Booster Beam Dynamics Christian Carli On behalf of the team working on and contributing to the study: M. Aiba did all ORBIT simulations (Booster modelling

Booster Beam DynamicsChristian Carli

On behalf of the team working on and contributing to the study: M. Aiba did all ORBIT simulations (Booster modelling …. injection) R. Garoby: elaboration of longitudinal painting scheme, Booster with Linac4 in the CERN complex B. Goddard and team: injection hardware, help implementing it in ORBIT M. Martini carried out ACCSIM simulations M. Chanel: experience on PS Booster operation and 160 MeV measurements S. Cousinau and her colleagues at SNS for help and advice on ORBIT F. Ostiguy for advice and help with ESME F. Gerigk, D. Montillot …..

Linac4 Machine Advisory Committee - Booster beam dynamics C. Carli 2/1529th January 2008

Overview of the presentation

Introduction PS Booster Overview PS Booster with Linac4

Injection studies Active longitudinal Painting Simulations of longitudinal Painting ORBIT Simulations of the PS Booster Injection

Benchmark effort ACCSIM versus Measurements

Integration in the CERN complex Further potential Topics and Studies Summary


Introduction - PS Booster Overview

Constructed in the 70ies to improve the performance (intensity) of the PS Four superimposed rings (chosen instead of fast cycling synchrotron or 200 MeV Linac) 16 cells with triplet focusing, phase advance >90o per period Large acceptances: 180 m and 120 m in the horizontal and vertical plane,

respectively Multiturn injection with betatron stacking and septum at 50 MeV Acceleration to 1400 MeV (was 800 MeV at the beginning) Double harmonic RF (now: h=1 and h=2), resonance compensation, dynamic working

point … Intensities of up to 1013 protons per ring (i.e. four times design) obtained Maximum direct space charge tune shifts larger than 0.5 in the vertical plane Vesatile machine - many different beams with different caracteristics

Sketch of the PS Booster with: - Distribution of Linac beam into 4 rings - Recombination prior to transfer


Introduction - PS Booster with Linac4

PSB to PS transfer energy increased in two steps from 800 MeV to 1400 MeV: To alleviate direct space charge effects at PS injection Creation of a bottleneck due to direct space charge in the Booster at low

energy PS Booster with Linac4:

Increase of PS Booster injection energy from 50 MeV to 160 MeV and 2 by factor 2

(Beam stays (a bit) longer with large direct space charge detuning) Goal: Increase of intensity within given emittances by factor 2

> Nominal LHC beam with PS single batch filling, Save generation of ultimate LHC beam with PS double (and single ?) batch filling, Increase of number of protons available for other experiments

H- charge exchange injection and Linac4 beam chopping: Opens possibility for painting (in all three planes ?) and further gain in performance

Losses and activation ?? Losses (on septum) inherent to conventional multiturn injection disappear Losses (during times with large direct space charge detuning) at higher energy

Some bottleneck at transfers PS Booster -> PS and PS -> SPS ?!


Introduction - PS Booster with Linac4Beams to delivered: Beams considered challenging:

In addition various beams (less challenging ?): LHC single bunch beams: Pilot beam, Probe beam, single bunch physics

beam Beams for PS physics: Slow ejection east hall, AD, nTof … MD beams … Many of these beam may be even easier with Linac4 (Painting of small

longitudinal emittances, no sieve reducing the Linac2 intensity needed … )

Beam PSB intensity per ring

(1012 protons/ring)

PSB Transv. Emittances

(rms, normalised)

PS intensity after injection (1012 protons)

LHC nominal – single batch PS filling (3 out of 4 PSB rings used)

> 2.76 and < 3.25 (2 bunches per ring)

2.5 m (H) 2.5 m (V)

> 8.3 and < 9.8

(6 bunches) LHC ultimate - double batch PS filling (6 bunches from two shots)

>2.04 (1 bunches per ring)

2.5 m (H) 2.5 m (V)

>12.24

(6 bunches) LHC ultimate – single batch PS filling (3 of 4 rings, feasible with losses ?)

