Minutes of the ABP-LCE team meeting on 24.10.03
present: EB, WH, JJ, AK, EM, FR, DS, LV, FZ
(1) Outstanding Actions
-----------------------------------
ACTION -> Check sign of real part of impedance (DS).
Perhaps done?
ACTION -> EXCEL spreadsheet for collimator impedance (EM).
Done (see below)
(2) Automatic collimator impedance calculation (EM)
---------------------------------------------------
EM prepared a program which computes various impedance
properties of the LHC collimators, using as input the
excel spreadsheets sent by Ralph Assmann in regular
intervals. There are 4 primary and 16 secondary collimators,
all made of carbon with an assumed resistivity of 14e-6 Ohm m.
Length of the primaries is 0.2 m, length of the secondaries
1 m. Rotation angles and plane (H,V,S) are defined for
each collimator.
ACTION -> Ambiguity in the angle definition needs to be clarified.
FZ asked why some angles differ by odd amounts from
0 or 45 degree, etc., reminiscent of sextupole rotation angles
in the SLC RTL.
For the longitudinal impedance, EM used the formula for
a round pipe. This should be a pessimistic estimate.
ACTION -> check possible longitudinal 'Yokoya factor'
relating round and rectangular cases
For the transverse impedance, EM includes Yokoya factors, Luc's
expression for the inductive bypass, the tensor transformation
by Francesco, and the beta functions at the collimators.
He sums over all collimators to compute and plot the total
real and imaginary impedance, as well as the ratios between
the two planes. The standard collimator version 6.4 has a 20%
larger impedance in the vertical direction, while the so-called
'option 1' promises an equal and 2-3 times smaller impedance
in both planes. EM's program also gives the coherent tune
shift on the stability diagram for the most unstable coupled
bunch mode. The nominal LHC beam is unstable for both collimator
versions, but much less so for option 1. Some discussion
ensued as to whether the tune shift was consistent with
Luc's earlier calculation. EM's program at the moment does not
yet handle the case of coating.
A further output of the program is the resistive heating of
the collimators, which is of the order of 140-200 W/m for a
half gap of about 2 mm.
The power scales as the inverse of the gap size.
ACTION: put numbers into ZBASE (EM)
JJ pointed out that a conversion program is available from
EXCEL to mathematica format.
(3) TMCI threshold in the SPS (EM)
----------------------------------
A few weeks ago an instability was observed when a low-emittance
beam of 1.2e11 protons was injected into the SPS. The beam loss
occurred after a few ms, and was suppressed by a high chromaticity
of 1 or 2 units and also by increasing the longitudinal emittance.
This instability could be either a head-tail or a TMC instability.
EM presented his general framework for TMCI instabilities, and
compared it with earlier MOSES calculation by Elena Shaposhnikova,
for leptons at 3.5 GeV. His simplified formula gives nearly the
same minimum threshold as Elena's calculation. The dependence on
the bunch length is qualitatively the same (quadratic), but there
is a factor of 3 difference, which is tentatively attributed
to the difference between Gaussian and parabolic bunches. It is
remarkable that the simplified formula extends over a huge
range of the MODES calculation, over which various different mode
mergers define the threshold.
The formalism can be applied to the e-cloud instability,
where EM found that shorter bunches are more stable. Also the
longitudinal TMC instability can be treated in the same manner,
after including the potential well distortion from space
charge and wake fields. Linear coupling can increase the threshold,
but so far no experiment was done to demonstrate this for
the e-cloud instability in the SPS.
(4) ZBASE GdfidL work (DS)
--------------------------
DS computed the res. wall impedance for a 10 cm long
carbon collimator. He plotted the ratio of the computed
real and imaginary impedances divided by the Piwinski
prediction and found a difference by at least an order
of magnitude (Gdfidl giving the smaller number), which
increased for larger frequencies. After contacting
Warner Bruns the conclusion is that GdfidL cannot be
used to compute the impedance for carbon, since it is
still too good a conductor (the ratio of epsilon/kappa
determines the time step required). A factor 1000 times
shorter time step would be required, which is not
practical. Presumably HFSS has to be used instead.
As a further check, DS is running a GdfidL job with
much increased resistivity.
