25 ns performance - longitudinal plane T. Argyropoulos, E. - - PowerPoint PPT Presentation

• T. Argyropoulos, LIU Day 2014

Other means to increase the SPS 25 ns performance - longitudinal plane

• T. Argyropoulos, E. Shaposhnikova, Jose E. Varela

LHC Performance Workshop (Chamonix 2014)

22-25 September 2014

Acknowledgements: H, Bartosik, T. Bohl, F. Caspers, H. Damerau, A. Lasheen, E. Montesinos, J. E. Muller , D. Quartullo, H. Timkó, C. Zannini

Outline

 Status of the 25 ns LHC beam in the SPS before LS1 (2012)  Main performance limitations in the SPS for HL-LHC parameters

• RF power
• Emittance blow-up due to longitudinal instabilities

 Possible additional means to increase the 25 ns beam intensity in the SPS after approved LIU upgrades

• At SPS – LHC transfer
• During acceleration ramp

 Summary

25 ns beam in SPS before LS1

Achieved in the last 2012 MD at 450 GeV/c: 4 batches with intensity of 1.35x1011 p/b (double RF and Q20 optics) and bunch length ~ 1.7 ns

• To avoid losses bunch length required for transfer to LHC: τ4σ ≤ 1.7 ns (BQM max

1.9 ns)

• This result is used as a reference point for scaling to the higher intensities
• High losses for injected

intensities above 1.4x1011

• The 200 MHz RF voltage during

the ramp was increased close to the limit of 7 MV (from beam loading)

• Longitudinal instabilities

Beam transmission

• J. Esteban Muller et al.

HL-LHC: limitations for SPS-LHC transfer

• HL-LHC request: 2.4x1011 p/b at SPS flat top with τ4σ ≤ 1.7 ns, but
• larger longitudinal emittance is needed for beam stability (𝜁𝑚 ∝ 𝑂𝑐

1/2)

• limited RF voltage due to beam loading and potential well distortion

(𝑊 ∝ 𝜁𝑚

2 and 𝑊 𝑗𝑜𝑒 𝑄𝑋𝐸 ∝ 𝜐−3)

 After upgrade we can reach

• 2.7 A (2.1x1011 p/b)

without performance degradation

• ~10 MV should be

available for 3 A (2.3x1011 p/b)

• But 12.5 MV are required

Solutions:

• Increase acceptable

longitudinal emittance 𝜻𝒎

• Reduce longitudinal blow-up

(impedance) After 200 MHz upgrade (2020): 2x4 + 4x3

RF voltage at transfer to LHC

• Ref. point

Note: single bunch scaling for LD & PWD from present experimental results (ref. point)

Vind for τ = const (LD & PWD)

SPS longitudinal impedance model: RF cavities (200 MHz + HOM, 800 MHz), BPMs, kickers, resistive wall, unshielded pumping ports, Y – chamber, beam scrapers Search for high frequency impedance  Measurements at flat bottom with long bunches (25 ns) and RF off

Peak at 1.4 GHz

• Vacuum flanges (different

types, ~500 in the ring)

Uncontrolled emittance blow-up: possible impedance source

Uncontrolled emittance blow-up: microwave instability?

• Simulations of a single bunch on the SPS flat top as a function of intensity using

the SPS impedance model (including the vacuum flanges)  compare with measurements Single bunch at the SPS flat top (meas. from AWAKE MD in 2012)

• Good agreement of these

measurements with particle simulations

• Signs of microwave (mw)

instability

• Main contribution from the

1.4 GHz resonant impedance from the vacuum flanges from simulations: Nth = 2x1011

Q20 – Double RF – V200 = 2 MV (low voltage before bunch rotation) MD to find the mw instability threshold

Simulations for 6 bunches (25 ns spacing) at SPS flat top Intensity threshold as a function of bunch length for 1 & 6 bunches

Qualitative agreement of simulations with measurements:

• Nth of 6 bunches is ~ twice

lower than of single bunch

• Only a few bunches are

coupled, no coupled bunch modes  indeed in measurements 25 ns and 50 ns spaced bunches are coupled, but batches spaced by 225 ns are decoupled

