Minutes of the ICE section
68th meeting on Wednesday 24/10/2012 (08:40-10:30, 6/2-004)
ICE members: Benoit Salvant (BS), Christian Hansen (CH), Carlo Zannini (CZ), Hugo Alistair Day (HD), Elena Benedetto (EB), Elias Metral (EM), Elena Wildner (EW), Frank Schmidt (FS), Giovanni Iadarola (GI), Giovanni Rumolo (GR), Jean-Luc Nougaret (JLN), Kevin Shing Bruce Li (KL), Nicolo Biancacci (NB), Nicolas Mounet (NM), Olav Ejner Berrig (OB), Serena Persichelli (SP), Tatiana Pieloni (TP), Werner Herr (WH), Xavier Buffat (XB).
Present/Excused: BS, CH, CZ, HD, EB, EM, EW, FS, GI, GR, JLN, KL, NB, NM, OB, SP, TP, WH, XB, Alessandro Vivoli, Alexander Molodozhentsev (from KEK), Alexej Grudiev, Alexey Burov (from FNAL), Andy Butterworth, Bettina Mikulec, Cedric Hernalsteens, Chandrashekhara Bhat (from FNAL), Christian Carli, Christine Vollinger, Daniel Valuch, Daniel Wollmann, Eirini Koukovini Platia, Elena Shaposhnikova, Fanouria Antoniou, Frank Zimmermann, Fritz Caspers, Gabriel Mueller, Gerad Tranquille, Gersende Prior, Giulia Papotti, Giuliano Franchetti (from GSI), Georges Trad, Hannes Bartosik, Harry Renshall, Haryo Sumowidagdo, Helmut Burkhardt, Humberto Maury Cuna, Javier Fernando Cardona, Ji Qiang (from LBNL), John Jowett, Kazuhito Ohmi (from KEK), Lajos Bojtar, Massimo Giovannozzi, Mathilde Favier, Michel Martini, Miriam Fiterrer, Octavio Dominguez Sanchez de la Blanca, Rama Calaga, Raymond Veness, Raymond Wasef, Riccardo de Maria, Roderik Bruce, Rogelio Tomas, Sandra Aumon, Simon White, Tatiana Libera Rijoff, Theodoros Argyropoulos, Tom Mertens, Uli Wienands (from SLAC), Valeri Lebedev (from FNAL), Vittorio Vaccaro (from Naples), Wolfgang Hofle, William Andreazza, Yannis Papaphilippou, Yuri Alexahin (from FNAL).
1) Newcomers / visitors
- None.
2) Comments on the minutes of the previous 67th meeting + Actions
- No particular comments on the previous meeting and on the minutes (not yet available).
3) General infos
- SL meeting:
- 2012 BE department meeting from PaulC will take place on Monday 3/12/12. Each group should present 6 transparencies => Please send me few slides / pictures of highlights for the different activities for 02/11/12. I will put them on some slides that I will post on the ICE site but I will send only 1 slide to OliverB.
- News from LHC:
- Many fills these days close to lumi record, i.e. > 7E33.
- The large bandwidth transverse damper has been used sometimes and also in collision it seems as they observed some decrease in the lumi lifetime (see for instance last Monday LHC 08:30 meeting):
- Info from GA this morning: even with the damper at high bandwidth we still see instabilities at the end of the squeeze (B1V). Wolfgang has increased the bandwidth but he has kept also the gain at low frequency constant or even higher. The damper pick-up data were saved for the fill 3212 => To be followed up.
- Landau octupoles at injection => Discussion with RogelioT and NM: it seems (from Rogelio et al.) that the DA at injection is at ~ 6 sigmas, i.e. close to the limit (as might start to be smaller than the mechanical aperture) and that it is dominated by the Landau octupoles (~ 6.5 A) used to avoid the transverse emittance blow-ups observed on some batches last year. Reminder: ~ 6.5 A in the Landau octupoles an injection creates an rms tune spread of ~ 2E-4 for ~ 2 microm (and it is linear with the transverse emittance) => It is not negligible and similar to what we have at high energy in fact. This value was never optimized until now as it worked and we had other issues (and furthermore since last year we changed the sign of the octupoles). From the theoretical and simulation point of view, this study was foreseen for next year in the framework of HL-LHC as this could be a potential worry with the future very high intensities. We agreed with RogelioT and NM to try and reduce this value "adiabatically" from one fill to the next. We could try 6 A in one fill, 5 A in the next if fine, and so on. This would be certainly very interesting => GA informed to see if we could do that in physics or if we need to ask for MDs.
