Lessons from SPS studies in 2010

Lessons from SPS studies in 2010

E. Shaposhnikova

Chamonix’11 session 09: LHC injectors upgrade

Outlineo Review of the SPS MD studies in 2010o Expectations for proposed SPS upgradesAcknowledgments:

SPSU SG: G. Arduini, J. Bauche, C. Bhat, F. Caspers, S. Calatroni, P. Chiggiato, K. Cornelis, S. Federmann, E. Mahner, E. Metral, G. Rumolo, B. Salvant, M. Taborelli, C. Yin Vallgren, F. Zimmermann + speakers

+ OP shifts for help in MDs RF: T. Argyropoulos, T. Bohl, E. Ciapala, H. Damerau, W. Hofle, E.

Montesinos, G. Papotti (OP), J. Tuckmantel, U. Wehrle, G. Hagmann +…

LIU/TF: R. Garoby, B. Goddard, V. Mertens BT: M. Barnes, B. Balhan, R. Barlow, J. Borburgh

228/01/2011 SPS lessons

SPS MDs in 2010 (related to the SPS upgrade)

• Nominal LHC beam → e-cloud, beam stability– 25 ns, 50 ns, 75 ns, (150 ns) spacings

• Ultimate LHC beam → e-cloud, beam stability, transmission – 25 ns, 50 ns spacings

• Single bunch with small and nominal transverse emittances– nominal intensity (stability in double RF system)– ultimate intensity (TMCI, longitudinal instability)– very high intensities: up to 3.5x1011 injected (transmission)

• Low γt optics – single bunch, from nominal to very high intensities, TMCI

Note: not all data has already been analysed

10/11/2010 3

Summary of SPS MDs in 2010

Week Date Spacing Max inj. int. Comments

17 27-29.04 25 ns nominal scrubbing run, dedicated SC

22 02-03.06 25 ns ultimate 1-4 batches, dedic. SC (part)

29 20-21.07 25 ns nominal practically lost

35 03-04.09 50 ns25 ns

ultimatenominal

1-4 batches

42 19-20.10 25 ns50 ns

nominalnominal

36-72 bunches; dedicated SC → 1-2 batches

45 09.11 50 ns nominal floating MD

46 17-18.11 75 ns nominal

28/01/2011 SPS lessons 4

Ultimate 25 ns beam

• Large efforts in whole injector chain, in the PSB and PS in particular (thanks!) to prepare ultimate intensity LHC beam in advance

• From 12:00 on 2.06 till 20:00 on 3.06 (too long for the same people involved) – stopped due to MKE heating to 70 deg

• Injection: up to 1.9x1011/bunch, longitudinal emittance ~0.4 eVs, transverse emittances ~5 μm

• 12, 24, 48, 72 bunches, then 1,2,3,4 batches

• Nothing broken but issues with MOPOS and FBCT • No e-cloud observations (no StSt liner in the ring)• Significant bunch length variation on flat top

24/06/2010 SPSU SG 5

Ultimate 25 ns beam

• Increasing number of batches - decrease of bunch intensity on FT: 1.6 x1011 in 1 batch, 1.52x1011 in 2 , 1.48x1011 in 3, 1.4x1011 in 4.

• To reduce losses the 200 & 800 MHz voltage programs were modified: 0.65 → 0.75 eVs (constant bucket area)

• Voltage on FT - 5.5 MV (nominal and max 7.5 MV) – to reduce effects of heating and outgassing

• Very large beam losses: 30 % at the beginning, 20 % at the end• Chromaticities: ξH=0.2 and ξV=0.3

• Transv. emittance blow-up during the ramp: 5 → 10 μm

(horizontal blow-up larger, also increased for more batches)• Beam was very unstable longitudinally on flat bottom (already with

12 bunches )24/06/2010 SPSU SG 6

Ultimate 50 ns beam

• w35, at the end of MD block (Friday, 12:00-20:00) - in || to the LHC set-up (with 150 ns beam)

• nominal 25 ns and 50 ns also (for e-cloud comparison)• ultimate intensity (1.75x1011/bunch) injected → maximum

1.52x1011/bunch on FT, ~15% losses• H/V emittances on FT – 2.7/2.8 μm for nominal beam,

injected 3.2/3.9 μm for ultimate • chromaticity (0.05/0.18) increase had no effect on losses • voltage programme as for 25 ns nominal


e-cloud vs bunch intensity for 25&50 ns spacing (MD w35)


C. Yin Vallgren et al

- A factor 3-5 difference between 25 ns and 50 ns beams- Some increase of e-cloud current with intensity for 50 ns beam (losses?)

