Summary notes of the forty-third meeting of the LHC Commissioning Working Group
Tuesday April 8th, 14:00
CCC conference room 874/1-011
Minutes of the Previous Meeting and Matters Arising
There were no comments on the minutes of the 42nd meeting. Roger announced that this 43rd meeting was focused on the 5 TeV run.
Overview of LHC Strategy for 2008/9 (Roger)
Roger presented the schedule and an indication of where we are. The sector 5-6 is 3 weeks late, two other sectors are doing well, and are even slightly ahead of schedule. The strategy for 2008-09 was determined in agreement with the DG, Lyn, and Jos Engelen. According to this, in 2008 we would aim for collisions of 43 on 43 bunches at 5 TeV. Following the 5-TeV run, the machine will then have to be trained up to 7 TeV, in a period from before Christmas 2008 to about Easter 2009.
Brennan asked if the exact dates of the shut down were already known. He mentioned that time would be needed for installing additional MKV kickers. Roger answered that there were no official dates for this shut down at this time. Django commented that also an ion run should be scheduled at the end of the proton run in 2009. Roger and Stefano were not aware about any decision for an ion run. After the meeting, Massimiliano recalled the current official strategy for ions, which is, he thought, to have a heavy ion run after the ‘first long high energy proton run,’ whatever that may be, with the corollary that the machine should be ready to deliver initial lead beams by the fall of 2009.
Roger recalled that the presentation by Werner in the 40th LHCCWG meeting on 12 February (“Bunch Filling Schemes for Early Runnig Scenarios,” Werner Herr) discussed various possibilities for colliding different numbers of bunches in LHCb in 43-bunch operation. A decision should soon be taken on the actual filling scheme which we will use in 2008. The questions for 5 TeV include the circuits and currents, e.g. which circuits are needed, which ones are not. Roger mentioned that magnets at zero field might not be trivial either, as it may not be sufficient to have zero excitation current. Another complication is that the experimental magnets during the 5 TeV run will be powered at their 7 TeV settings.
Walter remarked that if time suffices all the magnet circuits can be commissioned. Roger observed that it would indeed be stupid to find out only in 2009 that certain circuits are not working. Circuit questions would be addressed later in Mike’s presentation.
For the 2009 run, we would need to train the magnets to 7 TeV, to commission the remaining circuits that were not yet commissioned in 2008, and to implement 75-ns and 25-ns operation, plus possibly ions. Gianluigi added that also a decision on 50-ns spacing would be needed, in particular since this requires some preparation in the injectors. Massimiliano and Elias elaborated after the meeting that the 50-ns spacing option will be clarified at the injector level around the middle of this year, and that in particular a check of the 50-ns beam in the SPS is foreseen for week 33. The option of 50-ns spacing should be presented and discussed at the LTC. Mike commented that if we get beam in July, we might go a lot further than 43 bunches. Ralph agreed with Mike, and he recalled the PEP-II commissioning, where it was extremely easy to increase the number of bunches from 1 to 100s or 1000s of bunches. Ralph also stressed that we should test as much as possible in 2008, including crossing angles. Roger concurred that we should bring on the crossing angle with the 43x43 scheme. Both Mike and Ralph generalized these ideas by recommending pre-shutdown tests with beam in preparation of 2009. Thorsten asked whether regular machine development periods are foreseen. Roger replied that in regular LHC operation every month there would presumably be a few days of machine development and a few days of technical stops. Gianluigi added that machine development time is also needed for the injectors – e.g. for producing the 50-ns beam. Roger asked whether there existed already a clear plan for the injectors. Gianluigi responded in the affirmative, yes, the injectors will conduct 50-ns tests, as well as also studies with different bunch intensities on the same cycle. Ralph commented that studies of electron-cloud effects in the LHC would need 25-ns bunch spacing. Jorg cautioned that with such bunch spacing the machine protection for many bunches must first be set up. Ralph, however, reminded the team that if one bunch is safe, many bunches would also be safe, since already one bunch is considered an “unsafe” beam and, therefore, all the relevant protection must be put in place and guaranteed already for a single bunch.
Beam Parameters for 5TeV Operation in 2008 (Massimo)
Massimo’s talk considered aperture, luminosity, overall target parameters, and the next steps. He first described the general assumptions, which primarily are to consider the beam parameters of commissioning phase A, no crossing angle, and the same values for beta* as for phase A. Roger cautioned that the absence of the TCTVBs (due to contamination) in points 2 and 8 preclude any beta* values below 6 m in this IPs. Ralph confirmed that 6 m beta* was certainly safe, while the exact limit was still being worked out.
