Skip to content

Minutes of E-Beam Meeting #5 (draft)

The meeting was held at CERN (892/1-B09) on 10/07/2019 - See indico

Participants

A. Rossi, A. Levichev, D. Gamba, G. Tranquille, G. Arduini, L.V. Joergensen, L. Ponce, M. Zampetakis, N. Biancacci, R.A. Fernandez, R. Corsini, V. Kain, S. Hirlaender

Actions

  • D. Gamba: Circulate link to the new E-Beam website
  • D. Gamba, A. Rossi: Update list of future meetings
  • D. Gamba, L. Ponce: Organize follow-up AD optics studies following ideas of Pavel Belochitskii
  • N. Biancacci, D. Gamba: Prepare e-cooler lecture for LEIR operators
  • R.A. Fernandez: Coordinate the consolidation of the LEIR e-cooler software control system for operation, especially the set-up of high level parameters in collaboration with BI and the verification of unknown knobs (e.g. LEIRBEAM/DX_EC and LEIRBEAM/BeCool)
  • A. Rossi: Set up a framework for simulating the e- beam in the e-cooler
  • D. Gamba, A. Latina: Set up a framework for simulating the electron-ion interaction in LEIR
  • S. Hirlaender: Explore the possibility of using reinforced learning algorithms for e-cooler optimization (in collaboration with A. Latina, D. Gamba)
  • N. Biancacci, R. Scrivens: Study the possibility of optics coupling compensation
  • A. Frassier: Follow-up on the fast switch implementation for e- gun gird/control electrodes (tests in LEIR to be coordinated with OP)
  • G. Tranquille: Installation of coupling transformer for grid electrode modulation (from ELENA)

Future topic (A. Rossi)

Adriana presented the the list of future topics which is now available on the newly created E-Beam website.

Follow-up plans from the AD e-cooler review are still under discussion by the reviewers. If the option of a new e-cooler is chosen, there will be a technical design review next year, for which more detailed studies are needed.

The electron lenses project is still at its design stage, and mainly dedicated meetings are organized among a restricted group of participant - see for example e-lenses category on indico. More general meetings have to be expected after summer.

Other topics in the pipeline are a review of e- beam diagnostics used in various e-cooler and e-lenses devices, and a review of test stand facilities available at various institutes.

The discussion resulted in the following decisions:

  • Postpone the meeting on operational experience of ELENA e-cooler to September if no other slot is available during summer.
  • Foresee a meeting on Status of ELENA toward the end of the year (e.g. November).
  • Foresee a meeting beginning/mid September on AD e-cooler review follow up, in particular effects of optics improvement ideas (suggested by Pavel Belochitskii at the AD e-cooler review) on the cooling performance (see indico)

Operational experience with e-cooler (N. Biancacci)

N. Biancacci recalled the use of the e-cooler in LEIR, in which up to 7 injections of ions are accumulated, one every 200 ms.

Operationally the optimal setup is obtained when: - Each injection beam has an energy spread of the order of \pm 0.2%, and is injected at a delta energy of +0.1% w.r.t. to the bunching RF frequency. - Each injected beam is cooled and dragged by the e-cooler to about -0.15% w.r.t. RF, where is stacked with the other injections. - After the last injection, the beam is dragged by the e-cooler to the nominal RF frequency, bunched and then accelerated.

The e-cooler is typically optimized by looking at the available Schottky signal at h=100. One of the most important parameter to optimize the final emittance and lifetime of the stacked beam is the orbit overlap between ions and electrons. To be noted that minimum emittance also results in instabilities for high intensity beams, therefore a small angle (order of 1 mrad) between e- and ion beam is normally set for the NOMINAL cycle. Such an angle increases the final energy spread of the beam but it allows the beam to be better placed in the longitudinal stability diagram. Currently, such an optimization is performed by scanning the ion orbit/angle to produce cooling maps, procedure that can take several hours. G. Tranquille asked if the presented cooling maps were measured with single injection beams. N. Biancacci confirmed this, adding that no off-energy dragging were used for those measurements.

N. Biancacci showed the measured impact on tune due to e-beam interaction with the circulating ion beam. This is corrected once a year with a static correction in the requested tune along the cycle.

It has been observed that a missing injection, or lower beam intensity coming from the LINAC3, has an impact on the equilibrium ion velocity, resulting in sudden ion losses. The reason for this effect is not understood, yet. The operators normally react to this by changing the e- velocity acting on e-cooler e- gun, grid and control voltages. G. Tranquille and A. Rossi commented that such a correction has an impact on the e- distribution due to space charge and e- gun geometrical layout.