Intensity limitations for 25 ns beam - 2010

intensity/bunch

Origin Leads to Present/future cures/measures

0.2x1011 longitudinal multi bunch instability due to longitudinal impedance

- beam loss during ramp- bunch variation on FT

(FB, FF, long. damper) - 800 MHz RF system- emit. blow-up → RF

0.5x1011 e-cloud due to the StSt vacuum chamber (δSEY=2.5, 1.3 is critical for SPS)

- dynamic pressure rise- transv. (V) emit. blow-up- instabilities- losses (via high chrom.)

- scrubbing run (δ→1.6)- high chrom. (0.2/0.4)- transv. damper (H)- (50/75 ns spacing)- a-C coating (δ→1.0)

1.2x1011 not known exactlye-cloud/impedance/SC(?)

- flat bottom/capture beam loss (>5%)

- (lower chromaticity)- WP, RF gymnastics- collimation

1.5x1011 beam loading in 200 MHz RF system

- voltage reduction on FT- phase modulation

- Feedback & FF - RF cavities shortening

1.6x1011 TMCI (transverse mode coupling instability) due to transverse impedance

- beam losses- emittance blow-up

- higher chromaticity- high voltage- transverse high bw FB 928/01/2011 SPS lessons

Intensity limitations

• Single bunch – the same for all beams – space charge – TMCI (transverse mode coupling instability)– loss of Landau damping due to incoherent frequency shift– longitudinal instability

• Multi-bunch – worse for higher total intensity– e-cloud → Migel’s talk– beam loss (many reasons)– longitudinal coupled bunch instabilities – beam loading in the 200 MHz and 800 MHz RF systems– heating of machine elements (MKE, MKDV kickers, …)– vacuum (beam dump and MKDV outgassing), septum sparking (ZS was a main limitation in 2008 and 2009)


Transverse Mode Coupling Instability (TMCI)

• Many MDs (B. Salvant et al.)• PS intensity up to 3.5x1011,

εL =0.35 eVs, τ=3.8 ns • SPS end FB: (2.25-3.3)x1011

for ξV in range 0.0-0.3, ξH=0.25

• Threshold scales (matched voltage) ~ εLη

• Cures: high chromaticity, εL, impedance reduction, transverse wide-band FB (W. Hofle&LARP)

• Higher TMCI threshold is expected (from theory and simulations) for smaller transverse emittances due to space charge effect


N=1.8x1011 injected – 5.5% lossN=1.4x1011 - no visible loss on FBThreshold ~ 1.6x1011 for zero ξ

B. Salvant

Very high intensity single bunch

• First MDs: PS εH/V ~1.3 μm, then εH/V=2.5/2.6 μm

• Need to blow up inject. transv. emittance (in TL or PS) to reduce losses and blow-up in the SPS

• εH/V=3.3/3.6 μm for N >1.8 1011

• εH/V=5-6 μm for N >2.5 1011

• Longitudinal instability for N >1.4 1011 (in single and double RF)


εH/V =6/6 μm

εH/V = 3.15/4.25 μm

losses for small injected emittances

Some data for space charge

• ppbar time - ∆Q=0.07• Protons at 14 GeV/c (H. Burkhardt et al., PAC 2003) ∆Q=0.14/0.18 with

10% losses (N=1.2x1011, 3 ns, εH/V=3.43/3.75 μm – 30%?, )

• Nominal LHC bunch ∆Q=0.05, ultimate ∆Q=0.07• 50 ns nominal intensity beam with single batch injection in PS (2008):