> 4.08 (2 bunches per ring)

2.5 m (H) 2.5 m (V)

>12.24

(6 bunches) CNGS - single batch PS filling >8 and <12.5

(2 bunches per ring) 11.5 m (H)

4.6 m (V) >32 and <50

(8 bunches) ISOLDE up to 16

(1 bunch per ring) 12 m (H) 7 m (V)

-

Remark: Intensities needed depend on losses in downstream accelerators (at present ~15% loss PS and SPS for nominal LHC beam).


Injection - active longitudinal Painting With Linac4: similar RF system than at present

Double harmonic fundamental h=1 and h=2

systems to flatten bunches reduces maximum tune shifts

Injection with d(B)/dt = 10 Tm/s (no need for injection with small ramp rate with painting any more)

Little (but not negligible) motion in longitudinal phase space.

No way for painting from synchrotron motion (large harmonic numbers and RF voltages ruled out)

Need for active painting (aim: fill bucket homogeneously) and energy modulation


Injection - active longitudinal Painting

Principle: Triangular energy modulation (slow, ~20 turns for LHC) Beam on/off if mean energy inside a contur ~80% of acceptance Nominal LHC: intensity with 41mA (!!!) after 20 turns High intensity: several and/or longer modulation periods

Potential limitations: Linac4 jitter, debunching of Linac4 structure in Booster


Injection - Simulations of

longitudinal motion

Simulation with ESME for nominal LHC beam: Energy modulation 1.1 MeV, plus rms spread 120 keV About 98 % of particles at the and buching factor ~0.60.

High intensity (reduced bucket hight due to direct space charge): Similar results with reduced E modulation

After 10 turns after 20 turns (completed injection) after~3000 turns


Injection - first Results with ORBIT

Combines: Transverse painting as designed by team working on new injection hardware Active longitudinal painting

In Case of Dispersion Missmatch: Blow-up of horizontal emittance Large/small emittance dependant on position in longitudinal phase space Investigations with/without dispersion mismatch

Evolution of rms emittances after injection - blow-up (extrapolation ?) may be too large(slightly larger relative blow-up for 99% and 100% emittances)

from M. Aiba



from M. Aiba

-20 x(mm) 20

-15 y(mm) 15

-180o phase 180o-2 E

(M

eV

)

2

-6 y’

(m

rad

)

6-3

x’

(m

rad

)

3

-180o phase 180o

0 y(m) 100

0 x(m) 100-20 x(mm) 20

-15 y(mm) 15

-180o phase 180o-2 E

(M

eV

)

2

-6 y’

(m

rad

)

6-3

x’

(m

rad

)

3

-180o phase 180o

0 y(m) 100

0 x(m) 100

Dispersion at injection matched D mismatch (line ends with D=0m, Dbooster~-1.4m)



from M. Aiba Only nominal LHC beam

considered so far No simulations for high

intensity yet 240 000 macro-particles over

5000 turns (~5ms) Feasible within a reasonable

time for the job to run No acceleration, nor injection

foil scattering, nor machine imperfections

No dramatic effect of dispersion mismatch

Blow-up difficult to extrapolate (straight line, saturation …) Simulation over more turns and

with more particles ?


Benchmark Efforts -ACCSIM versus Measurements

Benchmark measur’ts (M. Chanel): High intensity (1013 protons in one

ring) beam at 160 MeV plateau Time evolution of emittances and

intensity Long bunches (see fig.) tune shifts

~0.25 Short bunches (second harmonic RF in

phase) -> more losses Different working points ….

(2)=(100, 47) m

(2)= (93, 52) m

Simulation (M. Martini) with ACCSIM: Only short times (computation time) In general overestimation of growth

rates (except long bunches & and hor. plane)

Insufficient statistics ? Attempt to compare with ORBIT ?

