New meetings are in the OMC indico
BPM aberrations, June 2014
Jaime Coello, Lars Jensen, Andy Langner, Thibaut Lefevre, Meghan J. McAteer, Rogelio Tomas and Robert Westenberger.
Presentations
[Robert]Strong (1,2) line in horizontal BPM spectrum of 2012 kick data
[Lars] LHC BPM data and LS1 objectives
Summary of actions
From Robert's presentation and assuming that the used software
has no bugs the BPM aberrations cannot explain the observed spectral line (1,2).
Thibaut suggests to look at tune monitor data and damper BPMs (to check
if data is still in the database and to consider this for future MDs).
Question from Rogelio: can beam size have an impact on BPM response? Thibaut answered that
the impact of beam size should be well below the impact of centroid offset.
Lars' presented the 3 corrections applied to BPM raw data. They are applied in the following order:
1st ambient temperature, 2nd electronics (gain and offset) and 3rd geometry polynomials depending on BPM type.
This order allows to invert the third step and apply new more precise polynomials
for 2012 data.
What to do in 2015?
The consensus is that the effect of BPM aberrations remains small and the
the situation will be better in 2015 with controlled temperature and the new more precise polynomials.
The proposal is to test simultaneously with beam the orbit and the turn-by-turn BPM data to verify that
implementations are correct. This should be done early during the commissioning in 2015.
Concerning the largest amount of BPM data that can be acquired a dry test should be carried out in 2014.
This specially important to be able to acquire all the flat plateau AC dipole data, which will be extended in 2015.
Non-linear studies, April 2014
Meghan Jill McAteer, Ewen Hamish Maclean, Tobias Hakan Bjorn Persson, Jaime Coello, Yngve Inntjore Levinsen, Rogelio Tomas, Piotr Skowronski, Robert Westenberger.
Presentations
[Jaime]2-beam correction in the GUI
[Robert]Resonance Driving Terms
[Meghan]PSB resonance compensation measuremets
[Yngve]Non-linear Coupling
Summary of actions
GUI actions: Make the code run for all IRs, including the IRs where beams do not share beam pipe (IR3, IR4, IR6, IR7). Display MQX magnets on the charts
assigning kqx to MQX.3.
RDTs actions: Analyze and compare more experimental files with the help of Ewen. Observe the issue with the frequency shifts of the non-linear lines with respect to the plain nQx+mQy computation. Compute the complex spectral lines in GetLLM with the help of Tobias. Follow up the modification of Drive with Jaime.
Meghan: Very nice introduction to RDTs, some typos to be fixed (done).
Yngve: Promissing results but further debugging and verification is needed maybe with a toy model. f1001 and C- from measurements. Other sources: sext+ skew sexts (or copuling) and skew oct. First action: extended slides with more cases and clearer descriptions.
Plans for 2014 January 22
Meghan Jill McAteer, Ewen Hamish Maclean, Tobias Hakan Bjorn Persson, Yngve Inntjore Levinsen, Rogelio Tomas, Andy, Piotr, Viktor.
Presentations
[Tobias]Coupling measurements improvements
Summary
Progress during summer 2013 (August 14)
Alexander Charles Sherman, Meghan Jill McAteer, Ewen Hamish Maclean, Tobias Hakan Bjorn Persson, Yngve Inntjore Levinsen, Simon White,
Rogelio Tomas
Presentations
[Alex]Improvements to Drive and action measurement
[Viktor]News on sfotware development
[Tobias]Optimizing coupling measurement
[Meghan]PSB turn-by-turn measurements
Summary
Optics reconstruction from AC dipole measurements in the energy ramp
Maria kuhn, Andy Langer, Meghan McAteer, Ewen Maclean, and Rogelio Tomas, Georges Trad (CERN)
Presentations
[Ewen]Beam 2
[Andy]Beam 1
[Maria]Emittance measurements
Summary
Ewen and Andy presented the details of the beta-funcion computations from the last MD data
during the energy ramp. This allowed to spot a single inconsistent measurement in B2H that seemed
to bias the reconstructed emittance. It is clear that the measurement resolution is
not suficient to judge on emittance variations during the ramp. The reslution is first limited
by being able to perform a single measurement per energy.
A source of systematic error (1-6%) is the recent finding of MQYs having a wrong calibrations.
The algorithm could also be improved
to pair more and/or better chosen BPMs. The AC dipole will be improved during LS1 to acquire
more turns per acquisition. This forecasts an improved resolution measurement after LS1.