An important issue is the longitudinal parameters. One possibility may be to assume similar parameters as for 7 TeV. However, the hardware for generating the longitudinal blow up will not be ready for the commissioning. Frank commented that the longitudinal emittance should grow with beam energy to avoid a decrease in beam stability due to loss of Landau damping, so that at 5 TeV we may accept a slightly smaller longitudinal emittance than at 7 TeV. He added that if the beam was found to be unstable longitudinally probably the blow up could be implemented during the 2008 run, since the longitudinal blow up had already been successfully tested in the SPS as LHC test bed. Andy confirmed that the blow up is under development, and not ready on day 1, but that, if it really were shown to be needed, it could be set up during the run. Ralph asked for the consequences of a missing longitudinal blow up. Stephane pointed out that presently the blow up affects bunch length and energy spread differently, adding that without blow up we may expect 4 cm rms bunch length and a correspondingly smaller luminous region.
Massimo now compared the apertures at 7 TeV and 5 TeV with beta*=2 m, no crossing angle, and re-scaled parallel separation. Stephane asked whether we could also have the apertures at 5 TeV with crossing angle. Massimo replied that, yes, this information could be provided, and that apertures with crossing angle had already been shown for IR8 by Werner. Ralph remarked that for 10 m beta* the IR aperture seemed to be already close to the arc aperture even without crossing angle, which he considered somewhat surprising. Stefano commented that the arc aperture must also be computed for the different emittance at 5 TeV, which might explain part of the puzzle. Massimo summarized Werner’s analysis for IR8: Down to beta*=2 m, the aperture and the beam-beam separation are sufficient for bunch spacings of 50 ns or above. Stephane strongly cautioned that all aperture margin is lost if we opt for beta*=2 m at 5 TeV. He recommended to squeeze no further than about 3 m in order to leave the same aperture margin as what we had been assuming for the 7 TeV run with 2 m beta*.
Massimo next presented the aperture as a function of beam energy, for beta*=2 m, to give an idea of the dependence on energy. At 4 TeV the aperture is about two third of the 7-TeV aperture.
Now turning to the expected luminosity, Massimo recalled that the target value depends on (beta gamma), which implies a rescaling of the 7-TeV values by a factor 0.71, assuming the same longitudinal parameters. The longitudinal distribution of luminosity is an exponential with an rms length equal to the bunch length divided by sqrt(2). Massimo presented a summary table with beam parameters and luminosity values for the various IPs. The next steps will include an update on the longitudinal parameters, with the help of the rf goup, and an assessment of reduced radiation damping. The optics configuration for the 5-TeV run is already available at http://lhc.web.cern.ch/lhc/ .
Implications of 5-TeV Operation: Collimators, Other
Ralph commented on implications of the 5-TeV run on collimation. A substantial gain in quench margin and relaxed requirements on the cleaning efficiency are obvious benefits. Also the damage thresholds are relaxed, by about a factor of 2. He emphasized that with 43 bunches at 5 TeV the total energy stored in the beam was similar to that in the Tevatron. In other words, in 2008 we must handle only Tevatron class beams, and need not enter into a new regime. Detailed collimator settings for 5 TeV will be worked out.
Stephane remarked that the s.c. magnets at 5 TeV feature a larger fraction of p.c. component (4% vs 1%), and that the magnet behavior may be less well known, as the magnets were measured for 7 TeV cycles.
5-TeV Settings (Mike)
Mike discussed the currents required for 5 TeV. He extracted the maximum strength for all circuits, including squeeze to 55 cm in IR1 & 5, and squeeze to 2 m in IR8. Default transfer functions are presently taken from the layout database. Mike highlighted that these cannot be used for beam operation, but FIDEL values are not yet fully available at 5 TeV. Luca Bottura is helping to get FIDEL set up for 5 TeV, though he is no longer formally responsible. The effect of the FIDEL correction could be a few percent of the magnet strength. Oliver pointed out that errors of this magnitude would imply an enormous beta beating.
For a number of circuits MAD does not supply a strength value. Stephane commented that the maximum strength of the octupole spool pieces should be different from the value listed. In Chamonix 2006 Massimo had presented a list of magnets for which the initial commissioning could be skipped (see Chamonix XV: “Electrical Circuits Required for the Minimum Workable LHC During Commissioning and First Two Years of Operation”, M. Giovannozzi). Frank commented that ROF and ROD are the Landau octupole circuits. Massimo commented that the beam might be unstable for TOTEM beams if we do not have the Landau octupoles. Both Ralph and Frank added that even a single bunch at reduced intensity could lose transverse Landau damping. Ralph emphasized that we might be lost if the beam turned out to be unstable and we had not commissioned these circuits. He advocated that we definitely commission the lattice octupoles .
