From: "Francesco Ruggiero" To: "Oliver Bruning" ; "Werner Herr" ; "Thys Risselada" ; "Frank Schmidt" ; "Daniel Schulte" ; "Frank Zimmermann" Cc: "Jean-Pierre Riunaud" Subject: Meeting on future LHC beam-beam studies: 14 July 2003 Date: Monday, July 14, 2003 8:05 PM ----------------------------------------------------------------------- Discussion of joint LCE-LOC tracking plans with beam-beam: 14.07.03 Present: OB, WH, FR, FS, FZ Excused: TR, DS ----------------------------------------------------------------------- FR starts the meeting by recalling that LHC tracking results with beam-beam are needed for the new LHC Design Report (deadline end of August) and for future discussions with the experiments to compare alternative crossing schemes (nominal H-V versus H-H or 45-135 degrees, possibly in conjuction with active compensation of long range collisions). The most recent tracking studies in LHC-PN-310 refer to LHC optics V6.2: meanwhile we have a (slightly) different crossing scheme and bunch filling scheme, and we should study dynamic aperture and working point for nominal as well as for Pacman bunches. Updated results for self-consistent orbits, tune spreads, and Q' on a bunch-by-bunch basis are available by WH. Possible strategies to investigate dynamic aperture and optimum working point include tune scans by fast tracking with lumped beam-beam elements followed by longer term investigations with element by element tracking. FS reports that an arbitrary transport matrix can be used in MAD and can be converted into Sixtrack for fast tracking with lumped elements. In connection with 45-135 crossing scheme, FZ says that betatron coupling can be cancelled by a proper choice of phase advance between IP1-IP5. OB says that the LOC priorities include optimization of the working point at top energy, scan of the phase advance between IP1 and IP5, and study of the tolerances for linear imperfections. TR will perform beam-beam tracking studies for the nominal crossing scheme with linear and nonlinear imperfections to optimize the working point. Many building blocks required for the tracking are already available (beta* knobs by WW, crossing angle and orbit correction by WH, triplet and D1 errors, etc,). Therefore TR may get results very soon, say by September-October (anyhow before end 2003). FS offers to assist with MAD-X and beam-beam tracking runs. FR says this is too late for the new LHC-DR, but it would be the right time scale for the forthcoming discussion on alternative crossing schemes and compensation strategies. Since the work plan of TR refers only to the nominal crossing scheme, the LCE team will name a person in charge of collaborating with TR and extending his results to other crossing schemes. ACTION -> LCE team: name a beam-beam tracker to collaborate with LOC. For each crossing scheme, it is agreed to perform: - 1D tune scans (parallel to the main diagonal), - by "fast" tracking over 10^5 turns, - including triplet nonlinearities, beam-beam, orbits, etc. - with ONE random seed (always the same), - for a nominal and a typical Pacman bunch. Once the best working point is determined for the nominal bunch: - repeat systematic tracking with 60 random seeds, - for nominal and Pacman bunches. The phase advance between IP1-IP5 should be optimized for each crossing scheme: this implies another scan. FR raises the question of injection oscillations and their effect on beam lifetime in case of poor dynamic aperture due to long range collisions. WH remarks that although dynamic aperture in collision is lower than at injection, increasing beta* is a powerful knob in case of problems. OB notes that even at injection, beta* and crossing angle can still be increased, although their effect is not as large as in collision. He also says that bringing the beams in collision should have a more drastic effect than injection oscillations. FR and FZ are not convinced of this. FZ remarks that injection oscillations may by transmitted to the other beam by the long range beam-beam. These considerations may require further follow-up, although the relatively fast damping of injection oscillations by the feedback system will probably reduce/solve these problems. Possible strategies to address beam-beam studies experimentally include more or less extended visits to Fermilab and RHIC. Francesco Francesco Ruggiero wrote: > > Dear colleagues, > > I propose to meet next Wednesday 16 July, say at 14:30 in room 6/2-004, > to discuss joint LCE-LOC tracking plans with beam-beam (see point 3 in > the enclosed LCE minutes). Can you please confirm your participation? > > Thank you > Francesco > > Francesco Ruggiero wrote: > > > > ---------------------------------------------------------------- > > Minutes of the ABP-LCE team meeting of 11.07.03 > > > > present: TdA, WH, AK, FR (secretary), HT, EV, LV > > excused: EM, DS, FZ > > ---------------------------------------------------------------- > > > > 1) Minutes and follow-up of pending actions > > ------------------------------------------- > > FR recalls that we have to produce a final draft of the Collective > > Effects chapter for the new LHC Design Report by 29 August. We therefore > > aim at a first preliminary draft by the end of July. At the moment there > > is only a (rather complete) draft of section 5.08 on Beam-Beam Effects > > by WH. > > WH, FR, LV, and FZ will meet next Tuesday afternoon to coordinate the > > different contributions. > > > > ACTION -> everybody > > > > STILL TO BE DONE > > > > ACTION -> HT will publish soon a comprehensive report on the SPS MKE > > kicker impedance, power deposition, and collimator impedance: a draft > > will be available before the end of the week (when he will leave CERN). > > > > DONE (see attachments 1 and 