DS also simulated the impedance and wake from a smooth
tapered step, and got a reasonable result, except for
long-term growth which appears to be a numerical artifact,
since it decreases for a finer mesh.
He checked the impedance for an unshielded LHC bellows,
comparing the ZBASE data from ABCI with new GdfidL results.
The wakes look quite different. Nevertheless the impedances
computed by the two codes seem similar.
FR encourages DS and other young ZBASE users to check
the element description with real world.
There was no urgent desire to re-distribute the workload
for the ZBASE review and update.
(5) LHC Ions (JJ)
-----------------
A substantial revision of parameters has taken place.
No longer are there 200-MHz capture cavities. The longitudinal
emittance was reduced accordingly. The emittance is later
blown up to reduce transverse IBS growth rates. At top energy
the IBS growth rates are comparable to the radiation damping.
Therefore, no emittance blow up is expected. The radiation
damping time for ions is half that of the protons. JJ is
studying the possibility of changing the partition number
to shrink the horizontal emittance. There is no crossing
angle, and the geometric luminosity loss factor is 1.
JJ showed beautiful 3-D animation of beam sizes and apertures
around the IP, which can be created with a mathematica
package.
(6) Update on e-cloud simulations (EB)
--------------------------------------
EB compared simulations of emittance growth due to
electron cloud with and without conducting boundary
condition. The latter reduces the growth by up to
50%. It also affects the centroid motion and the
head-tail oscillations along the bunch. If the
emittance growth for low electron densities is real
or not is still an open question, but a benchmarking
study is underway in collaboration with Ali Ghalam
at USC. Restricting the QUICKPIC code to 1, 2 and 4
interaction points (IPs), instead of the usual quasi-
continuous description, the emittance growth in QUICKPIC
is seen to depend strongly on the number of IPs.
This looks similar to the HEADTAIL behavior. Detailed
comparisons will be performed between the two codes.
In particular the low-density region is of interest.
A study of the geometric collimator wake field
using HEADTAIL is also in progress.
(7) Report from PEP-II MAC and Factories'03 (FZ)
------------------------------------------------
PEP-II and KEKB pursue upgrade programs which
in the short-term should increase the luminosity
by a factor of order 4 and in the longer term
should push the luminosity to 1e36. The recipe
is more bunches, higher bunch current, higher
beam-beam tune shift, and lower beta-star (plus
shorter bunches).
DAFNE considers strong rf focusing as approach
for the future.
Electron cloud observations at KEKB suggest that
the solenoids have no effect for 4-ns bunch
spacing, unlike for 6-ns and especially 8-ns
spacing.
There is a concerted effort to benchmark
strong-strong beam-beam codes, involving
5 codes SLAC, LBNL, KEKB and Cornell. Many
of the codes were also benchmarked against
one or several colliders. FZ prepared a
summary table in response to a suggestion
by D. Rice. K. Ohmi showed that the beam-beam
limit in the strong-strong simulation arises from
an incoherent effect, and that it can be reproduced
in a much faster 'quasi-strong' simulation.
Upgrades for the PEP-II multi-bunch feedback
system work with a time resolution of 0.5 ns,
and can damp the m=1 head-tail mode in addition
to the rigid bunch motion.
(8) Report on other ongoing e-cloud activities
and ZBASE (FZ)
-----------------------------------------------
G. Bellodi performed a very detailed comparison
between ECLOUD and POSINST parametrizations of
secondary emission. The simulated line densities
is almost a factor 10 different between the two
models. POSINST includes re-diffused electrons,
and the energy distribution of the true secondaries
is markedly different.
The ELCOUD'04 workshop is scheduled for April 2004,
and CERN is one of its sponsors, contributing
to the web page and to the poster design.
The progress on the ZBASE assessment is slow.
Two persons promised to send MAFIA files and
data for the cavities (D. Angal at ASTEC), for the
Y chamber (D. Li at LBNL). W. Hofle sent information
on the transverse damper, and asked for ABP/LCE
resources to update the MAFIA computations.
B. Spataro lost all his LHC MAFIA files due to
a computer crash in Frascati.
Attached: slides by EM, EB, FZ.