Summary notes of the thirty-sixth meeting of the LHC Commissioning Working Group


Tuesday December 4th, 14:00

CCC conference room 874/1-011

Persons present


Minutes of the Previous Meeting and Matters Arising

There were no comments on the minutes of the 35th LHCCWG meeting. Responding to a remark by Massimo, Frank explained that several earlier comments on a draft of these minutes which he had received by email, e.g. from Stephane, Massimo, Frank S., Jan, Samy, and Stefano, had already been included in the finally distributed version.

Massimiliano announced that after numerous discussions with a lot of people he discovered that there are a number of interesting filling schemes with 43 and 156 bunches, which could, for example, deliver half the maximum luminosity to LHCb without losing any luminosity in ATLAS and CMS, taking into account that ALICE may be happy already with only one or two colliding bunches.


=> ACTION: Follow up on early filling schemes (Massimiliano, W. Herr)


Experimental Magnets Commissioning (Helmut) 

The presentation on the commissioning of the experimental magnets was split into two. The first half was delivered by Helmut, the second half, focusing on operational aspects, by Delphine.


Helmut’s talk covered overview, commissioning and procedures, and details. After listing a number of pertinent references, including LHC Project Report 929 from EPAC’06, Helmut recalled that most of the expected perturbations caused by the experimental magnets are small effects. In particular, the solenoid-induced coupling at the LHC is much weaker than at lower energy machines. The correction schemes are based on known and well-established algorithms. 


The strongest solenoid field is that of CMS with an integrated strength of 52 Tm, followed by ATLAS (12 Tm), and ALICE (6.05 Tm). ALICE and LHCb also contain dipole spectrometers (with a strength of 4.5 Tm in the case of LHCb). Helmut pointed out that the LHCb spectrometer cannot be ramped down to the corresponding injection values due to power converter limitations. As a result the crossing angle arising from the spectrometer would be 2-3 times larger at 450 GeV than at 7 TeV.


Witold pointed out that ATLAS would like to turn on and off the toroids and the solenoids separately, emphasizing that the toroid effects on the beam should be explored as soon as possible, and that a test of the ATLAS toroid impact should be folded in early on in the commissioning program. In other words ATLAS requests to decouple the solenoids from the toroids and to do the toroid checks early. Helmut suggested that this check be done at injection where the effect would be largest.


Helmut now addressed the question when exactly the experimental magnet commissioning should be done. Originally it was supposed to come between phases A.9 and A.10. Another option will be before phase A.7 if the 450 GeV collisions are meant for physics. An even earlier possibility could be after, or at the end of, phase A.4 (450 GeV optics studies).


Coming back to Witold’s remark, Massimiliano recommended to distinguish between toroid, solenoid, and dipole commissioning in the schedule.


Helmut presented a proposed procedure.  The LHCb and ALICE spectrometer dipole compensation would first be executed at 450 GeV, and be left off for the first ramps, before checking their effect at 7 TeV.


Massimiliano asked whether the spectrometer magnets and their compensation could be turned on in steps from zero. Mike replied that it is not possible to excite these magnets at low current.


Next the solenoids of ATLAS, CMS, and ALICE would be compensated. Helmut suggested commissioning one by one, to start with the weakest, to measure the coupling it generates etc, to apply the calculated compensation, and to iterate if needed. Solenoids are turned on at their maximum field. They and their compensation are kept constant, independent of beam energy, and at 7 TeV they will have a greatly reduced effect on the beam.


There will be control knobs for both spectrometer dipoles and solenoids. ALICE and LHCb require knobs to adjust the internal crossing angle. Each of the three experimental solenoids is treated by a knob with two states, corresponding to the respective solenoid being on or off, respectively.


Helmut next presented a few details of spectrometer bumps, referring to W. Herr’s Chamonix 06 presentation. The maximum amplitude of the LHCb bump is not quite a mm. Since it is a closed bump without interleaved quadrupoles, no dispersion is generated by this bump. Always all 4 magnets are driven together.


Massimiliano, mentioning the wish to execute van-der-Meer scans etc in LHCb, asked whether providing a zero crossing angle would be simple. The answer was that it is not possible with these 4 dipoles only, but would require a superposed external bump.