2). FR thanks HT for the excellent work done > > and for the interesting results obtained in such a short time. > > > > AK found a logical flaw in his test Mathematica and Fortran programs to > > check the fast resistive wall kick summation with discretized version of > > the convolution theorem: only preceeding bunches were included in the > > summation over previous turns. He will correct this and spend one day > > perusing the existing literature on digital signal processing, e.g. > > image compression techniques, where a similar approach is widely used. > > > > ACTION -> AK will use the new code, also including resistive wall with > > the inductive bypass effect (impedance is sufficient instead of wake?), > > resonator impedance and space charge, i) to clarify previous > > discrepancies with analytic estimates of resistive wall instability > > growth rates, by increasing the multi-turn memory for the wake > > calculation, ii) to study injection oscillations and emittance growth in > > the LHC including a model of the feedback system, iii) to analyze future > > SPS measurements in collaboration with H. Burkhardt and if possible iv) > > to interface his simulations of multi-bunch instabilities to the LHC > > impedance database. > > > > STILL TO BE DONE > > > > FR had raised the question of the LHC experimental background and > > possibly other effects associated with synchrotron radiation photons > > from one proton beam being back-scattered by the other proton beam. What > > is the cross section for this process and the associated high energy > > gamma flux? Do we need dedicated masks to reduce background in the > > experiment? > > > > ACTION -> FS+DS > > > > STILL TO BE DONE > > > > 2) Recent results on MKE kicker heating and LHC collimator impedance > > -------------------------------------------------------------------- > > For the MKE kicker there is no difference between the results of LV and > > HT, provided the same input data is used: in particular bunch spectrum. > > HT's model of the ferrite is the same as the one used by F. Caspers, but > > there may be some confusion when quoting power per meter or per magnet. > > With typical SPS conditions the estimated heating is 300-400 W/magnet. > > Lab measurements used to obtain a thermal model of the MKE kicker seem > > to disagree with the measured beam induced heating: the measured > > temperature rise and the thermal model indicate a power deposition much > > larger then estimated by LV and HT. The weak or missing piece of > > information seems to be the bunch spectrum (shorter bunches?). F. > > Caspers and the BT group are working to understand possible dynamic > > effects. There isn't much that the LCE team can still contribute. EV > > will analyze recent bunch spectra measured by the RF group. > > > > HT has successfully completed estimates of the LHC collimator resistive > > wall impedance by HFSS and by field matching. They agree with LV's > > estimates within a few %, even for Cu coated collimators. The effective > > impedances have large discrepancies as a consequence of different > > definitions: LV defines them such that tune shifts can be obtained by > > multiplying them by the average beam current, while HT has to multiply > > by the peak current (in agreement with standard definitions by Sacherer > > and Chao). However the final physical results, namely tune shifts and > > instability growth rates, are the same within 20%. This discrepancy is > > related to the use of mode power spectrum by Sacherer and Chao vs > > intensity spectrum by LV. > > > > ACTION -> LV will clarify 20% discrepancy of his approach compared to > > the standard approach by Sacherer and Chao. > > > > 3) Beam-beam studies > > -------------------- > > > > New LHC crossing scheme > > ----------------------- > > WH shows beam separations, bunch-by-bunch tunes and orbits with the new > > (nominal?) crossing scheme proposed by SF, which features nearly 0.5 mm > > transverse offsets at the IP's (see attachment-3). The relative beam > > separation is reduced by about 0.3 sigma beyond Q2 on both sides of the > > IP (if the maximum sigma of the two beams is used to normalize it). > > Using the sigma of the test beam, which is physically more relevant, the > > relative separation is reduced only on one side of the IP. > > WH underlines that the LHC experiments may not like the small offsets of > > the new crossing scheme and possibly react by questioning again also the > > alternating H-V crossing planes (since with H-H crossings the offsets > > would only be in the horizontal plane). > > > > ACTION -> FR will transmit this concern to the LHC Optics team and > > ensure > > that there is a common position of the ABP group, also in > > connection with recent statements by J.-P. Koutchouk at the > > last LTC meeting. See the following working plan discussion. > > > > New LHC filling scheme > > ---------------------- > > WH shows the new (nominal?) bunch filling scheme proposed by P. Collier, > > with the abort gap following 4 instead of 3 bunch trains. This new > > scheme and the corresponding correct tune comparison bunch-by-bunch for > > the old and new filling schemes are shown in attachment-4. There is > > essentially no difference compared to the old filling scheme, but orbit > > and tune patterns are now shifted to different bunch trains. > > > > After the meeting, WH mentions that the counting of missing head-on etc. > > is changed: this is entirely due to the loss of symmetry. > > > > 4 head-on 3 head-on 2 head-on HO IP2 HO IP8 > > old: 2553 253 3 -72 -186 > > new: 2385 360 63 -192 -294 > > > > Preliminary discussion of a working plan for beam-beam studies > > -------------------------------------------------------------- > > WH shows slides