Minutes of the ICE section
6th meeting on Wednesday 01/09/2010 (08:40-10:30, 6/2-004)
ICE members: Benoit Salvant (BS), Christian Hansen (CH), Carlo Zannini (CZ), Diego Quatraro (DQ), Hugo Alistair Day (HD), Elena Benedetto (EB), Ewen Hamish Maclean (EHM), Elias Metral (EM), Elena Wildner (EW), Frank Schmidt (FS), Giovanni Rumolo (GR), Jean-Luc Nougaret (JLN), Kevin Shing Bruce Li (KL), Maria Carmen Alabau Pons (MCAP), Nicolo Biancacci (NB), Nicolas Mounet (NM), Olav Ejner Berrig (OB), Tatiana Pieloni (TP), Werner Herr (WH).
Present/Excused: BS, CH, CZ, DQ, HD, EB, EHM, EM, EW, FS, GR, JLN, KL, MCAP, NB, NM, OB, TP, WH + Elena Shaposhnikova + Theodoros Argyropoulos + Alexej Grudiev + Fritz Caspers + Valeri Lebedev (from Fermilab) + Chandrashekhara Bhat (from Fermilab)
1) Newcomers / visitors
- Valeri Lebedev from Fermilab is at CERN for few days. The visit of Valeri is related to (for the 50% paid by ABP) investigations on possibility to damp the hump in the LHC with the transverse feedback taken into account the noise of the system and its bandwidth by model and data analysis. He will also give a talk at the next ICE section meeting on the Tevatron luminosity evolution model and its application to the LHC.
2) Comments on the minutes of the 5th meeting + Actions + Follow-up
- Comments: None.
- List of Actions. Some results on the LHC multi-bunch modes at 7 TeV/c have been obtained by NM and should have been discussed at the end of the meeting, but it was finally not the case due to time constraint. The slides are however linked here and could be discussed at another ICE section meeting.
- Follow-up of the PSB instabilities by Giovanni (see previous talk on 04/08/10) => Slides presented at the MSWG held on 27/08/10:
- There was another MD session to try and understand the discrepancy between the sum and delta signals as on measurements done in the past the bunch length seemed to be ~ 2 times smaller in the delta signal.
- The 3rd instability is not observed anymore as the gain of the phase loop between the cavities has been reduced (non ppm hardware change) and the gain of the TFB has been increased. The 1st and 2nd instability are still there.
- For the 1st instability, the two puzzling observations were that: (1) the number of "nodes" depends on the intensity (which is not predicted by the usual head-tail instability formalism) and (2) the bunch length is ~ 600 ns and the delta signal extends only to ~ 300 ns, i.e. about half of the bunch length.
- The last MD revealed in fact that if we look at a later time the delta signal eventually extends to the full bunch length => This observation, together with the (1) above, seem to indicate that the observed "nodes" are not really nodes of a (slow) head-tail instability and that the instability observed is a strong head-tail (or TMCI) one, as also mentioned by Valeri Lebedev when looking to these pictures. Reminder: The best way to check if one has a head-tail instability or not is to superimpose a few number of consecutive turns (say ~ 10) to clearly see the nodes of a standing-wave pattern. If on the contrary there is a travelling-wave pattern propagating from the head to the tail, there is no clear node and therefore no clear head-tail modes: in fact this is due to the coupling of different modes and which reveals a TMCI. However, one has to be careful as even in this regime if we superimpose many traces we could have the impression of having nodes but it is not the case! See for instance Figure 2 of http://accelconf.web.cern.ch/AccelConf/e06/Pre-Press/THPCH057.pdf: one could think that one has one node (and that this instability is therefore a head-tail instability |m| = 1) but it is not the case as it is a TMCI! See also https://ab-div.web.cern.ch/ab-div/Meetings/APC/2006/apc060202/E_Metral_slides.ppt for more details.
3) General infos
- No particular comment from anyone.
- SL meeting last week: I was not there.
- News on the LHC:
- The peak luminosity reached 1E31 cm-2s-1 when using 50 bunches against 50 bunches. Next step after the current Technical Stop is to go to bunch trains with the 150 ns bunch spacing. Note that it is currently planned (Mike Lamont, FOM, 31/08/10) to use only this beam until the end of the year.
- Today there will be a discussion with WH on the crossing angles to be used.
- Losses are observed during Q' measurement at top energy => Info from Stephane Fartoukh: 50 Hz, i.e. about dp=5E-4 looks a bit too much for chromaticity measurement in the presence of octupoles. Indeed with 100A in the MO we get a Q'' as high of 10 000 in the H plane (possibly even slightly higher due to the triplet for beta*=3.5 m in 4 IR's). Therefore if we are unlucky with the signs and the on momentum chromaticity is already small or even slightly negative, the off-momentum chromaticity can be as negative as -5 => Action 1: BS will perform HEADTAIL simulations to try and reproduce these losses and see if this can be explained by a head-tail instability.
