2. This will have an impact on the US-LARP resources available to other labs and other activities. In particular AP contributions may have to suffer. S. Peggs proposed an ambitious "LHC operational modelling" programme, requiring a tight interface with LHC Controls, Operations, and Beam Physics. This seems difficult to fit with the available resources.
(8) LCE participation to EPAC04 and other conferences
-----------------------------------------------------
FR said that the criteria for participation are:
i) invited talks, ii) young people (especially if they get an EPAC grant),
iii) quality and number of papers/other conferences/interest to go.
Using these criteria, he has proposed FZ, AK, EB, EV and DS (with low
priority) for EPAC participation. FR will attend off-quota as EPAC organizer. EM will attend APAC04 and ICFA-HB2004. DS would prefer to attend LINAC04.
Attachments:
minutes of discussion on Resistive Wall with F. Caspers prepared by EM
--
E-mail: Francesco.Ruggiero@cern.ch, Location: Building 9/1-008
Address: CERN, A&B Department, CH-1211 Geneva 23, Switzerland
Telephone: +41 (22) 767 3726 or 767 5272, TeleFax: +41 (22) 783 0552
WWW: http://wwwslap.cern.ch/~rgo/