• Nth increases with emittance

Q20 – Double RF – V200 = 7 MV MD for coupled bunch instability threshold 1 batch & different number of bunches

Uncontrolled emittance blow-up: multi-bunch case

Large longitudinal emittance at SPS flat top

• Large longitudinal emittance at flat top (> 0.55 eVs or τ4σ > 1.8 ns) 

problem for losses in the LHC

• Three solutions are considered:

1) Bunch rotation on the SPS flat top 2) New SC 200 MHz RF system in the LHC 3) Reduce uncontrolled emittance blow-up by impedance identification and reduction

Bunch rotation at flat top (1/3)

 Already tried successfully for single high intensity (~2.5 - 3.0 x1011) bunches (MD for AWAKE)  but for very small emittance (εl~0.3 eVs)  Maximum needed voltage available only in 2020  Larger bunch tails  more beam loss in the LHC ? Bunch length (ns) N = 2.8x1011 Single bunch MD for AWAKE in 2012

Bunch rotation at flat top (2/3)

 Starting rotation from V200 = 5 MV and assuming V200 = 10 MV available at flat top (2.3x1011 p/b and Q20 optics)  Simulations with the SPS impedance + FF and FB in the 200 MHz RF  bunch position variation along the batch agrees very well with measurements Bunch rotation for LHC beam can be tested in the SPS with limited 200 MHz RF voltage

Measurements with half intensity N = 1.3x1011 p/b

Bunch position variation along the batch

Bunch rotation at flat top (3/3)

LHC capture with 6 MV: simulated bunch position variation in the LHC Buckets Particle Losses less than 1.5 % per bunch  follow the beam loading effect of the SPS 200 MHz RF system  pessimistic estimations

• Avg. bunch length, τmean = 1.45 ns

Bunch 1 Bunch 36 Bunch 72

• J. Esteban Muller

Possible MD on SPS – LHC transfer

New SC 200 MHz RF system in the LHC

(more in talks of R. Calaga and R. Tomas, Session 6 - HL-LHC)  Clean transfer between SPS and LHC  Double RF system in the LHC  better stability?, flat bunches, …  Additional impedance in the LHC, reliability issues  Double RF system operation in the LHC with all the complications (phase control,…)

Impedance identification (1/2)

Efforts during the last 2 years to identify the impedance sources in the SPS ring

• Beam measurements
• Measurements and electromagnetic simulations of impedance for different

devices/structures in the SPS ring

Synchrotron frequency shift  inductive part Enamelled QF – MBA ≈ 97 Non-Shielded, enamelled BPH – QF ≈ 39

Vacuum flanges

Long bunches with RF off  resonant impedance 1.4 GHz

Impedance identification (2/2)

• Vacuum flanges are the best candidate with strong peak at fr = 1.4 GHz (observed

also from beam measurements) with R/Q = 9 kΩ (different types,~ 500) Group I Group II Confirmed by particle tracking simulations: Nth increases by a factor of 2 without the impedance of vacuum flanges More studies for confirmation as the main source of mw instability

Impedance reduction

• Reducing the longitudinal impedance will reduce the Vind and uncontrolled

emittance blow-up  most robust solution

• Preliminary ideas of reducing the impedance of the SPS vacuum flanges

 Partial shielding + damping

• R/Q reduction factor 8 could be achieved.
• Only Group I (half) could be acted upon.
• 15-30 weeks of work

 Flange redesign

• Minimum impedance. R/Q reduction factor 20.
• All flanges could be changed (≈550).
• 15 – 30 weeks of work
• Higher cost (new elliptical bellows, …)

Limitation during the ramp

• The situation after the upgrade of the SPS 200 MHz RF system (2x4 + 4x3) is assumed
• For beam stability from certain energy (depending on intensity, emittance and optics) we need

to have controlled longitudinal emittance blow-up

• Optimistic scenario based on single bunch instability considerations  maximum emittance 0.7

eVs scaled from single bunch

Power/cavity for 2.2x1011 correction for PWD Voltage

3 sections 4 sections power limit power limit εl = (0.4 – 0.7 eVs) εl = (0.4 – 0.7 eVs)