As well significant ACCSIMORBIT benchmark effort: Moderate agreement only so far

6

8

1 0

1 2

1 4

1 6

1 8

2 0

0 2 0 4 0 6 0 8 0 1 0 0 1 2 0 1 4 0 1 6 0 1 8 0 2 0 0 200 Time (ms)

1000

0

p

roto

ns

10

13

0 time(ms) 2000

rm

s*

2

0

Dashed: measurement (interpolatio)Solid: simulation

v* long bunch

v* short bunch

H* short bunch

H* long bunch - fair agreement


Integration into the CERN Complex - present Ideas, under InvestigationProposals made (R. Garoby et al.) for demanding Beams:

Nominal LHC Beam with single batch PS transfer: Splitting in the Booster, transfer of beam from three rings into hPS=7 buckets (rest

similar to present scheme with Linac2, but no double batch), details/options under discussion:

hPSB=1 (and hPSB=3) component to adjust spacing to hPS= 7 Shorter bunches due to ejection kicker, bunch rotation in PS, adiabatic RF voltage decrease … No hPSB=1 component and injection off the bucket center for blow-up

Ultimate LHC Beam (several options): Increase intensity (if feasible) of scheme for nominal LHC beam With double batch PS filling:

Like with Linac2 for nominal LHC beam, but higher intensities Tunes shifts and spreads in PS acceptable ? (some margin for bunch length)

“Exotic” schemes with single batch PS filling using all four PSB rings: Four Booster rings with hPSB=3 fill 12 out hPS=14 buckets … Four Booster rings with hPSB=2 fill 8 out hPS=10 buckets …

High Intensity (CNGS like): Schemes similar to present case hPSB=2 and hPS=8 with Linac2 (possibly higher

intensities)

Other beams with similar characteristics to present situation with Linac2: No problems for transfer to PS (just the injection into the Booster changes)


Further Studies Injection studies with validation & optimization of the painting scheme:

Add acceleration and injection foil, possibly machine imperfections Tracking over longer times, check parameters to avoid numerics problems Check filamentation of structure from injection - especially with dispersion mismatch Limitations: Linac4 energy jitter, debunching in Booster and associated energy spread Elaborate scenarios for the generation of all beams needed

Integration into the CERN Complex - Elaborate detailed scenarios for all beams needed

Check limitations of present Booster hardware: Instabilities (existing damper with higher intensities) (Cure beam loading problems of h=2 cavities for h=1 beams to remove a potential

limitations for ISOLDE beams) Injection losses and associated activation (“normal” losses, failure scenarios …):

Desirable, in collaboration with or by injection hardware team (resources ?) Possibly advanced Studies on slow Beam Losses, and Activation and Collimation

….: Feasibility (computational tools, resources) unclear Most (all) accelerators with large direct space charge designed without detailed

simulations Losses in ring during period with large direct space charge effects:

Prerequisite: Benchmark measurements/simulations (160 MeV) to judge feasibility Estimate losses and induced radiation Possibly devise simple collimation (scraping needed before PS transfer)


Summary Linac4 to improve Booster performance

Double 2 and, thus, intensities (within fixed emittances) by a factor two Versatility of the Booster must be maintained Active longitudinal Painting proposed: gain: ~10 % intensity for fixed

emittances Studies carried out so far

ESME simulations for longitudinal painting First ORBIT Runs of the complete Injection Benchmarks ORBITACCSIM and ACCSIMMeasurements

agreement not (yet?) satisfying Outlook

Complete injection simulations (acceleration, foils …) Validate longitudinal painting (energy jitter, Linac4 debunching) Complete study on integration of Linac4 & Booster into the CERN complex Injection Losses highly desirable Possibly check other potential problems (cavities, instabilities & damper …) Possibly advanced performance estimates with simulations:

Agreement simulations/measurements a prerequisite Study of Losses & Blow-up during Acceleration, Activation, Rudimentary

Collimation ?

Documents

Booster Beam Dynamics Christian Carli On behalf of the team working on and contributing to the study: M. Aiba did all ORBIT simulations (Booster modelling