Skip to content

Minutes of E-Beam Meeting #6 (draft)

The meeting was held at CERN (864/2-B14) on 18/09/2019 - See indico

Participants

C. Carli, R. Corsini, R.A. Fernandez, A. Frassier, D. Gamba, S. Hirlaender, L.V. Joergensen, A. Pikin, L. Ponce, A. Rossi, O. Sedlacek, G. Tranquille

Actions

  • D. Gamba: prepare "official" optics (Twiss files) of AD for cooling efficiency studies.
  • D. Gamba: provide equations to implement LEIR e-cooler setup via higher level parameters (e.g. e^- beam velocities, if possible)
  • D. Gamba, A. Rossi: organize new E-BEAM meeting for beginning of November.
  • G. Tranquille: follow up on e-beam orbit measurement system in all coolers.

General Discussion and Approval of Minutes

  • Beta-cool: our main software is presently not maintained, S. Nagaitsev from FNAL told D. Gamba they do not use BETAACOOL, but some internally developed code.
  • AD Optics: Gerard asked Davide to make available an "official" AD optics (output of MADX) for cooling studies to be used by everybody.
  • Possible presentations for the next meetings:
    • "AD Improved Optics for Cooling" by P. Belochitskii
    • "Review of Measurements Tools for e^- Beams" by A. Pikin
    • "Simulations of AD e^- Beam with WARP" by Andrei?
    • "E-cooler Operation and Optimization with Reinforcement Learning" by S. Hirlaender
    • "List of e-cooler Test Beds and their Capabilities" by G. Tranquille
  • AD Cooler review: Gerard assumes we are moving forward with the design/construction of a new cooler. Adriana suggested to verify with Mike Lamont if further inputs are needed and by when.
  • HL-LHC e-lenses: Adriana pointed out that the e-lenses are still not in the baseline. This will be discussed (normally with a positive outcome) probably during the next Cost and Schedule Review to be held at CERN on 11-13 November 2019. In the meantime, discussions with Novosibirsk collaboration are ongoing to set up a plan for design and construction.
  • LEIR e-cooler Control: Reyes asked Davide to provide a formula to set up LEIR e^- beam gun voltages with respect to beam velocity/intensity. Gerard commented that there might be too many parameters to take into account, and a parametrization in velocity might not be possible. Christian added that a slope on the e^- beam energy along the cycle was implemented to compensate for ion-e^- space charge effects with the increase of the ion intensity.

AD/ELENA e-cooler requirements, and past and present performances (D. Gamba)

Davide presented an overview of the e-cooler design and present circuits layout (Courtesy of A. Frassier). During AD Design it was thought to have 6 s and 1 s long e-cooling plateaus. Such a short plateaus were computed assuming up to 3.5 A e^- beam current. Gerard commented that in reality we can only get up to 2.5 A with present gun, and current decreases over time (now probably around 2.1). From this consideration one should already expect almost a factor 2 longer plateaus with respect to the design.

In the AD cooler the cathode voltage is the reference voltage with respect to which all other voltages are referred to. A peculiarity of such an architecture is that if e^- beam is lost on the way, the gun PC can go in current limitation, effectively lowering the chatode voltage and so the emitted e^- beam current. To be noted that in the AD cooler there is no way to control the e^- current. Beam is started/stopped by fast HV switch on the grid electrode, and its intensity governed by I_e = P V_K^{\frac{3}{2}}, where V_K is the chatode voltage and P=0.58 \times 10^{-6} is the electron beam perveance.

The software tools used by OP for standard setup and adjustments of e-cooling performance in AD are:

  • LabView interface: possible to see e^- beam current and losses, but poor sampling rate (a point every second). It also allows to see e-cooler interlock status. The current delivered by the collector power converter is used to determine the e^- current. Alexandre pointed out that e^- beam losses can mainly be detected looking at vacuum activity.
  • WebScope: for fast acquisition of e^- beam current and losses. Not integrated in CERN control system, it allows to verify the good behavior of grid fast switches used to start/stop the e^- beam.
  • Schottky diagnostics: in different forms, it allows to have an immediate non-destructive information of cooling performance. Fast frequency oscillations at the beginning of cooling maybe an indication of bending magnet hysteresis, but could also due to space charge or e^- beam energy variation.
  • Chatode voltage adjustment: a simple but effective device to help the operator choose how to adjust e^- beam energy. Gerard commented that the present control allows to make steps of 1.2 V, which might not be enough for the 100 MeV/c plateau. Request improve control resolution to EPC was made, but never implemented. At this stage is more likely that this should be translated in a new specification in view of e-cooler PC consolidation, eventually linked with the new e-cooler for AD.
  • Ion/Electron beams orbit overlap control: a standalone application allows to regulate the ion beam transverse offset and angles with hardcoded knobs. In the past it was possible to observe both ion and electorn beam trajectory in the e-cooler, however this is presently not possible. Christian commented that a measurement of the electron beam position and ion beam position (with same instrumentation to minimize relative error) is highly desirable. Davide thought YASP will be used in the future, but Laurette is skeptical on its use for the time being due to historical peculiarities of the AD control system.
  • Emittance measurements: presently performed only via scraper measurements. Gerard pointed out that IPM (Ionization Profile Monitor) is installed in the AD and it has been recently improved by installing a double MCP to reduce the amount of gas to be injected: last year there was no time to test it, but it should work.

Davide went through some considerations of cooling efficiency by Tommy Eriksson. For both plateaus cooling has been proven to work, but its longer cooling time has never been understood. A key issue is/was the formation of long tails/halo, which affect transmission. A year-by-year analysis of emittance evolution is envisaged by Davide looking at the available scraper data.

Davide presented the evolution of cycle and plateaus duration over the last 20 years looking at Trim history data. From this data the main conclusion is that after a long shutdown the performance of all coolings seems to have degraded, but this might be just due to too little time reserved for machine optimization. Gerard commented that e-cooling time seems to be affected by the time for stochastic cooling: when stochastic cooling is longer, e-cooling can be shorter and tails/halo might be less pronounced.

From the presented plots, it also looks like the optics moved with the years, and best performance of the e-cooler (shorter plateau) correspond to a same particular optics: to be further investigated.

Davide also pointed out that the AD cycle contains several little ramp and plateaus that has been added over time to optimize for beam transmission. Those points make any optimization of the cycle difficult. One should probably consider to simplify the cycle description using the newly available features of the control system. Laurette commented that this will still take several year, as in AD there is still a mixture of old and new technologies, which is being slowly renovated.

For the ELENA e-cooler the situation is slightly different:

  • All (but filament) cooler power converters are referred to ground, so e^- beam losses are trickier to calculate (passes through ground).
  • Grid voltage can be changed, allowing to control the e^- beam current.
  • An Inspector application gives info on e^- beam current and losses. Poor sampling rate still due to PC used to measure current. A scope-like implementation as in AD would be welcome.
  • Schottky measurement only possible on standalone spectrum analyzer. Work to have a system integrated with CERN control system ongoing by BI and RF, independently.
  • Transverse emittance can only be measured via destructive scraper measurements as in the AD.

Regarding the cooling performance in ELENA, with the limited experience of last year, cooling times of the order of 1sec and transverse emittance reduction of the order of ~80% has been measured. This is compatible with expectations, and maybe thanks to careful design and characterization prior of e-cooler installation.

Davide provided also plots for expected geometric emittance damping and ~e-cooling times as a function of momentum. To be noted that the gain in reduced emittance blow-up one would get if cooling at higher energy is counteracted by longer cooling times (at constant e^- beam parameters).


Status and plans for hardware interventions on AD/ELENA e-cooler during LS2 (G. A. Tranquille)

  • For the AD cooler we expect to replace the collector (with the new design) and thermionic gun before the restart after LS2.
  • New collector for AD e-cooler is still in design stage in the hands of MME
    • idea of using standard pieces (need to arrive at 68 kV, with 100 kV feedthroughs - maybe difficult)
    • it is supposed to have it built by middle of 2020 for testing in test stand
  • For all cooler (AD, ELENA, LEIR) different ideas for modulating e^- beam energy and/or intensity so to see a signal on the BPMs is being explored.
    • First tests on ELENA e-cooler to start soon.
    • Procedure on how to make such a measurement for each machine to be redacted by BI.

Davide asked if the BPMs in the AD cooler are actually operational. Gerard confirmed that there is no reason to believe otherwise.