Now Helmut commented on the strength of the solenoid coupling, in particular addressing the question why it would be weaker than LEP: The LHC solenoids are 5 times stronger, but the beam rigidity is 20 times higher, so hat the net effect is 4 times smaller than in LEP.


Discussing the precision of the coupling correction, Helmut noticed that the LHC design tunes are not far from the coupling resonance, leading to a required precision of coupling correction of order 0.003 for the minimum tune distance (see EDMS LHC-B-ES-000910-00 by S. Fartoukh and J.-P. Koutchouk). Helmut suggested that we can correct more precisely, e.g. to the 1e-4 level if we can measure with this precision as we should be able to. The only solenoid that is expected to induce an intolerable coupling is the CMS solenoid at 450 GeV, for which the predicted coupling effect is about twice the tolerance.


Frank commented that under time pressure during commissioning perhaps we would only need to check the CMS solenoid effect at 450 GeV and not need to worry about the other magnets and about 7 TeV.  Mike assisted saying that it might be difficult to get high tune precision during commissioning. Helmut replied to both that these coupling checks are quick measurements, and that in his view everything should be done as carefully as possible at the LHC.


Helmut illustrated how the solenoids have been implemented in MADX by Thys Risselada. He closed with the example of a MAD-X prediction for the CMS solenoid compensation. Edge effects are included with a hard-edge solenoid model. The compensation was checked for 2 beams. A tiny 0.1% beta beat and 0.2% dispersion beating result from the global correction.


Frank asked about the plan for the external angle, and in particular whether we will initially collide with a crossing angle or head-on in LHCb and ALICE. Massimiliano remarked that a small crossing angle could even be beneficial for ALICE, avoiding any displaced collisions (the angle would imply several sigma separation at the parasitic encounters). Stefano replied that it is not foreseen to commission the external crossing angle until operation with 156 bunches has been established.


Experimental Magnets Commissioning (Delphine)

Delphine’s presentation reviewed entry & exit conditions, solenoid commissioning, and dipole commissioning.


The entry condition is that beams 1 and 2 are commissioned at 450 GeV. Both beams are circulating, with a good lifetime. The orbits are measured and corrected; the tunes and the coupling are also corrected.  The ramp is commissioned. The coupling is corrected at 7 TeV. There is neither a crossing angle, nor any separation bump. The commissioning is performed with one bunch per beam. The desired exit condition is that the solenoid induced is corrected and that there is no orbit distortion from the spectrometer bumps.


The commissioning at injection comprises establishing a reference orbit for each beam as well as a coupling reference for both beams. For each magnet the same procedure is applied. First the solenoid is switched on at standby current. Orbit and coupling are measured and corrected. The magnet current is then ramped to the nominal setting, where measurement and correction are repeated. A significant coupling is predicted for CMS only.


The machine coupling is expected to change during the ramp. It is assumed that the coupling has already been corrected throughout the ramp. The solenoid-coupling correction found at 450 GeV is simply extrapolated into the ramp. A check is performed at 7 TeV. Once commissioned the solenoid will stay on.


Massimiliano inquired whether we do not need to turn it off against when turning on the next solenoid. Delphine clarified that the previously compensated solenoids could and probably would stay on.


Now Delphine turned to the commissioning of the experimental spectrometers. Again a reference orbit is taken. The bump closure has to be corrected locally without using any external correctors. The reason for any non-closure would have to be investigated. First the bump is established for beam 1. Then this beam is dumped, and the measurement is repeated for beam 2, sequentially. Corrections are incorporated into the ramp. The bump closure is checked during the ramp as well as at 7 TeV. If the bump closure is not stable, the reason must be investigated.


Once the commissioned has been finished with one polarity, we need to switch off the 4 bump magnets, and change their polarity. The complete procedure then has to be repeated with the new polarity. Delphine emphasized that we need a clear description and clear references for the magnet polarity (in particular a common definition with the experiments).


With the crossing angle on, the CMS solenoid will induce an orbit distortion. Therefore, when turning on the crossing angle, a recommissioning of the solenoid compensation will be required. Also the ALICE and LHCb spectrometers will need to be recommissioned after turning on the external crossing angle.


