Wednesday
June 14th, 14:30
CCC
conference room 874/1-011
Ralph S. has sent a few comments on the minutes of the previous meeting relating to the feedback and corrector scaling on the ramp. The minutes have been augmented accordingly. Jan commented that strictly speaking one does not need to stop on the ramp for beam dump commissioning. Mike replied that some stops will anyhow be useful, but it still needs to be determined at how many different beam energies.
Roger reported that the last LTC discussed obstacles found in the beam pipe. As a result of these discoveries, now all magnets are being checked prior to installation. The vacuum group is looking at the problem of already installed magnets. This issue will be followed up in MARIC. It appears to be under control.
At the last LHC planning meeting, Lyn proposed collisions at 450 GeV in 2007. Steve asked Roberto to check the hardware aspects and Mike to look at the implications for beam commissioning. The new schedule will be decided at a TCC meeting on Friday 16th.
Mike pointed out that the proposed schedule definitely precludes a sector test. Roger confirmed this by presenting the schedule, which entails closure of the ring at the end of August, and limited hardware commissioning. The underlying theme of the new schedule is to leave out any non-essential items in order to accomplish 450-GeV collisions in 2007. The luminosity target is of order 1e30 cm-2s-1. The SPS collider had a few 1e28 cm-2s-1 at the time the W and Z bosons were discovered. At 1e30 cm-2s-1 the rate should be sufficient for detector calibration. Massimiliano remarked that the calibration is indeed useful for the detectors, but that this run should not be called a physics run.
Oliver pointed out that the 450-GeV optics could be partially squeezed e.g. to beta*=5 m for increasing the luminosity, and that there are no limits to the strength of the quadrupoles preventing such squeeze if the machine is not ramped (some power-converter modules would be switched off without any danger of spike during re-activation).
Frank announced that the 19th LHC MAC would be held June 15-17., including an LHC status report by Lyn, a collimation status report by Ralph A., and a presentation of the rf status by Trevor.
Rogelio first listed the team involved in these
optics studies and he acknowledged the magnet measurement effort, which has
provided a realistic error model of the machine. Measuring the betatron phase is easier than measuring the beta function
and fully independent of any optics assumptions. Rudiger
commented that the beta function can be computed from the betatron
phase, if the local optics is known, as has been done at LEP. Indeed Rogelio
next showed that there is a close relation between phase and beta beating. In
addition to betatron phase data, he considered
dispersion measurements based on radial steering. Rhodri commented that the
assumed BPM performance corresponds to the specification and not to the
performance actually achieved.
Clarify actual BPM performance. ACTION:
Rhodri?
Both beta beating and dispersion correction are
done using the pseudo-inversion of the same response matrix relating the
measured phase beating and dispersion errors at all BPMs,
plus the tune errors, to the 210 independent quadrupole
errors. The matrix is highly non-symmetric. The pseudo-inversion is computed
with an SVD algorithm. The correction is not guaranteed. No explicit orbit
errors are included, but the chromaticity sextupoles
are taken to be randomly misaligned with 2 mm rms. The measured quadrupole errors are used (not centered, not Gaussian) and
another random error of 5 units was added on top, representing the uncertainty
of the measurement. Spoolpiece displacements of 0.5
mm rms were also included.
Pessimistically a betatron
phase measurement error of 0.25 degree per BPM was assumed, while the simulated
rms error is about 30% smaller (but its distribution
is not Gaussian). The initial horizontal beta beating before correction is of
order 25% in x, but 60% in y. The difference is due to the random offsets in
the sextupoles.
After the meeting, Stephane explained that the asymmetry observed between the
H and V beta-beat does not come from an asymmetry between beta_x
at the SF and beta_y at the SD. Both are equal to
about 180 m, but the SD are twice as strongly excited (being twice less
efficient to correct Q'_nat because D_x at QD is only 1.1 m compared to 2.1 m at QF). In the
model shown by Rogelio, the dominant source of beta-beat comes from this source
via feed-down effect in the MS for an assumed 2mm r.m.s
CO in both planes at commissioning. Once this will be corrected we should come
back to the more reasonable budget of 20-25% in both planes.
The correction works well, using up to 5
iterations. After correction, the peak beating ranges from 5 to 15%, and
thereby is within the specified tolerance. If the target dispersion change is
set to zero, the dispersion is not affected by the beta beating correction.
This is much better than if dispersion is not at all taken into account, in
which case the dispersion error greatly increases during beta correction.
Rogelio concluded that beta correction must also consider the dispersion. For
most of the error seeds the peak dispersion lies within the tolerance.
Rogelio next looked at the implied changes in quadrupole strength. Largest changes are applied to the
matching quadrupoles. The strength changes of
individual quadrupoles KQ4-10 are at the 1%
level. All strengths required fall
within the range of the power supplies.
For some of the seeds the dispersion is not
correctable. Possibly, the non-convergence is linked to spurious dispersion at
the interaction-region quadrupoles.
The viability of dispersion correction and the
problem of non-convergence still need to be clarified. Action: Rogelio.
