Wednesday
July 12th, 14:30
CCC
conference room 874/1-011
Concerning the optics to be used for the 2007 450-GeV calibration run, Stephane, who could not attend the last meeting, commented by email that the nominal beta* in IR1/IR5 is 17 m not 18 m (it was changed in Version 6.4 to comply with n1=7 in the matching section). He further remarked that the aperture in the triplets obviously is not the same for beta*=6m and 18 m. For 6 m, the triplet aperture is saturated without X-angle (which is not needed for 156 bunches) while for 17 or 18 m we have plenty of margin (~12 mm or more). Recognizing the wish to have an optics as nominal as possible for this early commissioning, in his presentation for the last LTC he distinguished between nominal or pre-squeezed (see below) injection optics and the "high luminosity" injection optics (beta*=6 m, flat beams) that he generated a few weeks ago.
Stephane also reminded us of the pre-squeezed optics with beta*=11m in IR1 and IR5 which he had presented at the LTC-31-03-2004. In the present scenario, its advantages are two-fold: First this optics will allow pushing the luminosity (~10^30 for 43 bunches and 4 10^10 protons per bunches), while leaving a nice aperture in the triplet w/o X-angle (~8mm) and in the matching section. Then and most importantly, this optics will also allow a smooth transition to the nominal LHC commissioning (by "just" switching on the X-angle) and it will ease the later squeeze at 7 TeV (which was its initial motivation). Stephane’s comments were added to the minutes of the last meeting.
As for the time schedule of the ion run, Massimiliano clarified, also by email, that if ion running is incompatible with the main goal of the 2008 run (which is to accumulate several fb^-1 of good physics data), then it is likely to move to 2009. He also added that, in such a case, it would not be excluded to run ions before protons in 2009. Following his recommendation, the statement in the last minutes was rephrased to “the issue of ion running in 2008/2009 is yet to be discussed (by CRAG)”.
Regarding the useful length of the 450GeV run for the experiments, it had been said that this depends on the speed of detector debugging. Massimiliano tried to stress that the 450 GeV in 2007 should be driven by machine commissioning, not by the potential use the experiments could make use of so many nb^-1 or pb^-1.
Roger reported that the LTC meeting of 05.07.2006 has decided to run with the beta*=11 m optics in 2007. The new LHC schedule was circulated to the TCC and now goes through the approval process. Oliver mentioned that Roberto is investigating the possibility of swapping the hardware commissioning of sectors 5-6 and 6-7 with the one of 1-2 and 2-3, in order to provide more time for beam-dump commissioning and reliability run. He suggested that Roger and Brennan further follow up this question.
Roger explained that the presentation originally foreseen for this 10th meeting, namely “450GeV optics – IR aperture and IR bumps” by Yannis Papaphilippou, had been postponed to July 26, where it will now be given in tandem with Stefano’s presentation on aperture measurements. On a short notice, two substitute presentations by Frank were added to the agenda of the present meeting.
Frank
reviewed the nonlinear optics measurements to be performed at 450 GeV, including pertinent procedures and requirements, based
on his presentation on ‘flat
bottom optimization and tuning’ at Chamonix 2003
and references therein (studies by Oliver, Stephane,
Mark Hayes,
For b3 a correction at the 50% level is required on an arc-by-arc basis. The proposed procedure consists in pre-setting the spool pieces according to magnet measurements, adjusting the chromaticity with the lattice sextupoles, and then applying 7 pi-bumps across individual arcs, and/or detecting the chromatic phase advance. The expected resolution is at least 5 times better than the requirement. An example from the SPS illustrates, however, that here the chromatic phase advance measurement is less clean than what is predicted for the LHC. The a3 introduces chromatic coupling and contributes to Q’’. It can be measured by detecting the change in closest tune approach as a function of momentum deviation. It should again be easy to meet the specification. In addition, the measured a3 errors are smaller than originally thought. For b4 only a global correction at the 30% level is needed. The easiest method is to separate the tune (avoiding admixtures from a3) and then to minimize Q’’. The detuning with amplitude can also be verified. The last circuits are those for b5. The global b5 minimization can be done by correcting Q’’’. The earlier study by M. Hayes indicated that a local correction at the 50% level is required. This appears difficult to achieve, unless kicks significantly larger than 3 sigma can be applied. Frank finally pointed out that the SPS may serve as testbed for off-momentum nonlinear-optics studies as its Q’’’ value is 3 times larger than the LHC target value. In conclusion, it should be easy to adjust the spool piece correctors with the required accuracy, the only possible exception being b5.
