Summary notes of the twenty-third meeting of the LHC Commissioning Working Group

 

Tuesday April 10th, 14:00

CCC conference room 874/1-011

Persons present

 

Minutes of the Previous Meeting and Matters Arising

Jan made an important comment on the draft minutes of the 22nd meeting. He pointed out that the aperture kicker can only be used with the safe beam flag. Therefore, at 7 TeV this kicker cannot be operated at a bunch intensity of 3e10 protons, which had been proposed for improving the resolution of optics measurements. At 7 TeV the maximum safe-beam intensity corresponds to a single pilot bunch. This leaves the ac dipole as the only tool available for exciting the beam in 7-TeV optics measurements.

 

Beam Excitation: 1. Using the AC Dipole (Javier Serrano)

Javier described the ac dipole system for LHC, its technology and operational parameters. His talk was structured as follows: introduction, the key stakeholders, specifications, the proposed solution, center frequency choice, ongoing developments, outstanding issues and 2007 planning. He first recalled some fundamentals of ac dipoles, i.e. dipoles with an oscillating magnetic field. The oscillation frequency is chosen near the tune frequency modulo a revolution harmonic. An important advantage of an ac dipole compared with a traditional kick excitation is the preservation of the beam emittance if the ac dipole amplitude is ramped up and down adiabatically.

 

In 2006 AB-BT and AB-CO have agreed that the latter could use the MKQA kickers as ac dipoles. Relays are employed for selecting between three different generators driving the same magnets as aperture kicker, Q kicker or ac dipole. The ac dipole project was approved at the 62nd LTC Meeting held on 13.09.2006, which endorsed the proposal for a quick installation that is ready for the LHC commissioning. The group of key stakeholders includes Rogelio and Stephane (ABP) who determined the required power levels, Rudiger, Jorg and Jan for machine-protection aspects, Gene Vossenberg and Etienne Carlier ensuring generator compatibility, again Jorg for operational aspects, and various contacts in US-LARP.

 

A simple formula describes the integrated field strength required. It is proportional to the distance of the excitation frequency from the tune line, to the beam energy, and to the desired kick amplitude, and inversely proportional to the beta function at the ac dipole. At 450 GeV, a 7-sigma amplitude with a tune distance of 0.025 requires a peak current of 1733 A, corresponding to a kick per turn of 12 microrad. 

 

The assumed tune range is determined by the nominal tunes at injection and in collision, plus an additional distance of 0.025 on either side, resulting in a tunability range of 0.08 in tune units, or 900 Hz. The oscillation is realized with an RLC resonator, the resonant frequency of which is chosen at the center of the tuning range. If other tune ranges turn out to be of interest for the LHC, the capacitors will need to be exchanged. Javier added that studies on variable capacitors and inductances are presently ongoing at BNL. The specified excitation pattern of the ac dipole is a sine wave with trapezoidal envelope. The rise and fall times are 200 ms or longer.

 

The proposed solution is a parallel RCL circuit. Mathematically this is equivalent to a serial circuit. A circuit quality factor is defined in the standard way. The quality factor depends on the circuit parallel resistance, and it determines the power level required for covering a certain operating range in tune. The RCL circuit is driven by audio-amplifiers via transformers. Various companies offer audio-amplifiers with several kW power. The maximum power transfer is achieved for an optimum value of the parallel resistance. The transformer serves several functions: it transforms the resistance of the circuit into the optimum value for power transfer from the amplifier, by offering electric isolation it allows using the amplifiers in mono-bridge mode while the magnet is returned to ground, and it can couple the power of more than one amplifier to the same load. Javier explained that the transformer does not change the Q of the circuit, but that it just trades current for voltage.

