Longitudinal beam loss mechanisms for LIU and LAGUNA beams

Longitudinal beam loss mechanisms for LIU and LAGUNA beams

E. Shaposhnikovawith input from T. Argyropoulos, T. Bohl, J. E. Muller, C. Lazaridis,

H. Timko

LIU-SPS collimation review CERN 21.11.2013

SPS Beam Performance

*Feasibility including operational viability (especially in the PS) remains to be demonstrated

11/21/2013 LIU-SPS collimation review 2

Operation SPS record After LIU (2020)

Beam type: LHC CNGS LHC CNGS LHC post-CNGS

SPS beam energy [GeV] 450 400 450 400 450 400bunch spacing [ns] 50 5 25 5 25 5bunch intensity/1011 1.6 0.105 1.3 0.13 2.2 0.17number of bunches 144 4200 288 4200 288 4200SPS beam intensity/1013 2.3 4.4 3.75 5.3 6.35 7.0*PS beam intensity/1013 0.6 2.3 1.0 3.0 1.75 4.0*PS momentum [GeV/c] 26 14 26 14 26 14PS cycle length [s] 3.6 1.2 3.6 1.2 3.6 1.2*SPS cycle length [s] 21.6 6.0 21.6 6.0 21.6 6.0SPS average current [μA] 0.17 1.17 0.28 1.4 0.47 1.9

SPS power [kW] 77 470 125 565 211 747

LIU-SPS collimation review 3

LHC beam

11/21/2013


Distribution of losses in LHC cycle

• Very large losses in the past, reduced with time (machine tuning, e-cloud scrubbing?)

• Can be a serious limitation in future for very high intensity required by LIU for HL-LHC

• Relative losses increase with intensity N, absolute ~ N2?• Losses occur

– at capture (bunch shape)– on flat bottom (full bucket due to injected bunch shape)– during ramp above 120 GeV/c due to multi-bunch instabilities

or/and controlled longitudinal emittance blow-up (1-2%)

11/21/2013


Capture loss due to PS bunch shape(after rotation in longitudinal phase space)

Shape can be improved by higher PS voltage: => tails are less populated but losses are there!

11/21/2013

H. Timko et al.,ESME simulationsof realistic bunchdistribution from PS tomography(no intensity effects included)

bestresults:2.5% loss

operation:5% loss


Capture losses: uncaptured beam after injection(200 MHz signal )

Injection at 26 GeV/c A few seconds later

11/21/2013

Uncaptured beam is always moving to the left.Energy loss (dp/p < 0) due to resistive impedance?

Uncaptured beam

Transmission of 25 ns beamin MDs with Q20 optics (2012)


J. Esteban Muller et al.

=> Large losses and also increase with intensity


Losses Transmission (from BCT and Larger) ~ 85-89 % for single batch 90-92 % for 3 or 4 batches continuous losses along flat bottom

larger inj. emittance with similar intensitiesÞ scrapping

on FB

T. Argyropoulos et al.

Single batch

11/21/2013


Voltage programs for Q20 (MDs) 200 MHz voltage program settings:

I. 2.5 to 4.5 MV - 4 dips at injections II. 4.5 MV constantIII. 3.5 to 4.5 MV – 1st injection and 2.5 to 4.5 MV – for the restIV. As in III + 500 kV at acceleration and flat top (avoid losses for

higher intensities)V. As in IV but no first dip

optimal

T. Argyropoulos et al.

Þ Voltage increase during ramp required for higher intensities

Þ Voltage close to the limit

Þ For higher intensitiesÞ particle lossesorÞ increase length of the

cycleÞ longer LHC filling time

11/21/2013

Acceleration: voltage and power for nominal intensity in Q20 and Q26

Voltage Power


=> Power will be at the limit also during acceleration above nominal intensity!


High intensity LHC beam• High intensity: Np = 1.36x1011 p/b at FT

• TWC 200 MHz voltage program: case III

Controlled emittance blow-up is difficult to optimise for high intensity beam

11/21/2013


High intensity FT (LBNO) beam

11/21/2013


Main intensity limitations in the SPSfor CNGS-type beam (LBNO)

• Equipment heating (MKE, HOM couplers, beam instrumentation…)

• Beam losses• Transverse damper (40 MHz bandwidth)• RF voltage and power, beam control:

• Beam stability and bucket area for (un)controlled emittance blow-up

• Maximum available voltage in the 200 MHz RF system: – 7.5 (8) MV

• Maximum available RF power in one 200 MHz TW cavity:– 700 kW for full SPS ring (CNGS-type beam)

11/21/2013

LIU-SPS collimation review

LHC and CNGS-type beams in the SPS

• FT/CNGS beam from PS: – practically debunched beam– 5-turn extraction– no bunch-to-bucket transfer– injection below transition

14

Nominal parameters of two main types of proton beam in the SPS

11/21/2013

high intensity run


Particle losses during cycle

• Capture– Beam structure from the PS: de-bunched beam with 200

MHz density modulation => no bunch-to-bucket transfer

• Ramp– Uncontrolled emittance blow-up due to instabilities

during transition crossing and at higher energies (Nth ~ 1/E)

– Limited RF bucket area due to beam loading with RF power limited to 750 kW

11/21/2013


Loss distribution during high intensity run (2004)

• Injection losses - 5%• Losses on flat bottom ~ 2%• Particles in the kicker gap - losses

at extraction ( 2%) => cleaning by trans. damper

• Capture loss 3-4%• Beam losses at transition: 4%• Continuous losses after transition:

5% - 2% => early beam dump - main intensity limitation for the 2004 record of 5.3x1013

11/21/2013

Critical losses


Bunch length along the batch during cycle for high intensity beam (5.6x1013 injected, 15% losses)

(AB-Note-2005-034 RF, T. Bohl et al.)

t=1.315 s

t=3.286 s

t=1.534 sγ>γt

t=4.163 s

11/21/2013


FT/CNGS acceleration cycle:voltage and power

Þ at the moment maximum available voltage is used due to uncontrolled emittance blow–up during transition crossing - any voltage reduction leads to beam losses

11/21/2013

Christos Lazaridis (CERN)

MD with CNGS Beam in 2012

● Goals of 2012 measurements :● Study beam stability● Verify present intensity limitations

MD data analysis was done by C. Lazaridis

200MHz RF Voltage Program

Beam momentum

Obtained profiles

~3600 bunches

2 PS batches after 2nd injection in SPS


CNGS Cycle


Nominal CNGS Cycle

• Injected beam is practically debunched

• Bunches not well defined after injection

• transition

• instability at the end of the cycle

Bunch Length

Batch 1Average and min-max

for each frame

Beam structure after injection

Batch 2


a.u.


RF Voltage optimization at injection

Nominal voltage1.99 MV0.6 MV

Nominal voltage

0.6 MV

Batch 1 - 0.6 MV/1.99 VAverage Bunch Length

ms

MV

SPS BCT


Batch 2 - 0.6 MV/1.99 V

• RF voltage at injection was varied in range 0.6 - 2.0 MV leading to changes in emittance• Trying to reduce capture losses=> Nominal 0.9 MV close to optimal

Greater bunch spreadMore capture losses

s

ns

Transition crossing and after

Average Bunch Length

Relative Bunch Length

Batch 1/Batch 2

Reference : 1450 ms

Reference


1457 msγ transition(1480 ms)1490 ms 1501 ms

• Bunch oscillations start immediately after transition (1480.2 ms)

Bunch number

End of cycle

Average Bunch Length

Relative Bunch Length

Batch 1/Batch 2

Reference : 1546 ms

Reference


1932 ms 2760 ms 3588 ms4278 ms

• Beam is very unstable at the end of ramp

• Small losses observed

Bunch number


Reducing beam losses• Losses due to instabilities around

2.8 s in the cycle• Maximum 200 MHz RF voltage

• Tried reducing voltage to improve stability• Constant bucket area• Increasing synchrotron frequency

spread inside the bunch

Nominal voltageModified

ms

Bucket AreaNominal voltage

Modified

Beam Intensity


For intensities ~3.7x1013 losses are 3.5%=> 0.2% reduction in high-energy losses

Proposed changes applied to CNGS cycle


Possible future improvements for beam loss reduction

• 200 MHz power upgrade – limit will be still at 750 KW (due to full ring), but with 2 extra sections

• LLRF upgrade of 200 MHz RF– Separate capture of each PS batch in the SPS would allow

voltage capture modulation (e.g. 0.9 MV increased to 2.5 MV) – Variable gain of 1-turn-delay feedback – Upgrade of the frequency range of the feed-forward system

(below 26 GeV/c)• Use of the 800 MHz RF system during cycle• Impedance reduction• Q20 optics(?)

11/21/2013


The 200 MHz RF system

Presently both voltage and power are at the limit: 7.5 MV used after transition crossing (due to uncontrolled longitudinal emittance blow-up) Improvement after power upgrade with 6 cavities (18 sections)

4200 bunches spaced by 5 ns

N = 4.8x1013 (Irf = 0.73 A)-> V=7.5 MV

N = 7x1013 (Irf = 1.06 A) -> V = 9 MV

11/21/2013

Voltage available for acceleration with Pmax=0.7 MW


Voltage during FT/CNGS cycle for two optics

Q26 Q20

=> voltage above present limit of 7.5 MV even for 0.4 eVs

=> after transition crossing some bunches have emittance > 0.6 eVs

11/21/2013

limit


Can new optics help to reduce uncontrolled emittance blow-up?

Voltage program for 0.6 eVs Slip factor (~ beam stability)

11/21/2013


Summary• LHC beam

– Capture losses are determined by S-shape of injected bunches – Better stability for larger PS emittance, but more losses as well– Flat bottom losses are defined by full bucket– High energy losses come from beam instabilities and controlled emittance blow-

up in conjunction with limited RF voltage (power)• FT high intensity beam

– Absence of bunch-to-bucket transfer will be always a source of capture loss– Beam control during transition crossing is difficult - a source of losses– Only small increase in available voltage can be expected after the 200 MHz

power upgrade -> limited voltage (bucket) at high energies• Relative losses increase with intensity -> high absolute losses can be

expected during HL-LHC era (and LBNO) in the SPS at high energies unless beam instabilities are eliminated at source (impedance)

11/21/2013


LIU-SPS upgrades• Main difference between the two beams:

– injection at 14 GeV/c and transition crossing => different beam control (LLRF)– CNGS beam fills whole SPS ring and LHC beam – only half => different requirements for beam power (continuous and pulsed

regimes)– bunch spacing => multi-bunch effects (instabilities, heating)

• CNGS-type beam will profit from planned SPS upgrades: – impedance reduction (shielding of MKE kickers, …) – e-cloud mitigation– 200 MHz and 800 MHz RF upgrade– beam instrumentation – low γt (transition energy) optics?

11/21/2013

Documents

Longitudinal beam loss mechanisms for LIU and LAGUNA beams