Thijs asked whether a single spool piece could be missing and he wanted to know the reason for checking the spool-piece circuits octant by octant. Oliver replied that the original studies had been motivated by a request to investigate the implications of missing all spool-piece circuits in one or several entire octants, e.g. as the result of a power-converter failure. The result of these studies was that missing circuits in more than 2 octants are not acceptable. Paul commented that polarity errors are the most likely to occur. Presumably all spool pieces following such error would have the wrong polarity. The bump methods presented may help to localize such error source. Ralph S. pointed out that the nonlinear deflections from the sextupole fields are already important for 3-sigma amplitudes and could perturb the measurement of the b5 correction quality. Mike asked why we worry about losing 1 sigma from the 10-sigma dynamic aperture, if the collimators are set to 6 sigma. Oliver agreed that only a global correction of b5 would probably be needed.
Oliver stressed that the required measurements are a strong encouragement for having a PLL tune tracking as early as possible. Stefano commented that the collimators need to be opened if large kicks are applied. Likewise the momentum collimators should be opened for off-momentum measurements. Ralph S. explained that the measurement accuracy can be improved more efficiently by using several bunches and averaging, rather than by increasing the charge of a single bunch.
Frank next discussed the issue of transfer function for the separation dipoles, based on input from Oliver, Stephane and Massimo. This problem was highlighted in a Chamonix XII presentation on ‘parameter evolution for the first luminosity runs’ in which Oliver estimated that an expected error of 10 units in the D1/D2 transfer function would lead to a 3 sigma orbit offset in the triplet, and emphasized that a local correction requires a careful distinction between D1/D2 transfer function errors, triplet alignment errors, and triplet gradient errors in conjunction with crossing-angle bump offsets. There are 6 BPMs between the two D1, which are common for both beams, and 2 additional BPMs, separate for the two beams and without adjacent quadrupoles, close to the D2s. The procedure sketched in 2003 consisted in triplet alignment optics with low-beta quadrupoles switched off to define a straight line with 6 or 8 BPMs; K-modulation of individual Q1, Q2 or Q3 magnets to find offsets w.r.t. to a BPM reference line, inferring the transfer function errors from differences in the lines for beam 1 and beam 2, and, after correcting D1/D2 powering errors, obtaining a new reference line for triplet alignment Two different triplet alignment optics were developed by A. Verdier in LHC Project Note 325. Both require an emittance equal to, or less than, half the nominal. The x and y tunes decrease by one unit if the first alignment optics is simultaneously used in IR 1 and IR5. A refined procedure, arrived at in discussions with Massimo and Stephane, would first correct offset and angle upstream of D2 and then adjust D1 to steer the beam with zero slope through the IP. Afterwards the D1 and D2 strengths need to be further adjusted to bring both beams onto the same line. Frank listed a number of potential problems: the mechanical & magnetic axes of the low-beta quadrupoles are misaligned by 0.1-0.2 mm in x, 0.5 mm in y, and the beam needs to be steered through the mechanical center; BPM offsets may cause error of up to 5 mrad, compared with total deflection angle of 1.5 mrad, which is larger than desired precision of 3e-4; and the BPMs might have different offsets for beam 1 and beam 2.
Oliver
explained that the motivation for looking into this issue had been the absence
of any cold measurements for D1 and D2 (the only magnet measured, from RHIC,
had a different material). All the D1 and D2 magnets are only measured warm.
Cold measurements for the real magnets were possible neither in the
Frank continued his presentation with several questions: in which aspect is the D1/D2 transfer function effect different from other closed-orbit errors of the two beams? Do we want to measure transfer functions with beam at 450 GeV and at 7 TeV? Do we want to use the alignment optics - if yes, with which beam and which BPM resolution? He closed with his own proposal for treating this problem, consisting of triplet alignment optics, K-modulation, and 2-beam orbit correction including the D1 and D2 magnets as additional correction elements.
Oliver stressed that the large beta functions make the transfer-function issue a problem primarily for the squeezed optics, and that the triplet alignment will need to be performed before we reduce beta* to values below 2 m. Oliver also commented that the 2-beam orbit correction is unlikely to maintain the optimum aperture as there are not enough BPMs in the IR for clearly resolving the origin of the orbit errors. He suspected a loss in aperture compared to the ideal orbit if such a correction is applied, but he suggested that the effect of the proposed orbit correction could be explored in simulations. Paul asked whether the triplet alignment errors and the D1/D2 transfer-function errors could be disentangled by analyzing the orbit behavior during the squeeze. Oliver replied that this is a question of accuracy and measurement resolution, and that the two effects scale in the same way with beta*. Finally, Oliver remarked that triplet roll errors are more important than the triplet alignment errors, and he recommended that both the expected effect and the roll-error measurement procedures be investigated with some priority.
Wednesday July 26th, 14:30
CCC conference room 874/1-011
Provisional
agenda
Minutes of
previous meeting
Matters arising
450GeV optics –
Mechanical aperture and momentum aperture (Stefano)
450GeV optics –
IR aperture and IR bumps (Yannis Papaphilippou)
Possible
commissioning constraints from background in LHC experiments (
Follow up of web
documentation? (Verena et al)
AOB
Reported
by Frank