- Info from NM => Cedric Villani (a french guy from ENSParis) won the Fields medal on nonlinear Landau damping: http://www.icm2010.org.in/imu-prizes/prize-winners-2010/fields-medal-cedric-villani.
- Info from Frank Zimmermann: "Dust" particles can produce observed LHC loss signature (rate & duration) => See IPAC10 paper on "dust" in the LHC by FZ et al. (http://accelconf.web.cern.ch/AccelConf/IPAC10/papers/mopec016.pdf), where they describe the relevant electric forces, charging rate, and beam lifetime, for round metallic macroparticles interacting with the LHC beam. First conclusion: the nominal LHC beam cannot pick up any "dust" particles from the bottom of the chamber (a lower current beam can do so even less). More interesting is the case of a dust particle falling down onto the beam from above. For the mass range considered in the IPAC paper, A~10^12 to 10^14 proton masses, the particles charges up so quickly that it is repelled by the beam field before the proton loss rate comes anywhere near the quench limit. FZ found a mistake for two curves in Figure 5 of the IPAC paper. The loss rate for A=10^12 shown there is correct, but the loss rate does grow for higher masses. Upon closer inspection at a much larger mass of 10^16 the peak loss rate increase so much that it is close to the quench limit of about 10^7 protons per second. Larger particles will lead to a quench when they fall onto the nominal beam. At the present much lower beam current, already macro-particles with masses of A~10^14 proton masses (instead of 10^16 at nominal current) will lead to loss rates near the quench limit. Note that in all cases the time duration of the high losses is of the order of, or less than, a few ms, very similar to the signature seen in the BLM data! All this seems to suggest that we might well see large macro or "dust" particles falling into the beam from above.
- Info from Mike Barnes: Mauro & Christina measured a SEY of a piece of ferrite (as used in the kickers) and they found quite a high value: ~ 2.5! But the ferrite was not baked out immediately before the measurements...
- Info from Volker Mertens and Miguel Jimenez: All the BPMs are isolated by enameled flanges (one in each side) and an additional layer of vetronite. In the procedure it is clearly written that these flanges shall stay isolated (electrically tests by Vacuum Crews) after assembly. What needs to be checked is the presence or not of the RF bypasses. Reminder: These enameled flanges break the (DC) electrical conductivity of the beam pipe. Similar things exist in the PS and, as a remedy RF bypasses have been installed.
4) First measurements of longitudinal impedance and single-bunch effects in the LHC (Elena Chapochnikova): ppt
- These observations were done during operation and not during dedicated MDs, therefore they were not studied in detail.
- Reminder on the longitudinal emittance of nominal beam in LHC Design Report:
- 0.7 eVs (inj.) → 1 eVs (after filamentation) at 450 GeV/c.
- 2.5 eVs at 7 TeV/c (controlled emit. blow-up during ramp)
- IBS growth rates in longitudinal and transverse planes were taken into account for the value of 2.5 eVs at 7 TeV/c. Furthermore, for the longitudinal stability, to have the same thresholds as at 450 GeV/c with 0.7 eVs and as at 7 TeV/c with 2.5 eVs, it was proposed to use 1.75 eVs at 3.5 TeV/c. An emittance of 0.6 eVs from the SPS is a good compromise to minimize the losses on the FB and ease the emittance blow-up in LHC.
- Before the longitudinal emittance blow-up performed in both the SPS and the LHC, a loss of longitudinal Landau damping could be observed during the ramp (at ~ 1.8 TeV/c) with a bunch length of ~ 1 ns (i.e. a longitudinal emittance of ~ 0.35 eVs). It was also observed at flat top: B1 was unstable with 0.65 ns (corresponding to 0.5 eVs) and B2 was less unstable but still unstable with 0.72 ns (corresponding to 0.6 eVs).
- The Loss of Landau damping leads to undamped bunch oscillations: quadrupolar in our case (the dipolar oscillations are damped by the phase loop). An estimate of the longitudinal impedance deduced from this Loss of Landau damping is Im(Zl/n)eff,m=2 ~ 0.09 Ohm. In the LHC Design Report, Im(Zl/n)eff ~ 0.07 at 450 GeV/c and ~ 0.076 at 3.5 TeV/c. However, the resistive part from the collimators was not taken into account. With out current impedance model (including all the collimators) the effective impedance (m=1) is 0.086 Ohm => See page 16 of https://impedance.web.cern.ch/impedance/documents/TransverseCoherentTuneShiftsInTheLHCandSBStabilityLimits.pdf, which is ~ the same value for m=2.
- If we have a constant bunch length (to have a constant filling factor) we should not lose Landau damping.