Stephane commented that roughly 3.5% of the crossing angle bump is converted into the other plane due to the CMS solenoid.


In her conclusions, Delphine pointed out that the LHCb spectrometer has already successfully been commissioned recently, when it was ramped together with the compensator following the LHC ramp. The minimum power converter current determines the minimum possible bump amplitude at injection. The maximum di/dt of the LHCb magnet has to be taken into account when generating the ramp function.


Massimiliano asked whether the LHCb magnet limits the maximum speed of the ramp or whether this is in the shadow of other constraints. Delphine answered that the shape of its excitation curve must be chosen such that it remains in the shadow. Mike added that this could be an issue, for a typical rate of 4 A/s. Massimiliano queried whether the same situation is expected for ALICE. Delphine responded that the ramp of the ALICE magnets would be measured in the following week. Mike mentioned that he will check these questions with the power-converter experts. Alick asked whether the combined effect of crossing angle and solenoid would be a concern only at point 5 or also at some other IPs, adding a second question how many times one will want to turn and off the solenoids. Mike confirmed that the solenoid-on and off tests are not required for IPs other than IP5, and he explained that this test would need to be done at least once for CMS.


After the meeting Jorg commented that to check the non-closure of the spectrometer bumps during the ramp seemed to him hopeless or at the least very difficult, since either there would be changes through the ramp that mask the (small effect) or the orbit feedback would iron out the leakage. In fact the most sensitive method would be to ramp up the spectrometers+compensators to their 7 TeV settings at 450 GeV. In that case one could profit from the large amplitudes and be more sensitive to any non-closure. He thought that without the crossing angle, this should be (almost?) possible: For LHCb the amplitude is ~ 10 mm, which seems within reach. In any case one should try to go as far as possible. If it turns out not to be possible to go all the way, then it may be better keep the currents at injection level during the ramp (spectrometers and compensators), and then ramp them up at 7 TeV.


Alternative SPS Filling Scheme (Magali) 

Magali’s presentation was prepared together with Elias Metral. After an introduction, she reviewed 2007 SPS MD results, and then drew some conclusions. One final slide on the nominal filling scheme with lower bunch intensity completed her presentation.


First addressing the “why and what?,Magali explained that during 2006 instabilities of the LHC beam had been observed in the PS at extraction. Two references are a presentation at the 75th APC (December 15th 2006) by R. Steerenberg on “observations of the high energy instability in the PS”, and a presentation in the 20th LHCCWG meeting (February 14th 2007) by E. Metral on “implications for the injectors”.


The reason for the instabilities was later traced to the use of two different cavities behaving differently (that is providing different voltages for the same reference). Regardless, investigations started and continued to find solution to future similar problems. Indeed two solutions were proposed: the first is a double-step bunch rotation presented by H. Damerau at the 75th APC; the second is an alternative filling scheme described by W. Herr and Elias in the 20th LHCCG meeting. Magali’s talk was a continuation of Werner’s and Elias’ presentations. The alternative filling scheme consists of groups of 2, 4 or 5 batches of 48 bunches instead of batches of 72 bunches as in the nominal LHC scheme.


This alternative filling scheme could be studied in an SPS MD on October 17-18, where Magali compared the performance of 4 injections of 72 bunches with that for 5 injections of 48 bunches. She swapped from one filling scheme to the other, using the same intensity per bunch and the same supercycle length, changing only the timings. The intensity was 1.1e11 ppb at flat top, which is about ~10% below nominal. Therefore, she also looked at 48-bunch train injections with 1.2e11 ppb at the end of the flat top. The nominal case was approximately obtained, with 4x72 bunches SPS, and a total intensity of 3.24e13 injected initially, decreasing to 3.01e13 protons at the end of the store corresponding to a beam loss of 7%.  With 5x48 bunches, the total initial intensity was 2.74e13 protons, shrinking to 2.56e13. The beam loss of 6.6% was about the same as with the 4x72 injection pattern. In a second set up, 5x48 bunches where injected with a higher initial total intensity of 3.1e13 that decreased to 2.7e13, amounting to a higher beam loss 12.3%. No comparison was possible for this higher (~nominal) intensity per bunch throughout 2007, since the nominal bunch intensity could not be obtained with the nominal batch configuration consisting of trains of 72 bunches. On the other hand, with trains of 48 bunches the nominal LHC bunch intensity was reached in the first attempt without a problem. The intensity variation along the train was similar for 48 bunches per train and 72 bunches per train. The emittances measured at top energy were also similar, and lay within specification (<3.5 mm mrad normalized). The emittances at PS injection and extraction were also similar in the two cases, and indeed a little smaller horizontally for the 48 bunches.  A snapshot of the BPM readout illustrates that in the nominal scheme batches are separated by 8 missing bunches (or a gap of 225 ns ~ at the limit of the LHC injection kicker) compared with 9 missing bunches in the alternative 48-bunches filling scheme (or a gap 250 ns, which leaves sufficient time for the injection kicker).