with a list of required beam-beam studies: those with > > green dots have already been done, those in red are still missing (see > > attachment 5: a list of proposed machine experiments is also included > > and should be actively followed-up). Among the latter there are new > > tracking studies according to a recipe discussed on p. 26 of LHC-PR 628: > > to compare different crossing schemes, we need bunch-to-bunch tune > > spreads and thus tracking for particles belonging to nominal and Pacman > > bunches. > > A possible strategy is to use fast tracking with lumped long range > > collisions for tune scans and then element-by-element long term tracking > > for a few selected tunes. FZ and FS could help with these two kinds of > > tracking studies, plus maybe DS and T. Risselada. > > > > ACTION -> WH+FR will organize a meeting next week (possibly on Wednesday > > 16 afternoon?) to launch and coordinate new beam-beam studies. > > > > ACTION -> FR will organize a dedicated meeting at the end of August or > > beginning of September to review all the available information > > on crossing and beam-beam compensation schemes, including the > > results of new beam-beam tracking studies. > > The conclusions should be then presented at an LTC meeting. > > > > WH remarks that alternating crossings at 45-135 degrees would certainly > > complicate LHC operation compared to H-V crossings. He also notes that > > the most severe problems are at injection, where the beams are already > > separated in both planes. > > FR thinks that we need to clarify quantitatively what "most severe > > problems" means: if the dynamic aperture at injection is strongly > > reduced by machine nonlinearities and long range beam-beam interactions, > > then we MUST include also the effect of injection oscillations! EV > > suggests to separate the beams also longitudinally by the two RF > > systems. However the total number of long range collisions (about 30 > > around each IP) would not be reduced. > > > > WH shows Q and Q' bunch-by-bunch variations for the effect of non-equal > > IP's on the beam-beam compensation by alternating H-V crossings: even > > for 15% differences in crossing angles or beta^*, there is still a good > > compensation and the tune variations are reduced to about 1/3 of the > > value for H-H crossings (these and possibly a few other slides are not > > included in the attachments). WH says that the H-V crossing scheme is > > also robust against bunch-to-bunch intensity and emittance variations: > > this is not the case for the H-H crossing scheme even with an active > > compensation system, unless the latter operates on a bunch-by-bunch > > basis. > > > > FR recalls the so-called "scallop effect" observed at the Tevatron, > > where Pacman bunches have better lifetimes than nominal bunches > > (presumably after extensive tune scans and optimization) and asks > > whether we can simulate or explain these observations. WH remarks that > > Pacman bunches suffer a lower number of beam-beam interactions, although > > they collide head-on with some transverse offsets and their tune > > footprints are more difficult to accommodate in the tune diagram for a > > machine optimized in view of the nominal bunches. At the moment it is > > impossible to perform strong-strong simulations with 3000x3000 bunches. > > B. Muratori has studied the case of 4x4 bunches with offset collisions > > and found no effect unless the beam-beam parameter is about 20 times > > larger than for the nominal LHC. This was indeed the case for LEP with > > bunch trains: LEP experience showed that bunches colliding with offsets > > as large as 0.5 sigma had a beam-beam limit 30% lower. A possible way to > > simulate more bunches is to reduce the number of macro-particles > > (leading to more numeric noise). Other creative self-consistent > > approaches would be welcome. No problem is currently anticipated with > > coherent beam-beam modes. FR mentions that we may get a new PhD student > > to work on beam-beam simulations. > > > > 4) Impedance interpolation routine for ZBASE > > -------------------------------------------- > > The wake interpolation routine by HT, which enables proper sampling of > > wide and narrow band impedance for wake summations, works correctly > > although the LHC impedance currently available in ZBASE is obsolete or > > wrong. Next week TdA will discuss with OB how to merge this routine with > > the recently modified version of ZBASE. > > > > 5) AOB > > ------ > > TdA informs the team that he has implemented a new MADX plotting command > > NOVERSION=TRUE to suppress MAD version number in a plot. FR remarks that > > version numbers in MAD plots are useful in case a bug is discovered at a > > later stage. WH says that the date of a published report represents a > > time-stamp to recover MAD version numbers. TdA remarks that version > > numbers are included by default and that the new command has to be > > explicitely used to suppress them. > > > > FR will prepare a proposal to measure collimator impedance at the SLAC > > ASSET test facility in collaboration with R. Assmann. > > > > Attachments > > ----------- > > 1. draft report on MKE kicker heating by HT, FC, and LV (HT-mkeheat.pdf) > > 2. draft report on collimator res. wall impedance by HT (HT-col_rw.pdf) > > 3. bb results by WH with new crossing scheme (WH-new-crossing.pdf) > > 4. new bunch filling scheme and coherent tunes by WH > > (WH-new-filling.pdf) > > 5. bb studies and experiments proposed by WH (WH-bb-studies.pdf) > > > > -- -- E-mail: Francesco.Ruggiero@cern.ch, Location: Building 9/1-008 Address: CERN, A&B Division, CH-1211 Geneva 23, Switzerland Telephone: +41 (22) 767 3726 or 767 5272, TeleFax: +41 (22) 783 0552 WWW: http://wwwslap.cern.ch/~rgo/