-----------------------------------------------------------------------
Minutes of the ABP-RLC section meeting of 01.04.05
present: RA, EB, AG, WH, EM, FR, DS, FZ
web site: http://ab-abp-rlc.web.cern.ch/ab%2Dabp%2Drlc/
-----------------------------------------------------------------------
(1) Minutes of last meeting and pending actions
-----------------------------------------------
DS had sent a comment on the minutes, concerning ECLOUD benchmarking,
which was included in the minutes of last week's meeting.
=> ACTION => Maintaining IP knobs of W. Wittmer, check with optics team
and with W. Wittmer (WH, FZ)
STATUS: DONE (WH). LOC will take over knobs prepared by WH and WW.
=> ACTION => compare the size of the longitudinal geometric wake with
RW longitudinal wake from A. Koschik (FR)
STATUS: PENDING.
=> ACTION => confirm bunch length, intensity, and collimator gaps during
tune-shift measurement (EM, FZ).
STATUS: FINISHED (See presentations below).
=> ACTION => Attempt to derive a general nonlinear theory including
the inductive bypass (EM)
STATUS: PLANNED (?)
(2) Further progress on e-cloud simulations (FZ)
--------------------------------------------------
FZ presented LHC heat-load results for 1 and 2 batches. For the drifts,
the heat-loads are the same, and for the dipole field there is a
small increase between 1 and 2 batches. However, for the quadrupole
field the heat load increases significantly during the second batch,
and the heat load is very noisy with large fluctuations from bunch
to bunch. The solution was obtained with the subroutine 'move',
written by DS, which considers a locally constant magnetic field
and extrapolates from two finite-length time steps to determine the
solution with an infinitesimal time step. The routine 'move' is the
only one able to simulate more than 1 batch in 1 week computing time.
For the first batch, FZ compared the results of 'move' with those
obtained from 'odeint2', which is a routine also implemented
by Daniel, based on the Bulirsh-Stoer method. A significant
difference in the build up of the electron line density is visible
during the first batch. The total heat load increases by about 30-40%
(for 8e10) due to the high heat load in the quadrupoles for the
second batch. FZ also showed results for a cartesian wiggler
field in the ILC, where odeint2 and the Runge-Kutta routine
ddqemr from the CERNLIB yield widely different results.
AG suggested to parallelize the ECLOUD code to speed up the
computation time for the other methods. EM pointed out that the
heat load for a 3rd batch may be larger still.
FR recommended to develop a criterion to decide which routine
is correct and to underline the consequent uncertainty on the
average LHC heat load at the LTC presentation of next week.
(3) Results of the data analysis for the SPS collimator (EM+FZ)
---------------------------------------------------------------
EM extracted the intensity and bunch-length data with the help
of G. Arduini. The intensity decreased by about 30% during
a coast as a result of the collimator inward-motion. The bunch
length was 2 times longer than previously thought (close to
0.70 ns instead of 0.35 ns). The collimator gaps were provided
by Stefano.
Using the equivalent Vos/Burov-Lebedev/Zotter theories,
for large gaps the measured tune shift agrees with the
prediction, for small gaps it is about a factor 2 smaller.
The error in the theory is estimated as about 50% (this
error reflects, e.g., the uncertainty in the beam
distribution), so that the measurement and theory appear
consistent.
RA pointed out that the collimator gap numbers are based on
the settings and are not taken from the sensor readback.
He also mentioned that the resistivity was only 10 microOhm m
instead of 14 microOhm/m, which will help to further improve
the agreement.
FZ remarked that the SPS experiment was not sensitive
to the inductive bypass, and that an experiment with
many bunches may be more sensitive. FR suggested to use
Schottky monitors in next years experiment, to detect
the incoherent tune spread due to the nonlinear wake.
FZ also reviewed the experimental conditions. He noted
a small mismatch between the electronic logbook and
the data logging times of Marek Gasior. For time of the 3rd point,
the logbook indicates that only a single jaw was moved.
FZ obtained the emittances from Federico Roncarolo.
They were much smaller than previously thought
(namly only about 1 micron normalized close to injection)
and then changed signficantly during a store, by up to
a factor 2.5. FZ retrieved the intensities from EM's
and GA's excel file and the bunch lengths using
GA's notebook. Small steps in the bunch length correspond
to trips of an rf transmitter.
Based on this information, FZ computed the tune shift
predicted from the Burov-Lebedev formula and from the
nonlinear Piwinski theory, including the expected errors.
In case of Piwinski, emittances, bunch length, and intensity
contribute to the error by comparable amounts. In case
of the Burov-Lebedev theory, it appears the intensity
error dominates.
The final comparison between measured and predicted
tune shifts shows a slope of 1.27 for the Piwinski case
and 0.74 for the Burov-Lebedev case. FZ concluded that
the true tune shift corresponds almost exactly to the
average of the two predictions.
(4) LHC collimator impedance with epsilonr = 5 (EM)
---------------------------------------------------
EM repeated the previous calculation for the LHC collimator
impedance for epsilonr=5, instead of 1, for graphite,
and found no discernible effect on the impedance.
The epsilonr=5 was the maximum value found by AG and
reported in the previous RLC meeting.
(5) SPS MKE kicker impedance with new term of A. Burov (EM)
------------------------------------------------------------
Introducing the new term of A. Burov changes the collimator
impedance above 1 THz only. The effect is not the same
as the effect of ac conductivity. When inserting an
ac conductivity in the Zotter formula the impedance does
not change. For the kicker, the new term of A. Burov has
a small but visible effect on the imaginary impedance,
but it cannot explain the difference between the Burov
and Zotter results.
It was recommended to check again, why the introduction
of an ac conductivity has no effect (and possibly the same
for epsilonr). ==> ACTION EM
The possible improvement by coating was discussed. EM
and FR discovered that at the collimator review a coating
thickness of 1 micron was assumed, while in the more
recent figures from EM, the thickness is 5 micron.
The 5-micron Cu promises a substantial gain in
performance. However, RA recalled that 5-micron
is excluded by heating considerations.
(6) Reports from other meetings
-------------------------------
EM, FR and FZ met with a delegation of the BDI group
(Claude Fischer and company) to discuss the impedance
of different LHC instruments. Among the TV screen
profile monitors two special devices exist at the TDI
and near the injection septa. It was discussed
how the impedance of the 'hublots' could be reduced.
Another interesting impedance problem is posed by
the ionization profile monitor, which consists of
two floating high-voltage metallic plates that will
'cut' the electromagnetic field accompanying the beam.
A prototype was tested in the SPS. Heat load in
the LHC would be higher.
A meeting on the LHC upgrade was held by P. McIntyre,
a professor at the Texas A&M university, FR and FZ,
discussing new s.c. magnet technology, magnet designs,
IR layouts, photon stops, e-cloud countermeasures, etc.
Preliminary designs exist for two strong quadrupoles,
which could be moved inward closer to the IP.
PM recommended that the possible gain in minimum
beta-star be explored until April 4.
Posted on the web: Slides by EM (3X), FZ (2X)
Web site: http://ab-abp-rlc.web.cern.ch/ab%2Dabp%2Drlc/