Magali concluded that it is easy to switch from 4x72 bunches to 5x48 bunches, within a few minutes. A difference is the maximum bunch intensity, i.e. only the alternative 48 bunches could reach the nominal intensity. The emittances are similar. The alternative scheme leaves more time for the rise of the LHC injection kicker. The alternative scheme does, however, result in a smaller total number of LHC bunches, which would translate into a reduction of the instantaneous luminosity by about 8%. On the other hand the alternative scheme offers as an additional advantage a shorter PS cycle (2 basic periods, or 2.4 s, instead of 3 basic periods or 3.6 s). Accordingly, the SPS cycle could be reduced from 21.6 s to 20.4 s.


Magali recalled that even if instabilities would reappear, another solution has also been studied, which is the double step bunch rotation that would allow retaining trains of 72 bunches. Possibly, the alternative 48-bunch scheme could become part of the beam commissioning sequence, e.g. to be used prior to the nominal 2808 bunches (2592 bunches), in case of problems.


As another possible step in the beam commissioning, the nominal filling scheme with batches of 72 bunches but lower bunch intensity would be a probable candidate. The generation of such beam was also explored in an SPS MD. The intensity was decreased in the PS booster by a factor 10 using the sieve and shavers. Minor optimization was required in both the longitudinal and the transverse planes. The nominal 72 bunches were easily produced with intensities 10 times lower than nominal.


Stefano commented that during the SPS MD on the alternative filling scheme, the cycle was not optimized for 5x48 trains. He asked whether the losses could possibly still be improved if such optimization were done.  Magali concurred, replying that, yes, possibly they could be reduced further. Elias added that one could certainly reduce the length of the cycle for the alternative filling scheme. For the nominal 72 bunches, the cycle was already well optimized.


Noticing that for the nominal scheme one could not obtain 4x72 trains with nominal bunch intensity, Stefano asked whether the higher bunch intensity available with the alternative scheme might not compensate for 8% lower luminosity. It was pointed out by Elias and Gianluigi that the limitation on the maximum intensity for the nominal scheme arose from MKDV dump-kicker vacuum problems, which is a separate issue that would not constrain the beam intensity which could be extracted to the LHC.


Wolfgang Hofle commented that the PS emittances quoted looked very small. Elias replied that the minimum emittances ever achieved, many years ago, were of order 1.8 mm mrad. Nowadays, 2-2.5 mm mrad was typical from the booster. At extraction the PS emittances are close to 2.5 mm mrad. He agreed with Wolfgang that the measured value of 1.5-1.8 mm mrad appeared small. However, he also cautioned that the PS emittances varied a lot throughout this year.


Ezio queried whether the outgassing problem would be solved for next year. It was reiterated that this problem would not appear when we extract beam to the LHC. Verena remarked that it might still be harmful during set up.


Ezio next asked whether the ultimate bunch intensity of 1.7e11 could be reached with the two schemes. Gianluigi replied that above the nominal intensity, the losses are increasing more strongly than linearly. The SPS presently is at its performance limit with 1.1e11 protons per bunch. He concluded that as far as the SPS is concerned no large difference exists between the two schemes regarding the ultimate bunch intensity of 1.7e11.


Helmut reminded the working group that other arguments in favor of the alternative filling scheme were presented by Werner and Elias at the 20th LHCCWG.