Other measurement techniques, like K-modulation, are discarded during the ramp.
Meeting on MD3 Non-linear exp with Masamitsu
Masamitsu Aiba, Yngve Levinsen, Ewen Maclean, Tobias Persson, Simon White, and Rogelio Tomas (CERN)
Presentations
[Meghan]Normal and skew sextupole polarity checks at injection
[Tobias]Chromatic coupling and smplitude detuning from BBQ
[Yngve1][Yngve2]Chromatic coupling, comparing to model predicted corrections
[Simon]Amplitude detuning from BPM data
[Ewen]Crossing angle scans
Beam 1 KSS corrections MODEL Vs Exp correction:
MODEL EXP
-0.052 -0.007
-0.033 0.009
-0.032 0.003
-0.031 -0.005
-0.002 -0.0003
-0.006 -0.008
-0.035 -0.006
-0.015 0.00000
Beam 2 KSS corrections MODEL Vs Exp correction:
MODEL EXP
0.054 -0.017
0.036 0.022
0.031 -0.014
0.023 -0.001
0.003 -0.010
0.002 -0.012
0.033 -0.001
0.018 0.023
BI K-modulation measurements, May 18, 2012
Nicolo Biancacci, Yngve Levinsen, Ewen Maclean, Tobias Persson, Federico Roncarolo, Georges Trad, Glenn Vanbavinckhove, Aurelie Rabiller, Piotr Skowronski and Rogelio Tomas (CERN)
Presentations
[Georges]K-modulation measurements in IR4
[Ewen]Comparison to BPM measurements
Summary
Georges presented the k-modulation measurements and first analysis in IR4 to determine the beta functions
at the wire scanner and at the undulator locations. Most of the measurements were very clean.
The largest observed beta-beating was about 8%. This is in good agreement with the peak beta-beating
of 10% measured at the BPMs. Two improvements of the analysis would consist in using the exact formula
for the tune shift and computing the average beta at the quadrupole.
The extrapolated beta function at the wire is in agreement with previous BPM measurements.
An improvement of this extrapolation would consist in either using MADX to match the observations
or use at least two measured betas together with the transfer matrices to avoid any assumption in the alfa functions.
Ewen showed a zoom of the IR4 measured betas at the BPMs together with Georges' values. The different measurements
are in agreement. The BPMs feature a considerable error bar due to the phase advance either close to 0 or 90 degrees.
This could be improved in the future by choosing sets of BPMs with optimum phase advances to estimate the
betas using the LEP method.
OMC activities review, April 25, 2012
Thomas Bach, Rama Calaga, Yngve Levinsen, Ewen Maclean, Tobias Persson, Glenn Vanbavinckhove, Piotr Skowronski, Marek Strzelczyk and Rogelio Tomas (CERN)
Presentations
[Tobias]LHC coupling feedback
[Thomas]JAVA GUI
[Yngve]Chromatic functions and plotting library
[Ewen]LHC non-linear beam dynamics measurements
Summary
#############################
Coupling feedback - Tobias
#############################
Rama is surprised by the lack of sync of the BBQ BPMs.
BBQ BPM data sync algorithms are only robust right at injection.
Plans & How to proceed:
-Sync at every injection
-Use injection data for all BPMs
-Need a trigger for injection (Verena already replied)
-Save automatically all data (first analyse manually)
-Fully automatized analysis of injection osc ~ 1month
-Implement the buffer so the BBQ incoming data is always sync ~ 1month (in parallel)
-Investigate other BPMs, ADT are good candidates
-Fully automatized on-line measurement -> end of summer
-"correct button"
comment, ro: move "correct button" to earlier, summer?
######################
Java GUI-Thomas
######################
Ram, unit tests should include python codes plus different platforms.
Thomas comments that this will come later.
Glenn agrees that removing loading screen is OK if 30s are saved.
Memory problem (heap space) identified, now one can load more than 50 files.
Plans:
-finish refactoring
-finish unit tests
-write integration test plan
-testing framework for external python scripts
-New functionalities and performance improvements
-Estimated time 1-2 months
Comments: Have 2 separate set of tests, fast & deployment tests.
Glenn: Model creator , alternatives: JMAD or extract strengths from LSA. Marek's opinion: "it is doable...".
Bottleneck is file loading (20-25s), Ram just made a new version of svd_clean reducing the number of svd calculations and possibly multithreading, to be tested (rogelio will put a test case).
Glenn: fix phSVD.py with the tune window problem!
