Minutes of the ICE section

5th meeting on Wednesday 11/08/2010 (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 + Alexej Grudiev + Emanuele Laface + Fritz Caspers + Helmut Burkhardt + Kazuhito Ohmi (KEK) + Rama Calaga + Rogelio Tomas + Theodoros Argyropoulos + Frank Zimmermann.

1) Visitor

- Frank Zimmermann invited Kazuhito Ohmi (KEK) to participate to this ICE meeting (good idea!). Kazuhito is at CERN for one month and he is studying in particular strong-strong beam-beam effects with his own code. I hope we could continue and collaborate with him in the future.

2) Comments on the minutes of the 4th meeting + Actions

- Comments: None.

- Actions: Still List of Actions.

3) General infos

- OB has been working on the problem with installing windows 7, in order to have more memory for the CST simulations. After many iterations with helpdesk, he now has a windows 7 - 64 bit running on a virtual machine. He got that via: https://winservices.web.cern.ch/winservices/Services/CVI/Request.aspx Unfortunately he sees now that CST Particle Studio is not centrally managed on 64 bit machines. It is not worth while to install a 32 bit version, as it has only a memory span of 4 GB. He has now asked the manager to install CST on 64 bit machines. The virtual machine can be set up so that everyone from the section can use it simultaneously. We might have to request more than one virtual machine. We can also ask to add more memory, e.g. 18 GB like they use in CST/Darmstadt. There will probably be a cost to using the virtual machines permanently, but it will certainly be cheaper than getting one machine each. Also OB has other programs that he is not sure can work on 64 bit systems. CST have now acknowledged that there is a real problem with the open boundary. They suggest that OB should use ports with many modes enabled, in the meantime. He would like to test the ports with a structure that looks somewhat as a pumping port, but which has a known solution, which is a very good idea. To be followed up.

- TP informed us that Nicola Mariani (EN/MME) and Christine Vollinger (BE/RF), working with Fritz Caspers, are using a method to characterize the ferrite (CZ will follow this up) and that the same method will be used for the Phase II collimator studies (with SiC). We said that we need a frequency range from ~ 8 kHz to ~ 2 GHz a least. To be on the safe side, we could therefore ask for [1 kHz, 3 GHz]. As suggested, we could try and invite her at some point to present the method and ferrite's results at one of the next ICE meetings. Action 1: EM.

 - Emanuele Laface informed us that from the Logging Database (under Timber), a better resolution is now obtained: 10 s instead of 1 min in the past (following a request from Gianluigi Arduini). Reminder: With the Measurement Database, we have still points every 400 ms (i.e. at 2.5 Hz), but it is available only during 7 days.

- SL meeting last week:

- CAS: Daniel Brandt confirmed that all our candidates have a place at the next CAS (Bulgaria, 19/09-01/10).

- Last (65th) LMC: It has been decided to increase the number of bunches without reducing beta* in order to reach the target luminosity of 1E32 cm^-2 s^-1 by the end of the year.

- News on the LHC:

- Record peak luminosity ~ 4E30 cm^-2s^-1. Still with 25b * 25b, but with smaller transverse emittances => Seem ~ 2 microm, both planes, both beams (the controlled transverse emittance BU was not done in the SPS but the scraping was kept). These small transverse emittances are better for the luminosity but also for the injection process, where losses are observed (as could be anticipated in case of aperture issues).

- Accumulated luminosity of more than 1 pb^-1, which is a milestone but only 1/1000 of the goal for the end of 2011 (1 fb^-1).

- Presentation by WH at the last (65th) LMC dedicated to "How to get to a peak luminosity of 1E32 cm^-2s-1" (https://espace.cern.ch/lhc-machine-committee/Presentations/1/lmc_65/lmc_65h.pdf).

- Current plan: Move to 48b * 48b at the end of August and then move to bunch trains (75 ns or 150 ns, still under discussion).

- Requests from Fritz Caspers:

1) Plot of all (longitudinal and transverse) impedances for a graphite tube (L = 1 m, b = 2 mm) without and with copper (or Titanium) coating (5, 10 and 20 microm) at 7 TeV/c => 2 displays: Log (as usual) + Linear over 1 or 2 decades (=> several plots to cover all the interesting frequency range). We could also perhaps redo the same thing in the real case of flat chambers.

2) Computation of the associated tune shifts (single-bunch and coupled-bunch) at 7 TeV/c.

3) Same as 1) but for beta ≠ 1.