Massimo suggested that we skip the octupole spool pieces, and that we might skip the decapoles too. Oliver remarked that we should also ask how much time the hardware commissioning takes, and if the time saving from skipping a circuit was really that significant. He qualified that it would be nice to have the decapole spool pieces operational, and that the more circuits we have the more we can test with the beam. The decapoles would also be useful for verifying beam measurement procedures
Walter commented that he had established a list of priorities. The Landau octupoles would be added to the list of essential magnets. He was assuming that the nonlinear inner triplet (IT) correctors were not needed. Massimo confirmed that these IT correctors indeed were not required for beta*>1 m. Ralph questioned why we would not commission all IT correctors to have a response in case of problems. Stephane clarified that we could include MCSX (this could help to correct coupling generated by feed-down from the b3 component in D1 with vertical crossing that might induce a linear coupling (c_) of order a few 1e-3 with opposite sign for the two beams), but if we trust the magnet measurements we could drop all other IT magnets. Ralph also asked whether there was a large time saving from not commissioning these circuits. Stephane responded that these circuits have no quench protection packages, and a small power supply, so that their commissioning should not take a lot of time. Walter replied that we are trying to speed up the hardware commissioning by all possible means, since the total available time has shrunk dramatically. Roger clarified that the final list of circuits to be commissioned was being prepared by Walter and Mike. Mike mentioned the QT circuits.
Oliver pointed out that the synchrotron frequency crosses 50 Hz during the ramp. The synchrotron tune of 61.8 Hz at injection decreases to 21.4 Hz at top energy. If we are unlucky, 50 Hz could be reached exactly at 5 TeV. He suggested that this parameter be checked with the rf group.
Stephane asked whether there was a strong requirement for 2 m beta*, reiterating his recommendation that beta* should stay at 3 m instead of 2 m, to provide a similar margin as at 7 TeV. Oliver suggested that we can always stop at any intermediate beta*value during the squeeze. Massimiliano reported the preference of the experiments which is to squeeze as much as possible, so as to use the protons as efficiently as possible. Stephane cautioned that stronger squeezing implied a larger risk. Ralph asked about the relevant parameter range. Stephane explained that the scaling should be 7/5x2 ~ 2.8 m for beta* to arrive at the same margin as what we had foreseen for 7 TeV in phase A. By contrast with beta*~2 m at 5 TeV, we would almost reach the situation of ultimately low aperture margin. Massimo asked whether we should then assume that we stay at beta*~3 m at 5 TeV. Stephane recommended that we target n1~10 with crossing angle.
Stefano commented that there could be a difference between test mode and physics running. Stephane replied that the machine should be safe even when testing. Ralph supported Stefano, and recommended that we explore as much as possible in 2008. Roger stated that we must first define which parameters we are aiming for. Massimiliano summarized the 2008 plan as physics run without crossing angle, and an exploration of crossing angle in studies preparing for the following year. Ralph suggested that physics would profit highly from pushing the boundaries.
After the meeting, Massimo clarified some points concerning the required circuits during the LHC commissioning. His analysis presented at Chamonix in 2006 had indeed included the 7 TeV runs and was not restricted to the 450-GeV engineering run. The outcome of that analysis is summarized below (from the conclusions of his Chamonix XV paper):
Circuits mandatory from Day 1: orbit correctors; quadrupole correctors: MQTs, MQTLs, MQSs; sextupolar correctors: MCS, MS; spectrometer compensators; all RF modules.
Circuits required from Day 2: decapolar correctors: MCD; octupolar correctors: MCO; skew quadrupole correctors in the triplets: MQSX; normal sextupolar correctors in the triplets in IP2: MCSX; lattice octupoles: MO.
Circuits not required from Day 1: skew sextupolar correctors: MSS; nonlinear correctors in the triplet quadrupoles.
Massimo recalled that at the time the gain in the hardware commissioning time obtained by skipping some circuits was not considered worth it. The above list is still valid and correct. Massimo also explained that the aim of Walter’s analysis is mainly focused on the 600 A circuits, for which the energy extraction system requires considerable effort. Massimo computed the powering level for these circuits based on the optics knowledge as well as the magnetic field measurements: This information will be used to assign each circuit to a different hardware commissioning class depending on the maximum current required (200 A. 400 A, 550 A). In principle, circuits with 80 A-120 A current levels should not pose any problem as their commissioning should be fast. Hence, strictly speaking, both MCOs and ALL triplet correctors should be hardware commissioned. As a last point, Massimo mentioned that Walter was trying to get ALL the circuits - even those which are not strictly needed for the initial commissioning - tested with current up to 50 A: This should ensure that in case of an emergency, these circuits would be available after some (quick, hopefully) training.
Tuesday April 22nd, 14:00
CCC conference room 874/1-011
Minutes of previous meeting
5 TeV Stage-A parameter list (Massimo)
Machine checkout (Gianluigi)
Dry runs (Mike)
Reported by Frank