Elias clarified the SPS problem: For nominal LHC intensities, outgassing of the dump kicker MKDV caused an interlock after 1 cycle; another problem was related to outgassing of the TIDVG dump triggering an MKP (injection kicker) vacuum interlock after a few minutes. Gianluigi detailed that this kicker had been replaced during the last shut down and did not yet recover the nominal performance. There exist some ideas how to improve it and how to reduce the impact. Limits from outgassing are also affecting the CNGS beam. He again stressed that the problem occurs only if the beam is not extracted to the LHC. 


Stephane remarked that the limited field flatness of the MKI kickers, which exhibits very fast fluctuations at the level of +/- 0.5%, can yield a difference up to 1 sigma between the injection oscillations of two consecutive LHC bunches  (which therefore need to be damped at the 40 MHz frequency). He added the question whether we would gain with 48 bunches and if we could inject after the initial kicker overshoot error. The answer was negative, since the LHC gaps are increased by 1 or 2 bunches only (40 missing bunches instead of 38 or 39), which is not sufficient to avoid the kicker overshoot.


LHC Transverse Damper Commissioning (Wolfgang Hofle)

Wolfgang’s talk was structured in four parts: overview of the system, status of hardware commissioning, beam plan for commissioning, and summary.


Wolfgang started by reviewing the principle of the transverse multi-bunch feedback, pointing out that several parameters will need to be adjusted using the beam, e.g. ones related to the beam time of-flight between the pick up and signal processing and the betatron phase. Measuring the transfer function will also be part of the set up.


The transverse damper system has three functions: damping (transverse injection oscillations), excitation (including “abort gap cleaning”), and feedback (multi-bunch instabilities).


Concerning the first point, Wolfgang showed the example of the injection kicker wave form measured by the BT group on a prototype. The kicker ripple is such that when injecting the last beam, the beam already circulating in the ring will be excited.  Gianluigi asked whether this was the case despite the long abort gap. Wolfgang answered in the affirmative. A simulation of the emittance growth due to the injection kicker ripple shows that without damper all 4 batches from the SPS are strongly affected as also, slightly, is one batch just in front of the newly injected group. The simulation demonstrates that the transverse damper greatly reduces this emittance growth.


Wolfgang now recalled the damper performance specification from the LHC Design Report, before turning to the damper hardware, comprising 20 electrostatic kickers and 40 wideband amplifiers.


He described that one will need to check the optics around IR4 as part of the damper beam commissioning. The performance for LHC optics version 6.5 was compared with that for the original optics assumptions. The maximum kick per turn at 450 GeV/c is about 0.3 sigma.


The overview of one damper system illustrated the damper components located in the tunnel (kickers and power amplifiers), UX45 underground (the controls amplification and interlocking), and the surface (low level electronics). There will be no access to UX45 at any time with beam, i.e. also not during beam commissioning.


Wolfgang next reviewed the signal processing hardware. The time of flight of the signal has to be matched with less than 1 ns precision. This matching is to be done during beam commissioning. A special coupler BPM is used to detect the motion of each bunch. The peak voltage for the ultimate beam at 7 TeV reaches values up to 140 V. The behavior for this large voltage could not be checked in the laboratory. Signals can be simulated however. Simulations can be compared with measurements on the surface.


The present status of the transverse damping system was summarized as follows. In the tunnel the kickers are installed with bake out completed. Power amplifiers are installed and tested with their water cooling working fine. The amplifier final characteristics are being measured in situ. Pick ups were also installed (BI group). Non-conformities in cable lengths are to be followed up. In UX45, drivers, HOM loads, PLC controls, and interlock crates are installed. Signal observation crates are still to come. The characteristics of 16 installed amplifiers were already measured, with excellent results. Concerning the surface installation and hardware commissioning, several things still need to be done. The schedule is tight, but possible within 5 months.