4) Luminosity scans in the LHC (Simon White): ppt

- Simon was invited to give this talk as there were some discussions in the past few weeks concerning the automatic luminosity scan he developed. Fast (of the order of few s) beam losses were often observed in stable-beam conditions, after some automatic scans. At some point, OP stopped to use Simon's application and performed manually "gentle's scans". In the same time, the transverse damper was put in operation and the fast losses disappear. The idea of this talk was to try and clarify the role of the automatic scan in the observed beam losses.

- Simon already gave a talk at yesterday's LHCCWG (https://lhc-commissioning.web.cern.ch/lhc-commissioning/meetings/20100810/LHC-BC-WG-Min10August10.pdf).

- Simon first reviewed the Van Der Meer method. Measuring the collision rates (given by any luminosity monitor, BRAN or experiments) as a function of the separation (moving the beams stepwise across each other), in both planes, provides a direct measurement of the effective overlap area. Critical parameters: intensity, beam displacement.

 - The beam displacement is done with a closed-orbit (4-magnet) bump, using the MCBC and MCBY => Subject to non-closure due to optics errors, hysteresis etc. Furthermore, displacing the beam at the IP creates a large offset in the TCT region. The limit was set in the past to 0.7 sigma at the TCT (not to compromise the collimators hierarchy and the triplet protection), i.e. 3 sigmas at the IP. Using the MCBX (situated in the common region of the 2 beams), the amplitude of the bump at the TCT can be reduced, but it is subject to hysteresis.

 - A software has been developed for IR steering, luminosity optimization and calibration => Dedicated routines for each purpose. All the trims are sent via LSA. Simon's application is called "Luminosity Scan Application". The main sources of systematics were indentified and qualified.

- The Optimize panel (which allows for fast automatic optimization) was tested last week at IP5 by Simon and it behaved as expected. The method is now operational. The optimization process takes few minutes per IP.

- Next steps:

- Need more experience: run like this for a couple of weeks.

- Extend to parallel optimization. Full optimization few minutes.

- Set limits with respect to the reference orbit.

- IP2:

- Colliding with an offset. Reminder: the beam separation in ALICE is always done horizontally (to ~ 4 sigma). Cannot apply automatic steering => Set reference with head-on collisions: Automatic optimization, control separation in both planes, easier to find collisions.

- Since few days, Simon's application is again used by operation and it seems to work perfectly even with high-intensity beams.

5) Observations of fast and slow beam losses in stable-beam operation in the LHC (Emanuele Laface): key and ppt

- Emanuele presented several cases of beam losses, comparing (during 1h30min) the evolution of the bunch intensities, the evolution (turn after turn) of the beam centroid in x and y at a Beam Position Monitor and the FFT, i.e. the beam spectra in x and y:

- 19/07/10 => Stable run.

- 09/07/10 => Fast losses, the most frequent case. The fast losses are associated with coherent motion, visible both on the BPM's raw data and on the spectra. It is more critical in the vertical plane but appears also in the horizontal plane, both with a tune line slightly above 0.315. See also http://emetral.web.cern.ch/emetral/ICEsection/1stMeeting_14-07-10/StatusOfTheLHCInstabilities.pdf, pages 15-16-17. The losses seem to be correlated to some lumi scan in IP2.

- 26/06/10 and 01/07/10 => Other fast losses.

- 08/08/10 => Fast losses but due to a beam dump. Frank Zimmermann reminded us that the dump system dumps the beam in 1 turn, but it seems that we observe BPM's data for ~ 100 turns. Gainluigi Arduini confirmed after the meeting that this BPM behaviour was already observed in the past in the case of controlled beam dumps. Note that there are several cases of unexplained dump triggers (from BLM's signals) and that they could be due to dust: Frank Zimmermann is following that.

- 26/06/10 => Slow losses with some coherent activity. Several lines are observed with time and it would be good to know what the beam-beam predictions. Action 3: TP and Emanuele Laface will put in operation an application showing for each bunch pattern (and beam parameters) the beam-beam tune footprint and the expected beam-beam lines.

- 03/08/10 => Slow losses without coherent activity.

- Next steps: Action 3 + Continue and collect more observations.

- There were also some discussions about the Schottky monitor which could be used (and was already used by some people in fact). Frank Zimmermann suggested to look at the Schottky monitor while doing a lumi scan in IP2.

6) EM simulations for the new PS wire scanner (BS): ppt

- Reminder:

- In the PS => 2 H + 2 V.

- Want to add a 5th one (H) in SS68 in January 2011. The goal is fabricate 2 tanks (1 for the new wire + 1 for calibration) for beginning of this fall. For the 4 others the tanks will be done during the shutdown in 2012.

- The new tanks are universal => No pb anymore of H and V, and there will be only 1 calibration (universal) => We will have 6 tanks the same. At the moment we have 2 tanks H and 2 tanks V, and the calibration depends on the H and V.