 

The design procedure starts from the power needs and power rating. After measuring R_p and Q of the magnet in question, the power is computed from the desired kick strength in a first approximation. Then a parallel capacitance is chosen to give the target resonant frequency. Next, the current at 450 Hz from the peak is simulated, and the power requirement scaled accordingly. Amplifiers and transformers are then chosen to deliver this power. Four different excitation frequencies between 0 and 20 kHz are possible due to aliasing with the revolution harmonic. The scaling with frequency is not straightforward. Javier presented varying contradicting considerations. The optimum solution is determined by prototype measurements and by simulations using these test results. So far exploratory tests were performed at 2.9 kHz and 8.2 kHz (the second frequency results in about two times higher Q and four times higher R_p). Javier presented a few photos of the ac dipole test stand in building 867. The secondary transformer loop is arranged in serial topology for coupling several amplifiers, similar to arrangements employed at BNL and FNAL. The measured parameters are then used to simulate the magnet response to an excitation by two 13-kW amplifiers at the two frequencies. At a small distance to the tune 2.9 kHz offers lower losses, but the situation is opposite for excitation frequencies further away from the tune, where 8.2 kHz is preferred. The transition point is found at a tune distance of about 0.017. In any case, the difference in magnet response is not huge for these two frequencies.

 

Both at CERN and at FNAL it is planned to measure the actual behavior of two coupled amplifiers and compare the result with simulations. BNL is meanwhile developing variable capacitors and variable inductors. Javier listed a number of outstanding issues, including the heating of the relay, validating the current-to-field conversion factor, adjusting the amplitude specification to accommodate a possible beta beating of 15%, and coordination with machine protection.

 

The goal is to have four AC dipoles (two per ring) ready by the end of the 2007. A workable prototype is expected for the end of the summer, and the installation of the four final generators in November. HW/SW commissioning is planned for December.

 

Walter asked whether the excitation phase could be changed. Javier’s answer was yes, it can, but that no obvious reference phase exists, except possibly for the rf. Jan and Rhodri added that, in particular, there is no reference to incoherent oscillation frequencies of the beam. Jan asked whether the ramp-up and ramp-down rates are frozen. Javier answered that there is nothing fixing these rates. The only limitation on the high end may come from the power deposition in the relay. It was discussed whether, if excited on resonance, the beam could be lost instantly. Frank remarked that this is not the case, since a limit on the amplitude growth per turn is given by the maximum deflection angle which can be applied on each passage. The latter had been quoted as 12 microrad at injection energy. Jan suggested that this limit be calculated more precisely. Oliver pictured a scenario where the beam ends up on resonance at maximum excitation, due to the detuning with amplitude. He proposed to consider using the transverse damper as a safety element. Javier commented that presently the machine protection relies on the interlocked BPMs at the entrance to the dump line. Javier also clarified that the generator of the ac dipole is not the same as the one used for the other kickers. In particular, its strength is 3 times smaller and it takes 200 ms to reach the maximum value. Javier suggested that the failure mode to compare with is a failure of a D1 magnet. Oliver commented that the BLMs are distributed all around the ring and some of them should detect beam loss in case of a problem. Jan asked whether the ac dipole will also be used with unsafe beam. Oliver recommended that the possible interference of the ac dipole with orbit feedback and collimation should be looked at.

 

=> ACTION: Study interference of ac dipole with orbit feedback and collimation (Oliver)

 

Beam Excitation: 2. AC Dipole as Multi-Purpose Instrument (Rogelio Tomas and Stephane Fartoukh)

Rogelio’s talk was prepared in collaboration with Stephane. It addressed the following topics: basic principles, aperture measurement, linear optics measurement (beta beating and coupling), measurements of nonlinear driving terms, emittance preservation, adiabaticity condition, and summary. He presented the general response of the beam to a driven ac dipole excitation. The beam amplitude is inversely proportional to the distance between excitation frequency and betatron tune. A great benefit of the ac dipole is that it generates long-lasting coherent oscillation with sizable and tunable amplitude. The first ac dipole was developed at the BNL AGS. Later an ac dipole was also implemented in the CERN SPS using the transverse damper. Rogelio showed an example measurement from RHIC. The ac dipole is a non-destructive tool in principle.