- Very nice fill on FR evening: ~ 1.5 microm at injection and ~ 1.6E11 p/b, i.e. very nice beam from injectors and then it was really very nice in the LHC => ~ 7.4E33.
- Sunday: ~ 1.4 microm and had to wait for ~ hour and then emittance became ~ 1.75 microm. Is it H or V? Tbc.
- Sunday => Injection delayed a bit due to kicker temperature MKIC5.R8. Sudden heating up during fill. This re-increase was already seen on some other fills To be followed up => HD and BS informed.
- Tuesday => TOTEM dumped the beam due to BLM losses at 31 sigmas! It seems that all the BLMs went up exactly at ~ 30-31 sigmas where the TCL5 has a hole...
- 517 WE Heraeus-Seminar last week with talks in particular from EW and TP: https://indico.gsi.de/conferenceDisplay.py?confId=1910.
- BB2013 workshop => Open for registration and abstracts submission: https://indico.cern.ch/internalPage.py?pageId=2&confId=189544.
- IAR (Interim Activity Report) for HL-LHC WP2 => For beam-beam and for intensity limitations.
- PS dummy septum meeting on Friday => Very nice talk from SP.
- Meeting on Friday with Mario Deile + BS + one colleague for possible modifications of LHC TOTEM Roman pots during LS1 => They want to rotate them etc. => BS will try and give them some answers by the end of November.
- HDWG on Monday devoted to longitudinal studies by RF colleagues.
- At the LIS meeting on Monday, Vincenzo reviewed the simulation status for the PSB. I noted in particular that he uses a new way to import any longitudinal distribution => Thanks to S. Hancock's Mathematica script => Seems very nice and might be useful for other codes.
- There will be a seminar next Monday 29/10 from SashaM to review the space charge activity (meeting kept small with people directly involved on SC activities).
- CERN/PS LARP space charge collaboration started by SimonG and GR.
- Info from BS:
- For couple of weeks, there were some pbs with the LHC ADT data PU (just the viewer). This was a setting used for something else. There is now a new viewer which should work.
- Yesterday, test with RF: they blown up the bunches as soon as they were injected to reduce the IBS (initially done for Q20).
- Info from GR: It has been approved at the LMC that there will be MDs in the injectors in 2013 in PS and PSB and SPS at 26 GeV/c.
- Reminder: 11/02/13 is the end of the run (i.e. there will be only 4 weeks of operation in 2013 after 1 week of re-start).
4) Longitudinal to transverse Landau damping (Alexey Burov): pdf
- Trying to understand the huge difference (factor ~ 6) between 2 consecutive fills on MD block #3 with one beam only (B2) at flat top before the squeeze, with very similar conditions, AlexeyB looked at the effect of the longitudinal spread on the transverse instabilities of the LHC.
- AlexeyB considered first the effect of the longitudinal spread alone and then its effect on top of the (transverse) contribution from the Landau octupoles.
- It is known that the longitudinal spread (from the nonlinear RF bucket) can improve (only) the transverse stability (adding more spread, i.e. more stability) and this is why increasing the bunch length (by increasing the longitudinal emittance or reducing the RF voltage) can help and can explain more stability than expected. Larger tails in the longitudinal (and transverse) profiles can considerably increase the beam stability. But the new analysis of AlexeyB (in addition to a quantitative description of the above mechanism) also takes into account the Q" effect (introduced from the Landau octupoles). Reminder on the Q" contribution from the LHC Landau octupoles => For the current + polarity, it gives Qx = 8000 / 100 A and Qy = - 3300 / 100 A.
- The 2 main new findings are that:
- The 2 longitudinal spreads subtract in some cases and therefore increasing octupoles can make the situation worse! The effect of a small tail in the longitudinal profile was also studied, revealing its potential importance in explaining better stability than expected.
- For long bunches, the longitudinal incoherent spectrum is not monotonic anymore with respect to the longitudinal action which can lead to longitudinal stability issues (as in the longitudinal plane it is the derivative of the longitudinal distribution function which matters, whereas in the transverse plane it is the longitudinal distribution function itself).
- Note that on slide 4, the 2 distributions (the one with "normal" tails and the ones with larger tails) have the same HWHM (Half Width at Half Maximum) and might lead to the same measured bunch length in operation, whereas the behaviour will be very different. ElenaS mentioned that the HWHM in the LHC is 1.25 ns / 1.7 = 0.74 ns, which seems very close to what AlexeyB considered.
- ElenaS asked what will happen when potential-well distortion will be taken into account as this can also modify the picture => To be followed up.