- Furthermore, the calibration could be done under vacuum which is not the case at the moment.

- William Andreazza (BE/BI) already discussed with Fritz in the past, and he wanted to use some existing ferrites. Fritz suggested to use the TT2-111R (TRANSTECH => Maryland. The name changed to SKYWORKS => To be checked) ferrite. Reminder: One should not use the 4A4 and 8C11 ferrite which is used for kickers. The maximum of the ferrite we want to use has a maximum mu'' in the GHz, whereas for the kickers the maximum is few MHz.

- Which thickness of ferrite do we need (if necessary)? => Do we put 0.5 or 1 cm (certainly between 0.5 and 0.8 cm)? Where exactly? => To be simulated. And this is what BS did.

- In the past, some SPS wire scanners burned due to the beam induced losses (http://cdsweb.cern.ch/record/624417/files/ab-2003-067.pdf).

- BS used the CST Particle Studio Wakefield Solver in time domain and CST Microwave Studio Eigenmode Solver in frequency domain, and he reminded us the main assumptions made in these simulations.

 - BS reminded us how to import a model from CATIA and what the current issues are.

- Simulations of the wire scanner tank alone =>  The power loss associated to the modes is very small (i.e. smaller than microW!). The very small power loss is mainly explained by the fact that the first trapped modes appear above ~ 800 MHz and the bunch length in the PS is quite long (compared to the SPS or the LHC) and therefore the bunch spectrum extends only to ~ 500 MHz in the worst cases.

- Simulations of the wire scanner tank + the wire and associated mechanical system reveal another peak at ~ 300 MHz and the total power loss in this structure increases to ~ 800 W for this mode! Comparing to the current PS wire scanners, the power loss is ~ 5 times higher, as ~ 160 W is predicted.

- This 300 MHz mode can be damped using ferrite. Reminder from Fritz Caspers: Putting some dispersive materials in the structure helps reducing losses due to HOM (as the power loss is proportional to the shunt impedance Rs = (Rs/Q) * Q, Rs/Q being independent of the material but Q decreasing with dispersive materials)!

- Reminder: To be the most efficient, the ferrite should be placed at the position of the maximum magnetic field. However, William told us that the available space is only at the top and bottom (see slide 21). Furthermore, they would like to use their 4S60 ferrite plates => http://www.ferroxcube.com/appl/info/absorbertiles.pdf and http://www.ferroxcube.com/prod/assets/4s60.pdf.

- Putting William's available ferrite, the power loss is reduced by 2 orders of magnitude. Furthremore, the effect on the longitudinal and transverse impedances at low frequency seem to be small.

- In the parking position and without ferrite, the power loss is 1 order of magnitude lower than in the IN position.

=> Recommendation sent to William Andreazza (BE/BI) after the meeting.

- Remark from BS after the meeting: He checked the ferrite dimensions and he realized that they should be 100 mm * 100 mm *  6 mm (with a hole in the middle with a diameter of 10 mm). In fact BS made his simulations with 100 mm * 100 mm *  10 mm (with a hole in the middle with a diameter of 6 mm). BS will re-launch his simulations but no major effects are expected (to be confirmed).

7) Actions to be taken for the next meeting

- Old actions.

- New actions:

- Action 1 (EM): Invite Christine Vollinger (BE/RF) to present her material characterization method and the ferrite's results at one of the next ICE meetings.

- Action 2 (NM): Answer to Fritz Caspers' requests:

1) Plot of all (longitudinal and transverse) impedances for a graphite tube (L = 1 m, b = 2 mm) without and with copper (or Titanium) coating (5, 10 and 20 microm) at 7 TeV/c => 2 displays: Log (as usual) + Linear over 1 or 2 decades (=> several plots to cover all the interesting frequency range). We could also perhaps redo the same thing in the real case of flat chambers.

2) Computation of the associated tune shifts (single-bunch and coupled-bunch) at 7 TeV/c.

3) Same as 1) but for beta ≠ 1.

- Action 3 (TP and Emanuele Laface): Put in operation an application showing for each bunch pattern (and beam parameters) the beam-beam tune footprint and the expected beam-beam lines on the tune spectra.

8)  Miscellaneous

- The next (6th) meeting will take place on 01/09/2010 => Agenda:

1) First measurements of longitudinal impedance and single-bunch effects in the LHC (Elena Chapochnikova)

2) A HEADTAIL analysis tool (KL)

3) Highlights from GSI workshop (BS)

- See preliminary agendas for the next meetings.

Minutes by E. Metral, 12/08/2010.