 

The ac dipole can be used for aperture measurements. Its merits are the safe beam excitation, with no need to refill between measurements, and the possibility of simultaneous optics checks. In other words, the ac dipole allows for faster and better, more controlled measurements. Oliver suggested that the method might be insensitive at certain minima of the response. It was suggested that perhaps the measurement would need to be repeated at two different tunes to sample all phases. Discussion with Stephane after the meeting clarified that this is not the case, and that a single measurement is sensitive all around the ring. Jan asked whether the ac dipole measures the dynamic aperture. Rogelio replied that the dynamic aperture is not directly measured, but rather the mechanical aperture.

 

Rogelio next described linear optics measurements, namely how beta functions and betatron phase advance can be measured with the ac dipole. The simplest algorithm has an intrinsic measurement error of 3-5%. The FFT of the BPM signal in the plane orthogonal to the excitation contains the full information on linear coupling. Examples from the SPS illustrated a successful coupling measurement, and a disappointing measurement of the linear optics measurement. Excellent optics measurements with ac dipoles were obtained at RHIC. Also nonlinear and coupling resonance driving terms were measured at RHIC. The results were in good agreement with model calculations.

 

Rogelio now turned to the important issue of emittance preservation: The two biggest concerns are chromaticity and amplitude detuning. He explained that excitation on synchrotron sidebands will blow up the beam, and he presented as a rule of thumb that the excitation frequency should be separated from the main tune line by at least Q’/2 sidebands (distance > Q’/2 Q_s). He stressed that the excitation within the bunch frequency spectrum is forbidden also in presence of amplitude detuning.

 

Next Rogelio presented pessimistic simulations of ac dipole measurements at the LHC. At 7 TeV in store, with Q’=10, and a 5e-4 amplitude detuning at 3 sigma, the emittance growth at tune distances larger than 0.01 is acceptable. Clearly visible in the simulations is a higher emittance growth near each synchrotron sideband. At injection the situation can be more difficult. With a large amplitude excitation (4.7 sigma) the emittance can be blown up by orders of magnitude during a 2000 turn ramp.

 

Oliver asked whether the 3rd integer resonance should be avoided as excitation frequency. Rogelio replied that according to the ac dipole theory exciting on the 3rd integer resonance is not a problem. He emphasized that the ac dipole can indeed be safely operated on many resonances. There are certain resonances which must be avoided, however. Oliver suggested that the ac dipole may offer a new approach for safely crossing some resonances. 

 

Exciting 7 sigma oscillations at injection requires Q’<7 and an amplitude detuning at 3 sigma lower than 1e-3. The benefits of using the ac-dipole for aperture measurement and optimisation in parallel, in a non-destructive way, would require a control of Q' of the order of +/- 10 units. Stefano commented that in view of these preconditions the ac dipole cannot be used early on in commissioning.  Frank asked whether we will for sure have all 4 ac dipoles at hand in the LHC commissioning. The answer by Javier was yes. Frank also reiterated that presently no means other than the ac dipole is known to measure the optics at 7 TeV.

 

Rogelio concluded his presentation by listing other ac-dipole applications, e.g. for measuring the chromaticity Q’, and the detuning with amplitude, dQ/J. The underlying idea is a measurement of the head-tail phase shift inside the bunch using a tiny ac excitation at the betatron tune. More details on these advanced measurements can be found in a report by Stephane [‘A Theory of the Beam Transfer Function (BTF) with Chromaticity Induced Head-Tail Phase Shift’, LHC Project Report 986].

 

Stefano cautioned that the ac-dipole measurement was said to be non-destructive, but he pointed out that for aperture measurements scraping will occur at some amplitude, so that perhaps the total gain in time is not that large.  Rogelio replied that the advantage of the ac dipole lies in the capability of increasing the excitation amplitude little by little, such that always the same beam can be used until the aperture limit is found. It was nevertheless conceded that the advantages of the ac dipole for aperture measurements may not be dramatic.

 

Beam Excitation: 3. Aperture Kicker (Frank S.)