εH/V=1.1/1.4 μm at 450 GeV/c (E. Metral) →∆Q=0.15

• Recent studies with high intensity single bunch (B. Salvant et al., 2010)

2.5x1011 → ∆Q=0.1 for ε=3.5 μm• LHC ions in the SPS: γ=7.31, Ne=1.5x1010, (50% more than nominal), ε=0.5

μm (1/2 nominal). In DR ∆Q=0.08 → ∆Q=0.24… but with 25% losses

→ Space charge limit alone seems to be more close to ∆Q=0.15

→ Interplay with other effects (multi-bunch) is probably also important

Reminder: significant loss reduction for nominal LHC beam by change of the WP (G. Arduini et al., 2004)


Transverse emittance vs bunch intensity(space charge limit?)


Transverse emittances versus bunch spacing (same total and bunch intensities)

50 ns spacing 25 ns spacing

15

0 1 2 3 4 52.0

2.5

3.0

3.5

4.0

SPS MD 2010/10/19, Injection25 ns, 36 bunches/batch, ~1.2E11/bunch

e_H(um)e_V(um)Number of BatchesC. Bhat

Vertical emittance increase with n batches, measurement at 0.55 s (26 GeV)


0 1 2 3 4 50.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

SPS MD 20101110, 450 GeV50 ns, 36 bunches/batch,

~1.2E11/bunch

Number of Batches

No emittance increase with n batches, small (<10%) blow-up during the cycle

Vertical emittance on FT (2010)


Transverse emittance

• For nominal beam with 50 ns and 75 ns spacing emittance on FT is mainly defined by injected emittance, only very small blow-up was observed after tuning

• Nominal beam at 25 ns – 15% blow-up from 0.55 s to FT in H-plane, no more blow-up in V-plane

• Large blow-up for 25 ns ultimate intensities was measured under not-optimised conditions

• No measurements at FT for 50 ns ultimate beam


Transverse emittance measurements in SPS

• Measurements during the cycle and along batch(es) are essential to study origin of emittance blow-up (if any)

• Measurements with Wire Scanners (WS) in 2010:– Average for all bunches (no bunch-by bunch)– One measurement per cycle (difference between “in” & “out”)– First measurement at 550 ms after injection

• BI improvements for 2011 (L. Jensen):– new electronics for 2nd WS (linear, now broken) with possibility to gate

acquisition (over 50-100 ns, as in the past)– cross-calibrations (WS 1&2, “in”&“out”, PS&LHC) – expert involvement (settings are critical) plus fellow(?)– BGI (rest gas) monitor – continuous beam profile measurements during

cycle, average for all bunches over 20 ms


Longitudinal multi-bunch instability50 ns beam, 2 RF systems, no RF noise

Short PS bunches Long PS bunches

19

T. Argyropoulous et al.


Multi-bunch instability due to loss of Landau damping?

• There is a narrow window for the injected parameters – losses increase for larger bunch length and beam is unstable for lower emittance (blow-up is required also for 50/75 ns beam )

20

0.46 eVs

H. Damerau et al.

50 ns beam


Effect of beam loading in the 200 MHz RF on emittance blow-up by band-limited noise


T. Argyropoulos et al., HB2010

Low γt - solution for everything?

• Successful MDs with a single bunch (H. Bartosik, Y. Papaphilippou et al.): γt=22.8 → 18, increase in η: 2.86 (@26 GeV/c) and 1.6 (@450 GeV/c)

• Expected increase in beam stability for the same bunch parameters Nth~ η for TMCI (observed!) and longitudinal instabilities (to be seen in 2011)

• To have the same long. parameters (bucket area) V ~ η. Already maximum voltage (7.5 MV) is used now for extraction to LHC → longer bunches with the same long. emittance and voltage → capture losses in LHC

• But probably emittance blow-up for the same intensity can be reduced:

loss of Landau damping Nth~ ε2 η τ. Since τ ~ (ε2 η/V)1/4 → ε ~ η-1/2 and τ = const for V=const.