#######################################
Chromatic functions calculation - Yngve
#######################################
New version of W functions calculation finished by Yngve.
Plans:
-Remove required input which is available somewhere else
-Plug it into the GUI
-Fix syntax, conventions: use Python macro, import * not good, convenient to
mark the main, tabs and spaces mixing is not good...
-Documentation for the plotting library is not available in the CCC... need some
way to make it accessible.
Ro: Maybe June (next MD) is a good ref
Glenn: dp/p is only written to the file when GetLLM is run by the GUI. This should
be carefuly taken care of when implementing getsuper.
######################################
Non-linear stuff - Ewen
######################################
Plan:
-Examine measured data this week -> comparison to model
-Examine issue of CO leak from IR bump
-Corrections?
-Analyze high kick data for fterms
-Update GetLLM to compute more fterms
-SOLEIL, need feedback from Laurent...
...
#####################################
Automatic matching - Piotr
#####################################
Online demonstration
The measured phase-advance can be excellently reproduced with large strs.
Two beams at the same time does not work yet.
This tool would be very useful for different applications as optics
correction and effective model generation. A suitable time to have this
ready would be summer 2012.
#####################################
K-modulation application - Marek
#####################################
K-modulation application has been already used in CCC.
Some work needs to be done to automatically output beta* results.
Marek needs more testing that should come during a dedicated MD
and during regular beta* measurements. Tobias Baer found an intrinsic
problem when trying to modulate some main quads like Q4 but this was known.
A good target for completion of the code is middle or end summer.
ATF DR optics measurements, October 22, 2010
Eliana Gianfelice (FNAL), Manfred Wendt (FNAL), Glenn Vanbavinckhove (CERN), Yves Renier (CERN), Yannis Papaphilippou (CERN) and Rogelio Tomas (CERN)
Presentations
[G. Vanbavinckhove]ATF DR optics measurements
Summary
Glenn presented the basic analysis of the ATF DR TBT data provided by Eliana Gianfelice.
The vertical kick is weaker than the horizontal (vertical kicker at a lower beta_y function).
There is some low frequency noise that does not significantly affect
the measurement of the fundamental tunes but could pose some limitations in coupling
and higher order measurements.
A synchronization problem could explain a tune jump along the accelerator in one of the files.
This problem will be looked at by Manfred and his team in December.
One BPM is not connected. 2 BPMs have a wrong calibration factor
(by factor of 2).
These measurements were taken under great preassure.
Ideally Yves, Manfred and Eliana will be able to collaborate this December to get
the CERN codes running in ATF. Of course Glenn and Rogelio could help remotely.
Beta-beating code review, September 23, 2010
Glenn Vanbavinckhove (CERN), Rama Calga (BNL), Ryoichi Miyamoto (BNL), Ewen Maclean (CERN), Carmen Alabau (CERN), Frank Schmidt (CERN) and Rogelio Tomas (CERN)
Presentations
G. Vanbavinckhove Beta-beat code review
Summary
Glenn presents status, plans and required fixes of the beta-beating codes.
Parallelization of FFTs and SVDs should speed up the on-line
analyses. Nag library has parallel SVD and FFT with interface to Python.
Frank is meeting the Nag expert and will learn about these
particular functions.
Frank suggests some kind of general training for the beta-beat code.
This can be considered next year when Glenn releases the new extended version.
Future maintenance might be an issue. We rely on the modular structure
of the codes for an easy maintenance.
Ram mentions standardization and new versions.
Frank suggests the implementation of ICA algorithm
via the FNAL colleagues. They offered some collaboration
to Ryoichi and Frank.
ICA seems publicly available (Ram).
Frank commits (cc to all)
to investiage a bit to find help from Fermilab on ICA.
PTC could improve chromatic functions computation. Waiting for feedback
from Carmen, Ewen and Frank. Also Ewen will study model improvements,
we should get from him when to include what imperfections.
BNL-CERN LHC AC dipole meeting, October 30, 2009
Mei Bai (BNL), Carmen Alabau (CERN), Rama Calaga (BNL), Wang Gang (BNL), Verena Kain (CERN), Ryoichi Miyamoto (BNL), Peter Odo (BNL), Javier Serrano (CERN), Yipeng Sun (CERN), Rogelio Tomas (CERN), Glenn Vanbavinckhove (CERN) and Frank Zimmermann (CERN).
Presentations
[Indico link]
Summary
Javier presented the implementation of Solenoid savers as
the cure of the 100 Hz sidebands observed during the measurements
during last summer. The high order harmonics are more likely to
be an artifact. The recent measurements seem to be more noisy
and it is hard to observe the benefit of the Solenoid savers.