Lastly, Rogelio showed a simulated measurement
of the local coupling using the same type of turn-by-turn BPM data. The actual
coupling is nicely reproduced. The simulated coupling correction reduces the
minimum tune approach from 0.01 to 0.001. Rogelio concluded that we can measure
local coupling with a good accuracy, thanks to the double-plane BPMs, the readings of which contain the relative x-y phase
information.
He cautioned the audience that correct
synchronization of the BPMs is crucial, as experience
in the SPS has shown. If the BPMs function, at most 5
iterations are needed for a complete optics correction, where each iteration may include injection and re-optimization of
chromaticity.
Rogelio expressed strong interest in following
up this type of optics measurement and correction for the LHC commissioning.
Rudiger commented that the measurements could be done
at higher intensity, and Jean-Pierre agreed, recalling that the BPM error is
much reduced already at intermediated bunch intensities. Oliver asked why a
kick amplitude of 4 mm was considered and not, e.g., one of 2 mm at higher
intensity. Stefano strengthened this suggestion by stating that there is no
aperture margin for mm kick amplitudes. Jean-Pierre remarked that the real
issue is to get the software ready in time and that corresponding steps should
be taken. Rhodri and Mike both explained that the spectral analysis is done at
a higher level than the primary diagnostics. Oliver noted that this might be an
opportunity for using the MAD-X online model, and that MAD-X has a suite of
tools already available. Rudiger mentioned that he
was intrigued by the SNS success where almost all applications software was
written by the accelerator physicists. Frank suggested that the effect of the
measured quadrupole errors can be pre-corrected and
would not require beam-based measurement. The beam-based measurements are
useful for detecting unknown optics errors.
Mike asked why we want to correct the
dispersion beating. Stephane responded that the
spurious dispersion eats up the aperture margin. After some discussion,
Jean-Pierre and Stephane jointly concluded that most
of this aperture requirement comes from the bucket height and only 30% from the
momentum offset needed for chromaticity measurements. Only the latter
contribution might conceivably be relaxed. Oliver remarked that the spurious
dispersion will also have a negative impact on the collimation efficiency.
Stephane commented that it would be desirable to have
knobs for tuning the beta functions in the IRs. He
emphasized that we care much less about the beta beating in the arcs.
Gianluigi reported 1st results on
estimating the path-length difference between the two LHC rings. Two kinds of
worry are: (1)
different central frequencies for the two rings, (2) the need for
two different extraction energies from the SPS. Gianluigi
considered both random and systematic field differences. Using an
analytical formula and assuming a 2 mm rms closed
orbit error, he obtained a 3-sigma circumference difference of 0.6 mm. A more
accurate estimate can be computed from the measured magnet errors with MAD-X.
The angle error distributions for the two rings are almost the same. Stephane commented that this is unrelated to the magnet
sorting. The path length difference according to the second estimate is only
0.12 mm.
Pessimistically assuming a difference of 0.6 mm, the momentum SPS-LHC mismatch is 3.5e-5, which is three times smaller than the tolerance on energy matching. The tune difference between the two rings is 7.5e-5 per chromaticity unit. Both effects look small and acceptable.
Should nevertheless an unexpectedly large difference be observed in commissioning, the beams can still be safely captured in both rings, by adjusting the integrated dipole field independently for each ring, using the horizontal steering correctors, as pointed out by Oliver. There will be no need to operate with a different SPS energy for the two LHC rings.
Stephane commented that the maximum circumference change with orbit correctors is of order 0.5 mm due to aperture limitations, and that a difference between the two rings may arise from asymmetric end-laminations of the main dipoles, which shift the center of the magnet and may introduce a systematic asymmetry between the two rings. The expected shift is of order 1-2 cm, which would not be detected by the present magnetic measurement tools. Jean-Pierre added that the sagitta difference between inner and outer channel could also matter. Ralph S. asked whether there could be any impact of the transfer lines on the injection phase. It was thought that this would not be an issue. Oliver remarked that a bigger effect may come from non-reproducibility between cycles and he asked whether the latter would be different or the same for the two apertures. Stephane stressed that any systematic effect at the 1e-4 level is an issue.
After the meeting Stephane elaborated that the longitudinal magnetic center shall be known within 1-2 cm (not 1mm) to exclude effects from a systematic offset of this magnitude. Presently we can warrant only +/-10 cm by measurement (length of the short mole) by construction. He does not know if offsets by more than 1-2 cm can be ruled out.
Clarify whether the non-reproducibility is the same of different for the two apertures. ACTION:?
Roger announced that the next meeting on June 28th, during EPAC, may address 450-GeV collisions.
Magali encouraged the LHCCWG members to have a look at the documentation web pages, which were further developed and are now linked to the LHCCWG web site (“documentation” link), and to comment at the next meeting.
Wednesday June 28th, 14:30
CCC conference room 874/1-011
Provisional
agenda
Minutes of
previous meeting
Matters arising
Collisions at
450 GeV
AOB
Reported by Frank