• If LHC itself needs higher longitudinal emittance at injection (IBS, stability) → 200 MHz RF system in LHC?

• Fast cycles are not possible (3 s or 4.2 s acceleration time)• e-cloud effects - ?


Low γt - solution for everything?

No TMCI Small emittances• FB: no transverse blow-up for

– εH/V = 2.0/2.3, 2.6x1011

– εH/V =2.5/2.6, 3.3x1011

but too low voltage (1.8 MV) → losses (10-15%) and longer bunch (30%)

• Acceleration of 2.5x1011

– 5% capture losses– εH/V = 2.4/2.9,

– τ = 1.5 ns on FT• Very promising → more studies


Some results for 2010

• Low beam losses (5%) for nominal LHC intensity with low vertical chromaticity (0.1)

• No TMCI at injection up to ultimate bunch intensity. Ultimate bunch accelerated to 450 GeV/c with low loss (and low chromaticity). More problems/losses for small transverse emittances. Some transv. emittance blow-up.

• Ultimate LHC beam injected

– 25 ns: large losses, unstable beam on FT, large emittances – 50 ns: 10% losses

• No limitation from ZS after change of settings by BT group and from MOPOS at the end of run

MD limitations: MKE heating and outgassing (in MD SC), transverse

emittance measurements , MD time allocation

10/11/2010 24

SPS beams

Beam parameters

SPS at 450 GeV/c (maximum injected minus losses)

LHC LHC LHC LHC LHC LHC

bunch spacing ns 25 50 75 25 50 indiv

max bunch intensity 1011 1.2 1.2 1.2 1.4 1.52 3.2

number of bunches 4x72 4x36 4x24 4x72 4x36 1

total intensity on FT 1013 3.5 1.7 1.2 4.0 2.2 0.03

long. emittance eVs 0.7 0.5 0.4 0.8 0.6 0.4

norm. h/v emittance μm 3.6 2.0* 2.0* ~10 >3.2/3.9

6.0

25

* double batch injection in PS: 1.1/1.4 (E. Metral et al., 2008)


Questions

• source of limitations/bottlenecks (up to ultimate intensity) → e-cloud (25 ns), beam loading, coupled bunch instabilities, TMCI

• possible cures and mitigation measures → e-cloud mitigation, RF system modification, impedance reduction,

low gamma transition optics• p/b and emittance as a function of the distance between bunches

today and after upgrade– extracted emittance is a function of intensity, injected emittance,

distance between bunches, number of batches and machine settings • what should be done for delivering smaller transverse emittances at

ultimate beam current?→ beam from PS for studies, improvement of transverse emittance

measurements, e-cloud mitigation (25 ns spacing), transverse impedance reduction, transverse damper/FB improvement

10/11/2010 26

Summary

• Main SPS limitations for ultimate intensity with nominal transverse emittances have been identified and possible cures suggested

• MDs needed to see other possible limitations• MDs with small injected transverse emittance beam and

proper instrumentation are required to study origin of blow-up for high intensities

• Cures for longitudinal instability (800 MHz and emittance blow-up) are limited – consider other solutions to recover Landau damping

• Kickers give dominant contribution to impedance – reduction

• Low gamma transition optics is a very promising solution for beam stability, transfer to LHC should be studied

• Still more results to come for 2010 MDs


“Scrubbing run”

• w17 (April 27-29) – “scrubbing run”, cycle LHC1– 1 day lost (magnet exchange)– after careful OP and RF Low-Level set-up - losses 5%– chromaticity 0.1 (even 0.05)– heating and outgassing of different kickers: MKDH3,

MKP and MKE - limitation for 4 batches accelerated (M. Barnes talk at SPSU SG)

– H/V emit: PS ext. 2.0/1.5 → inj. SPS 3.2/3.3 μm, on FT 3.0/3.6; inj. later 2.3/4.2 μm


Horizontal emittance on FT (2010)


Lessons from SPS studies in 2010

Documents

Transcript of Lessons from SPS studies in 2010