The overall set-up and optimization of the e-cooler requires careful adjustments of several parameters, sometimes sparse in different workingsets and not necessarily linked to each other, making it prone to mistakes. Operation also remarked that is often difficult to understand if all e-cooler hardware is working properly, making troubleshooting time consuming and often requiring the (maybe un-necessary) intervention of an expert. A. Rossi suggested to set up a diagnostic tool, e.g. with Inspector, to monitor all equipment and easily show faults/non-conformities.

N. Biancacci stressed that from an operational point of view it would be useful to define and implement into LSA higher level parameters like:

  • e- relative velocity w.r.t. circulating ion beam
  • e- current

and to re-introduce the possibility of modulating the e- beam current in order to directly measure the overlap between ion and electron beams using the available BPMs, i.e. without the need of lengthy (several hours) cooler map scans. G. Tranquille commented that equations linking e-cooler gun voltages and e- velocity might not be trivial, e.g. the e- velocity strongly depend on the e- current due to space charge effects. A. Rossi proposed to set-up a simulation tool (e.g. WARP) as presently done for e-lenses, and to use it to measure normally unaccessible quantities on a test bench. This could be a long term study/development.

Some e-cooler measurable quantities seem to be logged only on Timber and not easily accessible by the operator. G. Tranquille commented that all signals going to Timber are on Samplers which therefore must be available in the control system. The e- current should also be available. N. Biancacci remarked that presently only one data point per cycle seems to be acquired - to be checked.

The possibility of having a predictive control of the machine optics was envisaged. Some LSA parameters (LEIRBEAM/DX_EC and LEIRBEAM/BeCool) have been found, but operation is not aware of their meaning. G. Tranquille is not aware of them, but he suggests to ask C. Carli or A. Huschauer.

On the longer time scale, there are ideas of using reinforced learning algorithms to help the set-up and optimisation of the e-cooler parameter. S. Hirlaender will start exploring those ideas in simulations.

V. Kain stressed that the main concern for operation is to find a failsafe way to understand if the e-cooler has a problem and how to get back to a working situation, possibly without the intervention of an experts. She also remarked a general degraded state of the e-cooler control system configuration, e.g. some elements of e-cooler magnetic system are specified in LSA as non-PPM, while they maybe should be, and this has been a source of issue in the past. G. Tranquille commented that BI is responsible for the hardware and for the general set up of the e-cooler, while OP should be responsible for the fine tuning and for the development of the necessary operational tools. BI remains at disposal for any questions and support. The discussion lead to several actions mainly act to consolidate the software to control of the e-cooler, which will be coordinated by R.A. Fernandez.

N. Biancacci concluded announcing a "LEIR e-cooler operation lecture" which is being organized by R.A. Fernandez and to be prepared by N. Biancacci and D. Gamba.

Status and plans for hardware interventions on LEIR e-cooler (G. Tranquille)

During LS2 the LEIR e-cooler will mainly undergo general maintenance.

The modulation of cathode and grid voltages have been always in place (designed to heat the e- beam on purpose by modulating its energy/current). The bandwidth of this system maybe not enough for measuring the e- beam with the BPMs. A coupling transformer for grid electrode modulation will be re-installed. This, historically proven, is the most promising way to measure the e- orbit with the BPMs.

New faster switching on grid and control electrodes will be installed. The main purpose of this intervention is to increase the safety and reliability of the hardware, but it will also allow to stop the e- beam faster at the end of the cooling process. It might also be possible to make some on/off modulation with this new system, which might give a signal on the BPMs. A. Frassier is doing some tests in the lab, and the system could be installed and tested in LEIR by 2020. A. Rossi asked which is the required accuracy for e-beam orbit measurement and stressed that one should verify the BPMs response for e- and ion beams. D. Gamba replied that one should be aligned better than 1 mm, and indeed the BPMs response should be checked somehow.

R.A. Fernandez asked to coordinate with OP any planned tests of the new switching system in LEIR.

LEIR BI Schottky System Status (O. Marqversen)

D. Gamba presented the slides (only the first one) on behalf of O. Marqversen. In LEIR, the Schottky system already set-up by Ole in 2018 will be upgraded with new hardware which will make the overall system faster. The operation team asked to know when the new FESA class will be available: Davide will ask Ole about his plans in this respect.