Longer acceleration cycle

• With twice longer acceleration cycle we can accelerate HL-LHC intensity but
• Dedicated LHC filling is 30% longer
• Average power increase in SPS
• More time for instability to grow

3 sections 4 sections power limit power limit εl = (0.4 – 0.7 eVs)

Power/cavity for 2.5x1011 - Nominal Power/cavity for 2.5x1011 - Longer

εl = (0.4 – 0.7 eVs)

Possible improvement by redesign of the magnetic cycle

Intermediate transition energy Q22

 If power limitation is still an issue with Q20  Beam stability is an issue for Q26

• Intermediate transition energy with γt = 20 (Q22) (see talk of H. Bartosik)

Beam stability (single bunch simulations at SPS flat top)

Double RF – V200 = 7 MV

For same emittance Q22 provides:

• Better stability

compared to Q26

• Worse stability

compared to Q20

Intermediate transition energy Q22

Power consideration for longer acceleration cycle

• Controlled emittance

blow-up will be still needed for stability

• Longer cycle also

necessary due to power limitations during ramp

• More margin in power

compared to Q20 3 sections 4 sections power limit power limit

Power/cavity for 2.5x1011 – Longer cycle

εl = (0.425 – 0.8 eVs)

Summary

• The SPS intensity is limited by available RF voltage (due to beam loading) and

longitudinal emittance blow-up (due to instability during ramp)

• Limitations are coming from both the acceleration ramp (SPS losses) and SPS-LHC

beam transfer (LHC capture losses)

• For the 25 ns beam this limitation is now ~1.3x1011 p/b and expected to become

~2x1011 p/b after the 200 MHz RF upgrade (more, shorter cavities and higher power)

• Possible measures to reach intensities required by HL-LHC (~2.4x1011 p/b) are
• ramp: double the duration of the acceleration ramp
• 450 GeV/c: bunch rotation or installation of the 200 MHz RF system in the

LHC

• ramp & 450 GeV/c: impedance reduction (after confirmation of sources)
• Q22 optics: can give additional flexibility between Q20 and Q26 optics, but

Q20 is still considered the main option

Thank you for your attention!

Back-up slides

Δ𝜐𝑛𝑏𝑦= 0.11 ns Δ𝜐 𝑛𝑏𝑦 = 0.12 ns Bunch length evolution

Stable beam at flat top with intensity of 1.33x1011 p/b

Example of 25 ns beam during MD on 2012

Before extraction

Dipole oscillations Quadrupole oscillations

Dipole oscillations Quadrupole oscillations Bunch length evolution

Instability at flat top for intensities more than 1.35x1011 p/b

Example of 25 ns beam during MD on 2012

Δ𝜐𝑛𝑏𝑦= 0.81 ns Δ𝜐 𝑛𝑏𝑦 = 0.8 ns

Before extraction

SPS longitudinal impedance model

200 MHz TW RF Vacuum flanges

Maximum bunch intensity with Scenario no performance degradation 10% longer bunches, more losses in LHC 4 cavities, 750 kW - now 1.35x1011 1.45x1011 4 cavities, 1.05 MW pulsing, new LLRF 1.5x1011 1.7x1011 4 cavities with 1.05 MW & 2 cavities with 1.6 MW, new LLRF 2.1x1011 2.5x1011

Note: all estimations are done with simplified models and scaling from present experimental results

Summary for the 200 MHz upgrade scenarios (flat top)

• R/Q reduction factor 8 could be achieved.
• Only Group I (half) could be acted upon.
• 15-30 weeks of work

Proof of concept

Partial Shielding details (1/3)

Partial Shielding details (2/3)

Holes for spot welding

Partial Shielding details (3/3)

spot welds

Simulation models

Group I Group II

Flange redesign

• Minimum impedance. R/Q reduction factor

20.

• All flanges could be changed (≈550).
• 15 – 30 weeks of work
• Higher cost