- AlexeyB then considered the combined effect of the transverse spread from octupoles and the longitudinal spread, assuming 200 A in the octupoles => It is found that the for l = 2 it is better (the stability diagram is larger than with the transverse spread only) whereas for l = -2 it is close to transverse only as pure longitudinal contribution almost cancels => It seems (tbc) that there will be always a mode which will be more critical where the pure transverse case should be considered.
- SimonW commented that it seems that the transverse case only is the worst case (the longitudinal distribution adding only ore stability) and that in reality the beam is much more unstable than predicted => Something else is missing.
- Summary:
- It is clear that longitudinal distribution is important and that changes of it could explain the high volatility which might have been observed in some cases.
- Longitudinal to transverse Landau damping (LTLD) can be strongly deformed or eliminated by Q.
- High volatility of the longitudinal tails could be an explanation for the measured volatility of the thresholds.
- So far, LTLD looks both potentially powerful and fragile. Proper diagnostics and tomography are needed for any attempt to take it into account.
- AlexeyB finally proposed a research program described in slide 8:
- AlexeyB is asking for help to have the longitudinal distribution function for each bunch and for each fill.
- Etc.
- Reminder from ElenaS: Tomography is discussed every year and there are always pros and cons and in particular the cons is the number of bunches etc.
- Comment from ElenaS (during discussion about future tests with either shorter or longer bunches to study the RF heating): we should not decrease bunch length during dangerous parts of the cycle and do this only during physics.
- Email from AlexeyB after the meeting (that he wanted to be added to the minutes):
- Dear all, I am going to add something to a discussion followed an important objection of Simon at the end of my talk this morning.
Let me remind you first that I expressed a hypothesis that the instability at the end of the squeeze happens due to a loss of the longitudinal halo responsible for a significant part of the longitudinally-transverse Landau damping (LTLD).
Simon's objection was following. Loss of the halo cannot make situation worse than without that longitudinal halo at all. However, the thresholds computed without LTLD (my NHT computations in particular) agree with 2*nominal impedance, obtained from tune shift and growth rate measurements. Thus, we may conclude from this observation that LTLD does not play a significant role, and the loss of some longitudinal halo should not matter. Dear Simon, please correct me if I lost something in your idea.
As a counter-argument, I suggested an assumption that the impedance is not 2*nominal, but higher, say 4*nominal, and the stability is provided by half by a longitudinal halo. This counter-argument assumes huge ~100% error bar in the growth rate and tune shift measurements (well, may be the real impedance is not 4, but 3 times higher than nominal, so ~50% error bar - still would be sufficient for that argument).
At this point the discussion was over and later Nicolas expressed his intention to look one more time at these error bars.
Now I am coming to my new argument in defense of my LTLD hypothesis for the instability. Namely, this hypothesis does not actually need to assume that impedance exceeds 2*nominal. It still can be 2*nominal, but with the transverse tails of the injected beam shorter than Gaussian. The reason we see an agreement with Gaussian model of the transverse tails is that the deficiency is covered by the LTLD. When the longitudinal tails are lost at the squeeze, the instability is developed because the transverse tails are too short.
In other words, to overcome the seeming contradiction, I can assume shorter than Gaussian tails of the injected beam. In this case I can be in agreement even with smaller than 2*impedance. On top of that loss of longitudinal halo, additional loss of the transverse halo may also happen, of course.
If this explanation is correct, what remedy does it suggest against the instability? I think RF reduction should work positively, making longitudinal tails thicker. The extrema of the incoherent frequency I showed you this morning should not be actually that harmful as for the pure longitudinal case, since we still have some LD from the transverse degrees of freedom. Another promising idea - to go to even higher chroma, where l*dQs is big enough. Due to this factor, at high enough chroma we could work without octupoles (to be checked in the following computations).
5) Actions to be taken for the next meeting
- Old actions.
6) Miscellaneous
- Deadlines and important dates for ICE.
- The next (69th) meeting will take place on 31/10/2012 => Agenda:
1) Review of the beam-beam COMBI (COherent Multi-Bunch Interactions) code and plans to include the impedance (TP, XB and WH),
2) Observations of LHC beam instabilities during physics cycle: Comparison old / new octupoles' polarity, comparison old / new collision scheme, is the instability reproducible? When exactly does it happen? etc. (XB, TP, NM, etc.),
3) First comparison between AlexeyB's instability model and measured instabilities: do we miss something? What are the next steps / MDs? (AB, NM, XB, TP, SW, EM, etc).
- See preliminary agendas for the next meetings.
Minutes by E. Metral, 27/10/2012.