Frank S.’s talk addressed the question whether we have an aperture kicker at the LHC. He covered kicker hardware, kick classification, examples, the expected dynamic aperture at injection and in collision, and conclusions. A photo proved the existence of the kicker and of the MKA/MKQ pulse generators. The circuit diagrams for the latter two were also presented. The kick voltage is limited to 900 V limit at the moment. The mode of excitation of each kicker module is selected by turning “keys” in the CCC. The 0 position corresponds to the Q kicker, the other two to aperture kicker and to ac dipole.  

 

Roger asked about the master of the keys. The answer was not clear. Jan expressed surprise, recalling it had earlier been agreed that the keys should be located in point 4. There was a consensus that OP and not BT should be responsible for the keys. Rhodri asked whether the kicker is interlocked to the safe beam flag. Jan explained that if the safe beam flag is based on two measurements of energy and intensity it should be 'safe'. If instead it is based on single measurements only (which could be the case) he considers it as 'unsafe'. Another question is when the keys are actually turned - only before using the kicker or earlier. Frank S. had discussed this issue with Rudiger, who did not know the answer. This complex of questions is to be followed up by the MWSWG.

 

=> ACTION: Follow up MP issues of ac dipole and aperture kicker (Jan)

 

The kick classification distinguishes between basic optics checks, requiring kicks below 2 sigma provided by the tune kicker, detailed optics measurement with kicks of 2-6 sigma, to be obtained from the aperture kicker or ac dipole, and dynamic aperture studies, calling for kicks above 6 sigma.  As an illustration for the usefulness of kicked-beam studies, Frank S. presented measurements of linear coupling and sextupolar resonance driving terms performed by R. Tomas and himself in the SPS, and the compensation of a skew sextupole resonance by P. Urschuetz and M. Benedikt in the PS booster. He also recalled SPS dynamic aperture studies by J. Gareyte, A. Hilaire, and himself, from around 1988/89, revealing particle loss at the chaotic border for different detunings with amplitude.  At injection the expected dynamic aperture of the LHC is around 11 sigma without beam-beam, and near 7 sigma with beam-beam. The only device which can provide oscillation amplitudes of this magnitude is the ac dipole. For example with a tune distance of 0.025 it could drive a 7-sigma oscillation. After the meeting Frank commented that since the ac dipole involves a continuous excitation, the beam oscillation is different from that generated by a single kick, which might affect the “apparent” dynamic aperture.  Frank S. next reviewed the dynamic aperture in collision, which is predicted to vary between 9 and 15 sigma without beam-beam and to shrink to about 5.5 sigma with beam-beam. At a tune distance of 0.01 the ac dipole can reach a maximum amplitude of 4 sigma at 7 TeV. On the other hand, the aperture kicker can provide a maximum kick amplitude of 5 sigma at injection and of only 1.4 sigma in collision.

 

Frank S. concluded that the ac dipole is the only tool capable of probing the dynamic aperture at least at injection. He pointed out that, if ever the need arises and machine protection permitting, we could increase the kick strength of the MKA aperture kicker, by upgrading the MKA power supply from 890V to 4 kV. This should result in a 22.5 sigma kick at injection and 6.3 sigma kick in collision. The only hardware change needed is at the ‘back’ of the pulse generator.

 

Jan recalled an earlier review of MPS issues which determined that the higher MKA voltage was too dangerous, as the “safe beam flag” is not really safe.

 

Hardware Commissioning: Vacuum System (Frank)

Frank discussed the commissioning with beam of the vacuum system. His presentation was prepared in collaboration with Gianluigi and Ralph. Most of the information was provided by Miguel Jimenez. The talk addressed the following topics: vacuum requirements, LHC vacuum system, pressure measurements, signals available in the control room, bake-out and NEG activation, 450-GeV vacuum pressure and beam lifetime, collimation vacuum, e-cloud test bench, effect of quenches on beam and insulation vacuum, beam operation at 450 GeV and beam measurements, higher beam intensity, 7 TeV operation, and He leaks.

