Minutes May 30th, 2008
Present: Gianluigi Arduini, Rama Calaga, Fritz Caspers, Elias Métral, Federico Roncarolo, Giovanni Rumolo, Benoit Salvant, Bruno Spataro.
Agenda
- Rama Calaga
"Impedance localization: HeadTail studies" (slides, minutes)
- Benoit Salvant
"Status of SPS impedance studies" (slides, minutes)
- Elias Métral and Bruno Spataro
Minutes
Actions from last meeting on April 17th :
Action 1: Elias will send the answer (when it comes) to everybody to see if one can run all the nominal operational beams (nominal LHC but also FT, CNGS etc.). Elena also mentioned that there will be a presentation at the SPS upgrade meeting on 13/05 to see the limitations from the SPS kickers heating etc. for upgrade scenarios.
Dealt with in SPSU meetings => http://paf-spsu.web.cern.ch/paf-spsu/meetings/2008/M-13-05/m15.pdf.
Action 2: Benoit and Elias will try and solve this possible discrepancy as well as the issue of Parabolic/Gaussian distribution to make a consistent data analysis:
Inconsistencies on assessment of bunch length during MDs delayed the study (see Benoit’s slides).
Action 3: Rama will perform HEADTAIL simulations to try and understand what happens starting with the simplified case of a single localized impedance, looking at the horizontal and vertical planes for a round and flat structure etc. On the basis of these results the type of constraint to be imposed will be defined and the analysis of the data should be done. Rama will ask Frank Zimmermann whether he has the 2003 data and/or whether he can redo the analysis with the constraints that will be defined so to have a comparison of the measurements in different years.
See Rama’s slides.
Action 4: Benoit will study the longitudinal plane in the same way as the transverse one taking the “real impedance” instead of a Broad-Band impedance model, starting with the wake-field from all the kickers. The situations in 2007 should be taken – i.e. with MKQH shielded and the “shielded” MKE kickers (up to now these shieldings were not taken into account).
Not started yet.
Action 5: Gianluigi and Elena will try and find someone to perform this study.
Presentation by Rama Calaga:
- The impedance localization program was applied to various headtail simulations outputs in which impedance sources were located in different places in the ring.
- This study was possible thanks to a new version of the HeadTail code developed by Diego and Giovanni, which takes into account the SPS lattice through a MADX input file. Rama checked that HeadTail and MADX simulated particles (very low intensity) followed similar trajectories.
- Case of a single impedance source (round chamber) located within the first 100 m of the machine: the main peak is clearly around the source. It is also important to notice that using all quads leads to less parasitic peaks in other locations of the ring.
- Gianluigi also noted that the sign of the peak seems consistent with expectations. From the first slide, vertical and horizontal tune shifts due to the impedance are both negative. That means that in both planes the impedance leads to less focusing. Therefore, the peaks should be negative in the horizontal plane, and positive in the vertical plane (as in the MAD convention focusing in V means negative K), which is what is obtained.
- Case of multiple impedance sources (flat chamber – actually the kickers) located in various places in the ring: the main peaks are again located around the sources. Including all quads again leads to a cleaner result.
- This time, the sign of the main peak in the vertical plane is not in agreement with what we would expect. However, it is now important to analyze this result in more details and check which quadrupoles are assigned a peak, and where the impedance source exactly is compared to the quads.
- Elias mentioned that the constraint applied in slide 9 should be K > 0 instead of K < 0. Everyone agreed.
- These results are very promising as the impedance localization technique seems to give a good idea of where the impedance sources are, even if parasitic peaks may lead to wrong conclusions.
- Conclusions and outlook: o The same analysis as the one applied previously to SPS measurement data was applied to HEADTAIL simulation data. It was checked that it is indeed possible to localize the impedance source(s), which is already a very good result. o However, the sign and amplitude of the peaks heavily depend on the location of the impedance. This analysis should only be used to locate an impedance source but no conclusion should be drawn from the sign and amplitude of the impedance contribution. o The next step will be to re-visit now the data analysis without constraint and this heavily depends on the cut etc. => It has to be done with care.
Presentation by Benoit Salvant:
- The SPS bunch length data from both Headtail monitor (R. Steinhagen, available in the CCC) and Wall Current monitor (T. Bohl, available in the Faraday cage) are very consistent (4 sigma bunch length ~ 0.8 ns). However, the reconstructed signals are significantly different (4 sigma bunch length ~ 0.7 ns for the WCM and 0.4 ns for the HTM). Action: Benoit will continue to study this with Ralph and Thomas.
- Total impedance from elements, which have a different geometry (flat chamber, round chamber or tilted for example) should not be exported through a single wake field (vertical dipolar wake field for instance), as the other 3 wake fields can not be directly inferred from that single wake field. Indeed, in this case, the dipolar and quadrupolar wake fields in both horizontal and vertical planes should be exported (i.e. 4 wake fields). In order to get the most up-to-date SPS impedance model, both Zbase and Headtail had to be modified.
- Also, obtaining the wake field data for the 20 kickers in Zbase represented a tedious task (30 minutes for each kicker with a high risk of making a mistake in assigning the data). Therefore, a quick import routine was implemented in Zbase to facilitate this process (now 30 minutes for all kickers, with no chance for the user to make mistakes – the mathematica routine still can make mistakes though!). Action: Benoit will continue to improve Mathematica/ZBASE/Headtail/MAD...
Presentation by Elias Métral and Bruno Spataro:
- Elias presented Bruno’s MAFIA simulations (time domain) of SPS BPMs and RF 800MHz cavities.
- Fritz mentioned that the impedance can not be accurately predicted from time domain simulations in which the oscillations are not damped within the time simulated (only the resonance frequencies can be predicted). Everyone finally agreed, and as it is not computationally possible to extend too much the simulated time, then frequency domain simulations should also be performed to assess the Q and shunt impedance of the resonances.
Action 1 : Bruno will perform frequency domain simulations of the BPMs (it was already done, and I will give you Benoit the corresponding resonator impedances to update your HEADTAIL simulations) and 800MHz cavities to find the resonator parameters.
- Fritz also mentioned that the electronics connected to the pickup play a major role in the impedance of the BPMs, and that they should be taken into account. Simulations are not easy; therefore everyone agreed that measurements should be performed. This will be done as soon as someone can help Fritz (may be the future PHD student from Daresbury).
Author: Benoit Salvant CERN
AB/ABP-LIS