Finally Wolfgang turned to the beam commissioning. In phases A1-A2 signal levels and calibration can be verified. Pick up signal delays from Q7 and Q9 will be equalized. The kicker strength will also be checked. In phase A3, after the RF capture, the RF front-end of the damper will be commissioned and optics checks be performed. Digitization and “f_rev tagging” (identifying the location of the abort gap / 1st bunch) will be commissioned. As for the optics checks, the phase advances between Q7, Q9 and the damper as well as local beta functions at the same places will be verified. The injection kicker pulse shape is scanned in this phase too. In commissioning phase A4 one can measure the decoherence without the damper, the open-loop transfer function, and then close the damper beam feedback loop after adjustments of gain, phase, and delay (3 parameters). Also the beam-blow up facility will be commissioned and the beam lifetime be measured as a function of the damper gain at 450 GeV. In phase A6, here defined as the first two minutes of the ramp, one can test the use of the damper as abort-gap cleaner (possibly this operation mode can be commissioned before starting the ramp), and check out the machine protection interlocks.


Alick asked what would happen if the signal from the damper would show the damper “to go crazy”. Wolfgang replied that one would rely on the BLM system (to detect losses) and the standard machine protection.


In phase A7 (ramp), the damper loop will continue with adjustments during the ramp. This again involves measuring open and closed-loop transfer functions. The abort gap cleaning can be verified up to 7 TeV during this stage. Wolfgang pointed out that it would be useful to have a good orbit feedback at this step of the commissioning, in order to obtain a good dynamic range for the damper.


In summary, the power system is well advanced, and the present focus is on finishing the low-level components.  Wolfgang stressed that he could do extremely useful things with beam from day 1 onwards. He emphasized that it was essential to commission the damper at the beginning of the beam commissioning.


Stephane commented on the need to know the phase advance between pick ups and kicker, underlining that the phase advance will change with dynamic conditions in the LHC. He asked whether this phase advance will be measured all the time, as it could change by 20 degrees easily (e.g. from fill to fill). 


Wolfgang answered that a precision of 5 degrees would be good, based on SPS experience. 20 degree errors are visible in the SPS, where they result in a degradation of damping. One could implement a function which follows the change if it is reproducible from cycle to cycle.


Stephane queried whether the phase information was not already contained in the signal of the damper. Wolfgang responded that in principle, yes, this was the case, but that inferring the optics parameters from the damper signals would need an appropriate adaptive filter.


Looking at the kicker wave form presented by Wolfgang, and in particular the signal of the perturbed bunches, Elias inferred that there was no large advantage in going to 5x48 bunches. The reason is that the entire first batch is perturbed and not only the first few bunches, which could well be indicative of a problem with the kicker. One of the perturbation frequencies is around 10-12 MHz. It is clear that the damper is vital for the injection process, and that the alternative filling scheme could not help for this aspect.


Stephane asked for the size of the emittance growth caused by a 10 degrees error. He also pointed out that the strength of the kick itself is of order 100 sigma. 1% kicker error translates into a 1 sigma oscillation.


Stephane or Elias asked for the optimum bunch train form. The answer was not obvious. Gianluigi inquired whether the kicker wave form was reproducible. At an LTC presentation from a long time ago, the kicker error had been supposed to be 0.5% and to only affect the first bunches.


Mike asked for how long we can live without the injection damper.  Wolfgang replied that 43 bunches probably will not require the damper, while 156 bunches may perhaps already need it. Certainly the damper is required for injecting trains of 72 bunches.


Frank asked for the sensitivity to the betatron tune. Wolfgang responded that this is about 1.5 times higher than the sensitivity to the phase error quoted earlier (the tune enters in the phase difference between two turns).


Mike pointed out that if the injection were stable, steering alone would suffice to avoid emittance growth. In this case parts of the damper commissioning could be postponed.


Gianluigi asked whether at the beginning of the beam commissioning we might need the damper as an exciter for beam measurements. He added a question about the minimum damper commissioning needed to do beam measurements. Wolfgang’s response was that if one only wanted to kick everything and did not worry about machine protection, one could hook up a signal generator.


Mike inquired whether there would be a connection from the tune PLL to the damper. Wolfgang answered yes, it had been foreseen to have such a connection, either on the surface or in the tunnel.  BI, which is pulling the cables, had opted for the surface. Wolfgang also clarified that the damper had no internal protection against unwanted signals.


Alick asked about the need for abort gap cleaning during commissioning. Mike reassuringly answered that the abort gap cleaning could also be commissioned later.