 

The LHC design pressure corresponds to 100 beam lifetime from nuclear interactions with the residual gas, which translates into an average loss rate of 6e4 p/m/s to be compared with a local quench limit of 8e6 p/m/s at 7 TeV.  The emittance growth time from multiple Coulomb scattering is 10-30 h at injection, depending on the gas species. The LHC vacuum system consists of sectors separated by valves. The experimental IRs accommodate 5 cold vacuum sectors on each side of the IP: the triplet, Q4, Q5, Q6 and the arc (2900 m long). The intermediate regions and the space around the IP are warm vacuum sectors. By contrast, in IR7 there are 3.5 warm vacuum sectors between the IP and the cold arc. The pressure in each sector is measured by gauges. Two types of gauges are used: Penning gauges and Bayard-Alpert gauges. Penning gauges are cold-cathode gauges, consisting of two electrodes and a permanent magnetic field. They were invented in 1937 by Penning, and are a precursor of the sputter-ion pump. They are characterized by a nonlinear dependence, and are considered to be “more stable, but less precise”. Bayard-Alpert gauges are hot-cathode gauges, consisting of three electrodes, a filament, a collector and a grid. Their invention in 1950 brought about a revolution in vacuum technology. Their current shows a linear dependence on the pressure. Bayard-Alpert gauges are suited for ultra-low vacuum measurements, and they are considered as “less stable, but more precise”.

 

Oliver asked in which sense the Penning gauges are “more stable”. Frank replied that he was quoting Miguel Jimenez.

 

The pressure in each cold vacuum sector is measured by Penning gauges, which are located in an attached short room-temperature section, about 20 cm from the cold region. As a consequence there is no simple rule to deduce from the gauge reading the actual pressure in the cold part.  Warm sectors are equipped with both types of gauges. There are at least two Penning gauges per sector (warm or cold), and 6-8 gauges per arc (every 3 or 4 cells). Each warm sector features an additional Bayard-Alpert gauge. The insulation vacuum is monitored by Penning gauges with auto-protection, i.e., these gauges switch off for pressures above 5e-5 torr.

 

Frank showed two overall vacuum layouts of the LHC, which had been provided by M. Jimenez, and two photos of Penning gauges and a vacuum valve taken during the CIC/EIC tour of Point 8.

 

Available in the control room will be all pressure gauge readings at a sampling rate of 0.1 Hz. Locally the sampling frequency can be increased to 1 Hz during MD periods. The gauge readings are accessed via the PVSS vacuum interface. Frank presented some typical LHC vacuum pages in PVSS. Also vacuum alarms are sent to the control room. The naming conventions for vacuum equipment were explained in a separate email by M. Jimenez. For example the device VGP.39.B5R8.C refers to a Penning gauge (VGP), located in a combined vacuum sector (.C instead of .B or .R for blue and red beams, respectively), in vacuum sector B5R8. The 39 indicates that the gauges is located at a distance of 39 m from the sector valve when leaving the IP.   

   

The vacuum is interlocked with the BIS and LBDS, e.g., valve movement or pressure above threshold lead to a beam dump. Conversely, valves can only be closed it there is no beam circulating. The vacuum-interlock aspects were discussed by Laurette and Bruno Puccio at the MPSWG on 14.02.2007.

 

Presently it is foreseen to bake out all room-temperature vacuum chambers of the LHC. At the time of discussion with Miguel, there was still a question mark concerning the ceramic chambers MKD in IR6 L&R due to scheduling conflict. Jan commented that these sections will definitely be baked.

 

Massimiliano asked how, if possible at all, the state of the beam pipe surface all around the ring will be tracked (activated NEG, saturated NEG, oxidized ... ?), as this has consequences for emission/desorption effects and comparison to simulations.

 

Concerning the collimation vacuum, the following information is available. The initial total outgassing rate of the collimators is 2e-7 mbar l/s. With a pumping speed of 20 l/s, the pressure should be below 1e-8 mbar. The pressure will decrease by a factor of 10 after one year. Sudden beam loss at the collimators can lead to a pressure spike due to beam-loss induced desorption. If the local pressure rises above 5e-5 mbar the valves will close. The pressure recovery time after a spike is determined by the pumping speed. The beam loss will also increase the temperature of the collimator jaw, giving rise to enhanced thermal stimulated desorption. The latter will persist for a longer time, but the resulting pressure increase is much less than the one resulting directly from the beam impact.  