- On the flat bottom, fast longitudinal emittance growths were observed for bunches with nominal intensity but small longitudinal emittance (corresponding to bunch lengths of ~ 1.1-1.2 ns) => Agreement with IBS calculations by Frank Zimmermann and Johh Jowett to be checked. When the longitudinal emittance was increased (with a corresponding bunch length of ~ 1.7 n), almost no longitudinal emittance blow up could be observed but losses were then observed on FBCT and not BCT, meaning that some particles are leaving the RF bucket.
5) A HEADTAIL analysis tool (KL): pdf
- KL presented a summary of his activities on HEADTAIL development, HEADTAIL LSF submission tools and PYHEANA (a PYthon HEadtail ANAlysis tool).
- HEADTAIL development:
- KL proposed a more professional file structure for future developments on the very important HEADTAIL code, clearly separating the objects and the variables, which both have attributes => Easier to maintain and understand in particular for newcomers.
- KL proposed to "Include nrutils.h" => This avoids redundant definitions of standard routines.
- The input file is now passed as argument => This makes the execution and scripting slightly more compact.
- This scheme has many advantages (for instance the reduced file sizes etc.) but one drawback: one must know the location of the subroutines, which forces a logical organization.
- HEADTAIL LSF submission tools:
- SUB-HDTL/SUB-ECLD => Fast and simple script that makes use of LSF job arrays (quasi-parallel execution of jobs).
- Required user input: filename, scan variable and scan range.
- PYHEANA (a PYthon HEadtail ANAlysis tool):
- KL developed a user panel.
- He applied the tool to PSB impedance studies and SPS ecloud studies obtaining a lot of very interesting results, such as head-tail modes vs. intensity (to be compared to coherent tune shifts measurements performed in the past), ecloud coherent and incoherent tune shifts etc., which deserve careful analyses.
6) Highlights from GSI workshop (BS): ppt
- The link to the workshop is: https://indico.gsi.de/conferenceDisplay.py?confId=1031.
- BS discussed few interesting talks made during the 2-day workshop in July 6-7, 2010 and made the following remarks (which I commented in red):
- Possibility to invite Lukas, Uwe, Oliver to our impedance meetings via webcast => Good idea. We could perhaps organize impedance meetings every 1-2 months with them via webcast => Action 2: BS will organize these meetings with Hugo and Carlo (and other interested people).
- Al Khateeb formula works at low frequency only with a non trivial Fortran code from Reiner Hasse or Tom? We provided our Rewall code to them. 1st task for them is to check with Al Khateebs formula.
- Oliver mentioned that it was great that CERN had so many people involved in impedance team .
- Oliver said it was a good idea to have the RF people strongly involved in longitudinal instabilities as the hardware (phase loop, imperfections) is so important. In this frame, having the longitudinal collective effects within the RF group is an asset.
- Oliver would like that Vladimir shares his results with us before HB2010 => I think nobody from our team was contacted yet.
- Oliver mentioned that he would not use HEADTAIL for longitudinal simulations. He uses his own code LOBO. He says the matching of the bunch to the bucket is critical and this seems to be true. Opportunity to work towards a new code? => It is a good idea: We (CERN and Frank Zimmermann for his code repository: https://oraweb.cern.ch/pls/hhh/code_website.startup) are still waiting for his code LOBO (source + user guide if any).
- They are debating the need for collimation in SIS100. Effect of ions? Oliver proposed to Ralph that some tests can be performed at GSI for ion collimation, but no answer.
- New diagnostics at SIS18 for longitudinal dynamics (Olexander Chorniy).
- Advice from Wolfgang Muller (TU):
- There should be no tetrahedral mesh for wakefield solver.
- Japanese PhD student tries to make I-solver work with wakefield solver and it seems there is not much gain yet compared to the classical TD solver.
- Very large and easy improvement when using this GPU acceleration method (more than with a cluster!!!!). Maybe useful for Carlo and Hugo? We should ask for a cost estimate and do a test with one input file.
- 64 bits OS!!!!!!
7) Actions to be taken for the next meeting
- Old actions.
- New actions:
- Action 1: BS will perform HEADTAIL simulations to try and reproduce the losses observed in the LHC at 3.5 TeV/c during chromaticity measurements and see if this can be explained by a head-tail instability.
- Action 2: BS will organize "impedance meetings" with Carlo and Hugo (and other interested people) via webcast with our impedance colleagues from GSI and TU Darmstadt.
8) Miscellaneous
- The next (7th) meeting will take place on 08/09/2010 => Agenda:
1) Transverse impedance localization (NB)
2) Tevatron luminosity evolution model and its application to the LHC (Valeri Lebedev from Fermilab)
- See preliminary agendas for the next meetings.
Minutes by E. Metral, 07/09/2010.