Gianluigi asked whether the damper had enough strength to compensate the kicker error within a single turn. Wolfgang replied that the damper can deflect the beam by 1 sigma over 4 turns at low frequency, but the kicker exhibits a larger ripple at a frequency around 10 MHz, where the damper strength is much less.


=> ACTION: Present an updated estimate of the emittance growth arising from the injection-kicker ripple to the LTC (Brennan) – After the meeting Brennan remarked that this point had already been a long-standing LTC action item and that an LTC presentation is foreseen for early 2008.


Stephane asked for the time needed to turn on or off the experimental solenoids. Mike replied about 5 hours.  In view of this time needed to switch on the experimental solenoids, Stephane expressed his doubts on the feasibility of the proposed measurement procedures. Indeed, even in a priori stable conditions (e.g. after many hours on the injection plateau), the coupling drift coming from the machine itself during the solenoid switching OFF/ON could be similar to, or even higher than, what is expected from the combined effect of all the solenoids (c_~0.007). After the meeting he further elaborated that in practice tune shifts of the order of 0.01 were observed everyday at LEP after a few 10s of minutes, possibly arising from small orbit variations, and LEP was a warm machine. For the solenoid compensation, if still deemed to be needed after further consideration, he would therefore recommend a local coupling measurement (all solenoids on), for instance a measurement of the driving terms followed by an identification of local sources of multipole errors.


LHC Operation Modes (Reyes) 

Reyes discussed LHC modes. She distinguished five types of modes: accelerator modes, beam modes, operation modes, access modes, and sector modes. Reyes’ presentation focused on the first three. She mentioned that the operation modes are sector based, indicating whether a sector is operational or not, which will ease the role-based access. The sector modes refer to a subset of accelerator modes, but are defined sector by sector.


Accelerator modes consist of modes with beam and modes without beam. Example acceleratpr modes are shutdown, cooldown, machine checkout (including dry runs), access, machine tests, calibration, warm up, recovery, sector dependent, beam setup (including beam commissioning), proton physics, and ion physics.


Beam modes comprise setup (possibly beam in transfer lines, but no beam in the ring), abort, injection probe beam, injection setup beam, injection physics beam, prepare beam, ramp, flat top, squeeze, adjust, stable beams, unstable beams, beam dump warning, beam dump, ramp down, cycling, recovery, inject & dump (for commissioning and injection studies), circulate & dump (no screens in the ring, beam dump via timing system), and no beam.


Alick asked for the length of time between a beam-dump warning and the actual beam dump, in particular whether the machine would wait for a confirmation from the experiments before dumping the beam. Mike replied that, yes, normally the accelerator would wait for a response from the experiments. Stefano asked if the modes start and change automatically. Reyes answered that the modes are changed either via the sequencer or manually, depending on the operator’s.  


Reyes next presented an example transition tree between beam modes.


The operation mode is introduced primarily to simplify the work of RBAC. It will therefore be sector based. 


Users of the LHC modes are the Safe Machine Parameters, the experiments, RBAC, various hardware, the LHC control software, the alarms, and the access system.


Distribution of the modes proceeds via four channels: high level publishing from LSA, DIP - the data interchange protocol for communication with the experiments, GMT – global machine timing (as SMP – safe machine parameters), and BST – the beam synchronous timing.


Alick asked whether the mode information is sent in parallel through all these channels, or whether only some will go to the GMT, BST and DIP e.g.. Mike responded that the modes will be distributed in parallel.


Among other modes feature the sector mode and the access mode.


A functional specification of LHC modes has been published as LHC-OP-ES-0005 rev 1.0, EDMS DOC 865811.


Alick asked whether during a sector test the accelerator mode would be sector dependent. Reyes replied in the affirmative, yes, indeed. Helmut commented that the mode name “calibration” was too general, since it referred only to the calibration of power converters. Gianluigi queried whether the accelerator modes are passed to the alarm system. Reyes answered yes.


Next Meeting

Tuesday December 18th, 14:00

CCC conference room 874/1-011


Provisional agenda


Minutes of previous meeting

Matters arising

Results, problems and lessons learned from

- TT40/60 extraction beam tests (Brennan Goddard)

- Transfer line beam tests (Jan Uythoven)




 Reported by Frank