minutes ABP-LCE meeting 04.04.03
present: Td'A, EB, WH (partly), JJ, MK, AK, BM (partly),
GP, DS, FR, LV, FZ
special guests: R. Gluckstern, B. Zotter
-------------------------------------------------------
(1) new action item and follow-up on old ones (FR)
--------------------------------------------------
ACTION FR -> change time of Joint ABP/CLIC Meeting
to avoid overlap (with Hans Braun)
ACTION FR,LV -> computation for higher-order head-tail
modes could not yet been done since the
impedance model was missing
ACTION FR -> create web page for LCE, pending
(2) impedance of LHC collimators (LV+DS)
----------------------------------------
impedance estimates and limits from beam stability considerations (LV)
LV reviewed the transverse impedance in the LHC without collimators,
for the original collimator design, and for the new collimator design.
He quoted impedances at 8 kHz and 20 MHz. With graphite collimators
the impedance at top energy exceeds 1000 MOhm/m. An important point
is that this happens at both high and low frequencies, which will
make it much harder to combat the incurring instabilities by a feedback.
If the feedback gain is increased to cope with the growth rate at
high frequencies the beam may become unstable at the low frequencies.
LV then showed the stability diagram due to Landau damping by head-on
and/or long-range beam-beam collisions. The long-range collsions stabilize
the beam for impedance below about 100 MOhm/m, the head-on collisions
up to 1800 Mohm/m. This is sufficient to suppress the effect of the
collimator impedance. The head-on collisions at only one interaction
point only provide stability through 450 MOhm/m, and the stability
for the ultimate LHC (2 IPs) is about 1000 Mohm/m.
The most critical point during an LHC cycle is the time between the
closure of the collimators and the beta squeeze until the beams are
brought into collision. A possible solution promising stability would be
to bring the beams into collision prior to closing the collimators and
squeeze with colliding beams. The tune spread by octupoles falls far
short of that required for beam stability. If instead of beam-beam
collisions a feedback is used to stabilize the beam during the squeeze,
the emittance growth time depends on the resolution of the feedback
pick up and for a 1-micron resolution it might approach an acceptable
value. FR suggested that higher-order instability modes should also be
looked at.
DS reported on discussions with I. Syrachev on numerical wake field
computations. MAFIA is not suitable. Microwave studio may be able
to do this type of calculation in 0.5 years time. The code GDFIDL can
do the computation now, and will be available on a CERN computer in
about 6 weeks from now. Warner Bruns, the author of the code, will
visit CERN in this time frame. DS and possible EM will pursue the
numerical calculations.
FZ suggested that the dielectric constant or permittivity of some
of the materials considered may be frequency-dependent.
(3) localized multi-bunch modes (AK)
------------------------------------
AK reported on a paper by G. Dugan published in PRST-AB 1999
on the possibilty of soliton-like solutions, where only a single
bunch is oscillating. Ingredients are a smooth approximation of
the lattice, a multi-bunch wake field and a nonlinear tune shift
with ampitude. FR remarked that GD's symmetrizing matrix should
depend on the wake field. Solving the system of nonlinear
equations using an iterative procedure both Gerry and AK
found solutions where a single bunch could be excited
for the time of the simulation, while all other bunches remain
quiet. The simulation for the LHC is more diffcult. A simplified
study using only 9 bunches did not show a large effect.
FZ suggested to apply the calculation to the Tevatron which observes
soliton-like excitations both longitudinally and transversely.
(4) strategy for collimator impedance (FR)
------------------------------------------
FR stressed the need to develop a coherent & integrated picture
and docmentation, incuding long-range coupled-bunch dipole
wake fields, short-range quadrupole wake fields, the effect
of the inductive bypass, etc. More than one person should
work on this problem, and the final result should be a clear
and precise recommendation to the collimation project.
Consistency of Measurements and simulations should be
demonsrated, and the information extracted from each measurement
clarified. New measurements should be proposed either at CERN
or SLAC if needed. Questions to be answered include scaling with
aperture, the impedance with copper stripes on the two sides of
a graphite substrate, revision of Fritz Caspers' bypass measurements,
effect of finite collimator length; is Yokoya's theory
applicable.
ACTION FZ -> repeat analytical estimate for emittance
growth from collimator wake
WH's presentation on crossing angle in IP2 was postponed to next week
due to lack of time.
(5) preliminary results on emittance growth with electron cloud (EB)
--------------------------------------------------------------------
EB showed the simulated emittance growth due to an electron cloud
of density 6e11/m^3 at injection into the LHC, for 1,3 and 5
beam-electron interaction points per turn. The growth is unacceptably
large in all cases. For 1 IP the growth looks incoherent, while for
3 and 5 IPs the growth is dominated by coherent motion. The emittance
growth varies somewhat with working point. JJ pointed out that
the effect of a localized impedances on the TMCI instability was studied
in the past, and that there can be a strong tune dependence. He also
mentioned a discussion of this effect in A. Chao's textbook.
The emittance growth is not due to a badly performing random generator.
To test this, EB implemented a different better generator provided
by DS, with similar results.
Zeroing the centroid position of all bunch slices at each IP the
emittance is suppressed for 5 IPs, but not for 1 and 3 IPs. In the case
of 5 IPs the emittance is even shrinking which is due to an artificial
'stochastic cooling' effect. Alternatively, EB removed the coherent
motion by 'symmetrizing' the electron and proton macro-particle
distributions. Also in this case the emittance growth is cmpletely
eliminated in the case of 5 IPs. This does not mean that the emittance
growth does not exist in reality, but it shows that a coherent motion
is involved. The HEADTAIL simulations were cross-checked against
simulations by K. Ohmi (KEK). His results are comparable. K. Ohmi's
numbers are slightly more pessmistic, which could be due to a
different size of the grid and/or cloud.
FZ's presentation on e-cloud at SNS and KEKB was postponed
to next week due to lack of time.