 

Oliver inquired on the time scale and procedures for the vacuum recovery after a pressure spike due to beam loss. This information must be obtained from Miguel Jimenez.

 

Frank reported that the installation of the LHC e-cloud test bench is postponed to the 2008/09 shut down. Roger asked in which sector the e-cloud test bench will be installed. Frank answered that a location in IR4 is foreseen according to the EDMS document LHC-VI-EC-0001 v.1.0.

 

Next Frank discussed the effect of a quench. Its consequences depend on the temperature transient of the cold bore during the quench. If the peak temperature stays below 30 K the effect is small, with only H2 molecules being released. At higher temperatures, also CO and CH4 will be liberated. Frank presented the result of a quench test at string-2, where the temperature of one magnet exceeded 30 K (see E. Blanco-Vinuela et al, “Experimental Validation and Operation of the LHC Test String 2 Cryogenic System”, LHC Project Report 681; and V. Baglin, “Vacuum Transients During LHC Operation,” Chamonix XIII).

 

Stefano asked for the implication of a temperature rise above 30 K. Laurette responded that the interlock “vacuum OK” may be triggered for the cryogenics system.  Jan asked for a confirmation that the beam vacuum is interlocked to the cryogenics. 

 

Also time scales and procedures for the vacuum recovery after a quench would be of interest. 

 

Concerning beam operation and beam measurements, no pressure change is expected for the 450-GeV commissioning beams. No special beam measurements are needed accordingly. Of course, the vacuum group will monitor the pressure readings. The predicted pressure evolution in the first year of LHC operation is superb. The initial pressure will correspond to a beam lifetime of about 300 h, and it will constantly improve with time (V. Baglin, “Running In – Commissioning with Beam,” Chamonix XII).

 

According to Mike and Jan, the reason why no vacuum issues are expected with beam is the careful preparation during checkout before the first beam injection. Jan elaborated that the good vacuum is established thanks to cryo-pumping in the cold sectors and bake out in the warm sectors.

 

At 7 TeV and with higher intensity new effects come into play. Synchrotron-radiation photons can lead to enhanced desorption. And the heating induced by secondary halo or nuclear cascade behind the collimators can degrade the vacuum in the cleaning insertions.

 

The last item Frank which discussed was the issue of helium leaks. Localized He leaks can cause quenches, without any visible effect on beam lifetime or emittance. The helium front propagates very slowly, with a speed of the order of a few cm / h, and it may take weeks or months before the front reaches the next pressure gauges. Three solutions were proposed in the past: (1) mobile BLM “snake” [B. Jeanneret], (2) debunching the beam and measuring the ionization current [A. Poncet, CERN MT/95-01 (ESH)], (3) an online measurement of the heat load to the cryogenics [V. Baglin, “How to Deal with Leaks in the LHC Beam Vacuum,” Chamonix XIV).

 

Frank concluded that many vacuum readings are available. Nothing special is expected for the commissioning. The beam lifetime should be ok. At higher intensity and energy, higher steady-state pressure may be observed behind the collimators. Sudden beam loss on collimators could cause pressure spikes which may lead to valve closures. Desorption by synchrotron-radiation photons may also be observed. A list of references was finally presented. 

 

A number of questions were left unanswered and require feedback from the vacuum group.

 

=> ACTION: Follow-up open questions on vacuum system (Frank)

 

Next Meeting

Tuesday April 24th, 14:00

CCC conference room 874/1-011

 

Provisional agenda

 

Minutes of previous meeting

Matters arising

Commissioning of accelerator system - BLM (Laurette)

Combined use of RadMons and BLMs (Thijs)

Commissioning procedures for the ramp (Mike)

AOB

 

 

 Reported by Frank