This will be further studied. Safety features have been added
to the AC dipole control equipment.
The Solenoid savers might be implemented in the LHC AC dipoles
next January 2010.
Verena presented the updates to the CCC control software.
New safety checks. The units of the AC dipole frequency are now
in tune units. It was discussed the need for a life-tune interlock
system to avoid users from exciting on resonance. This feature
is interesting but maybe not for commissioning.
The acquire button in the GUI triggers the AC dipole event and the
BPM turn-by-turn acquisition. A maximum of 10 turns difference between these
two processes can happen. It might be useful to display the OASIS
AC dipole voltage signal. The actual frequency of the AC dipole signal
will be easily accessible to the GUI.
Ryoichi presented independent simulations reproducing the most
relevant features of Ref concerning emittance growth. He also
showed that if high order harmonics are real they could be dangerously
close to the machine tunes.
The pessimistic white noise tolerance at store seems to be about the 1% (in time domain). This is OK for present AC dipole performance.
Rogelio presented some ideas of what the AC dipole beam commissioning
should address and when it could take place. It is rather obvious
that the AC dipole commissioning should take place in 2010.
The beam commissioning of 4 AC dipoles at injection and at higher
energies seems to be a rather lengthy procedure. It will be useful
to prepare time estimates.
Mei discussed the contributions to CM13. Ryoichi will help her in
summarizing the status from CERN but he will not attend CM13.
Wang is developing new measurement and correction algorithms.
We will be happy to hear from them when they are ready.
Yipeng presented the implementation of the AC dipole element in MADX.
First simulation data looks good. To be done: Introduce the vertical
AC dipole element, allow for more input parameters and update of the
MADX-SixTrack converter.
BNL-CERN LHC AC dipole meeting, July 29, 2009
Masamitsu Aiba (CERN), Massimo Giovannozzi (CERN), Verena Kain (CERN), Ryoichi Miyamoto (BNL), Steve Peggs (BNL), Frank Schmidt (CERN), Javier Serrano (CERN) and Rogelio Tomas (CERN)
Presentations
[Javier Serrano] Status of the LHC AC dipoles
[Rogelio Tomas] LHC AC dipole preliminary noise considerations
[Verena Kain] Status of the AC dipole operation software
Summary
First AC dipole dry runs and hardware tests were conducted. All the fundamental hardware
and software are in place. The first most likely working scenario is assumed to be at
injection with Qx=0.28 and Qy=0.31. The first use with beam should correspond to
a system commissioning in order to validate the AC dipole operation, probably also at
collision tunes 0.31 and 0.32. Capacitor bank has to be changed manually.
Miyamoto proposed to change AC dipole resistance in order to reduce peak strength but increase
bandwidth. The 200ms ramp is hard coded. This could be made variable in the future
but needs some work.
Verena agreed to use tune units for the frequency excitation rather than Hz.
Two AC dipoles can be simultaneously used but this needs to be tested.
The amplitude displayed on the GUI is in percent of 10V but the peak field
in the AC dipole depends on 3 parameters: Capacitor setting, Voltage and Frequency.
Tables with all this info should be provided to the users. Furthermore the
beam excitation depends on yet another parameter, the distance to the tune.
It might be possible to use the BBQ tunes for a safer GUI.
Amplitude detuning could be an obstacle to rely on these pre-computed tables.
However the safe operation of the AC dipole already has tight tolerances
on the allowed amplitude detuning (0.0005 at 3sigma, @ injection) Ref . Without forgetting that the
slow ramp of the AC dipole made it become a safe instrument.
In the AC dipole measured signal 100Hz sidebands around the main frequency and harmonics
of the main frequency are observed. The harmonics are most likely an artifact coming from the
measurement devices while sidebands could be real. In the most pessimistic case these
sidebands could cause 16% emittance blow-up. This should be avoidable if sidebands
are well known.
AC dipole only operates with "set-up beam flag" (before called "safe beam flag") which
is not fully safe at store energy for a 10^{10} proton bunch.
More dry runs and hardware tests are to come.
BNL-CERN LHC optics correction meeting, March 6, 2009
Rama Calaga (BNL), Masamitsu Aiba (CERN), Glenn Vanbavinckhove (CERN), Frank Schmidt (CERN), Javier Serrano (CERN), Rogelio Tomas (CERN), Ryoichi Miyamoto (BNL)